What explains the reduction of urban violence in Medellín, ?

Caroline Doyle

A thesis in fulfilment of the requirements for the degree of Doctor of Philosophy

The School of Business

January 2017 Acknowledgements

This thesis is the result of a challenging, rewarding and inspiring journey. I owe a debt of gratitude to the many people who have made that journey a life-changing experience.

First, I am immensely grateful to my supervisors, Professor Mark Turner and Assistant Professor Anthea McCarthy-Jones. They have played a supportive and inspirational role to this project and it is their scrutiny and inquisitive nature which have helped to shape this thesis. I am also grateful to the support of Professor Michael O’Donnell in the second part of this journey.

Second, thanks to my parents for their practical, emotional and monetary support throughout this journey. My parents have been an irreplaceable source of support and inspiration throughout my academic journey.

Third, I am extremely grateful to the all the paisas I encountered in Medellín who gave their time to speak to me about violence in the city and its reduction. Given the nature of the research, I have decided not to name some of the contacts who assisted me during my time in Medellín. But, I am grateful to all, especially my initial contacts in the city who provided me with accommodation, helped me to establish contacts in the city and for the use of their finca when I needed a quiet space to do some writing. Thank you to my paisa housemate for welcoming me, helping me with my paisa avanzada and for answering my many questions about Medellín. Lastly, I would also like to thank my colleagues at the Universidad de Antioquia and the Universidad EAFIT who provided me with timely advice, encouragement, support and inspiration both during and after fieldwork. The lessons I learnt through these colleagues deeply impacted on this project and made it such an enriching experience.

Finally, thanks to my colleagues at the University of Canberra and University of New South Wales who have provided me with much needed emotional and practical support throughout my doctoral studies.

ii

List of Publications Arising From This Thesis

Referred Journal Articles

Doyle, C. 2016. Explaining patterns of urban violence in Medellín, Colombia. Laws, 5(3).

Refereed Conference Paper

Doyle, C. 2015. Public policy or public action? Explaining declining patterns of urban violence in Medellín, Colombia. Paper presented at Australian Political Science Association Conference 2015, University of Canberra 28-30 September.

iii

Table of Contents

Acknowledgements ...... ii List of Publications Arising From This Thesis ...... iii List of Abbreviations ...... vii Glossary of Key Spanish Terms ...... x List of Figures ...... xiv Chapter One: Introduction ...... 1 Medellín, Colombia as Case Study ...... 3 Aims of Research ...... 4 Research Questions ...... 5 Importance of this Research ...... 5 Fieldwork in Medellín, Colombia ...... 6 Structure of the Research ...... 7 Chapter Two: Literature Review ...... 10 Urban Violence ...... 10 Emergence of Violence in Urban Environments...... 13 The grievance perspective and greed hypothesis ...... 13 Structural violence ...... 15 Different Causes or Circumstances to Explain the Emergence of Urban Violence ...... 16 Population increase and density ...... 16 Weak or ineffective state institutions ...... 18 Inequality and poverty ...... 19 Perpetrators of Violence in Urban Environments ...... 21 Use of violence ...... 25 Challenges to the state and ungoverned spaces ...... 26 Approaches to Reduce and Prevent Violence in Urban Environments ...... 28 Social disorganisation ...... 30 Broken windows ...... 31 Social capital ...... 33 Context-specific approaches ...... 35 Towards an Analytical Framework ...... 36 Chapter Three: Methodology ...... 38 Introduction ...... 38 Case Study ...... 38

iv

Selection and Justification of Case Study ...... 39 Measuring Levels of Violence in Urban Environments ...... 40 Fieldwork in Medellín ...... 41 Semi-structured interviews ...... 42 Participant observation ...... 44 Document collection and analysis ...... 48 Researching Violence and Conflict ...... 48 Triangulation of Data ...... 51 Chapter Four: The Emergence of Urban Violence in Medellín ...... 53 The Most Progressive City in Colombia ...... 53 Growing Inequalities in Medellín ...... 58 The Growth of Narco-Trafficking ...... 60 : Antioquia’s Robin Hood ...... 63 Escobar: Plata o Plomo? ...... 66 The Rise of the Militia ...... 70 Conclusion ...... 74 Chapter Five: The 1990s: Medellín, ‘the Most Dangerous City in the World’ ...... 76 Medellín: The Battleground of the Medellín Cartel ...... 76 ‘The Most Dangerous City in the World’ ...... 79 The Presidential Council: A Strategy to Address the Violence ...... 81 The Militia: Semi-criminal and Semi-legal Organisations ...... 87 Peaceful Co-existence Pacts with the Bandas ...... 91 The Paramilitaries Enter Medellín ...... 95 Conclusion ...... 100 Chapter Six: 2000-2015: A Significant Reduction in Homicides ...... 102 A New Wave of Violence ...... 102 Pacification of the City ...... 105 Integral Urban Projects...... 112 The Medellín Miracle ...... 118 A ‘Tenuous’ Reduction in Homicides ...... 122 ‘Reformers have Combatted Crime and Poverty in Medellín’ ...... 127 ‘The Most Innovative City in the World’...... 131 The Bandas have Captured the Medellín Miracle ...... 141 Conclusion ...... 148 Chapter Seven: Explaining the Reduction of Urban Violence in Medellín ...... 150

v

The Emergence of Violence in Medellín ...... 151 Policy Approaches to Urban Violence Reduction in Medellín ...... 153 The Perpetrators of Violence (Homicides) in Medellín ...... 161 Changing Patterns of Violence in Medellín ...... 164 Conclusion ...... 169 Chapter Eight: Conclusion ...... 172 Reference List ...... 181 Appendix 1: Semi-structured Interview Questions ...... 210

vi

List of Abbreviations

ACCU Autodefensas Campesinas de Cordoba y Urabá (Peasant Self- Defence Forces of Cordoba and Urabá)

AUC Autodefensas Unidad de Colombia (United Self-Defence Forces of Colombia)

BACRIM Bandas Criminales (Criminal gangs)

BCN Bloque Cacique Nutibara (Cacique Nutibara Block)

BM Bloque Metro (Metro Block)

CAP Comandos Armados del Pueblo (The Peoples Armed Commandos)

CAI Comando de Acción Inmediata (Centres of Immediate Attention)

CLASCO El Consejo Latinoamericano de Ciencias Sociales (Latin American Council of Social Sciences)

CONVIVIR Servicios Especiales de Vigilancia y Seguirdad Privada (Special Services Surveillance and Private Security)

CTI Cuerpo Técnico de Investigación (Colombian Office of the Attorney General)

DEA Drug Enforcement Agency

DECYPOL Departamento de Estudios Criminológicos e Identificación (Department of Criminological Studies and Identification)

ELN Ejército de Liberación Nacional (National Liberation Army)

FARC Fuerzas Armadas Revolucionarios de Colombia (Revolutionary Armed Forces of Colombia)

HASOW Humanitarian Actions in Situations Other Than War

vii

HDI Human Development Index

INML Legal y Ciencias Forenses Instituto Nacional de Medicina (Colombian National Institute of Medicine: Legal and Forensic Sciences)

IDRC International Development Research Centre

M-19 Movimiento Abril de 19 (19th of April Movement)

M-6 & 7 Milicias 6 y 7 de Noviembre (November 6 and 7 Militia)

MAS Muerte a Secuestadores (Death to Kidnappers)

MILF Moro Islamic Liberation Front

MMP Plan Piloto (Medellín Master Plan)

MPPP Milicias Populares del Pueblo y para el Pueblo Popular (Popular Militias of the People and for the People)

MPVA Milicias Populares del Valle de Aburrá (People’s Militia of the Valley of Aburrá)

NGO Non-Government Organisation

OSHM El Observatario de Seguridad Humana de Medellín (Human Security Observatory of Medellín)

PEPES Perseguidos por Pablo Escobar (People Persecuted by Pablo Escobar)

PRIMED Programa Integral de Mejoramiento de Barrios Subnormales en Medellín (Program of Subnormal District Improvement in Medellín)

PUI Proyectos Urbanos Integrales (Integral Urban Projects)

SIJIN Seccional de Investigación Criminal (Department of Colombia’s Criminal Investigation)

viii

SISC Sistema de Información para la Seguridad y la Convivencia (System of Information for Security and Cohabitation)

UN United Nations

UNDP United Nations Development Programme

UNODC United Nations Office on Drugs and Crime

Urbam Centro de Estudios Urbanos y Ambientales (Centre of Urban and Environmental Studies)

US United States

USAID United States Agency for International Development

WHO World Health Organization

ix

Glossary of Key Spanish Terms

Alcaldía de Medellín Local government of Medellín.

Alcalde Mayor.

Antioquia The region in which Medellín is located.

Antioqueños The people from the region of Antioquia.

Banda One name for a gang in Medellín.

Bazuco Cheap and highly addictive cocaine paste imilar to crack cocaine.

Bloque de Busqueda Search Bloc created in 1992. It had the objective of apprehending Pablo Escobar and his associates.

Campamentos de Paz Peace Camps set up in the early 1980s in Medellín following the ceasefire with the M- 19 left-wing guerrilla group.

Centro de Analisis Politico Centre of Political Analysis. A research centre of the University EAFIT in Medellín.

Chichipato A ‘low-life’ individual or gang member who commits crime in his or her own neighbourhood and normally to satisfy their drug addiction.

Combo One name for a gang in Medellín.

Comuna An administrative unit containing a group of neighbourhoods in Medellín. There are 16 comunas in Medellín.

x

Consejería para el Área Metropolitana de Presidential Council for Medellín and its Medellín Metropolitan Area.

Encapuchado Hooded vigilante, usually associated with social cleansing but sometimes as part of a death squad.

Favelas Shanty towns or slums

Frente Nacional Power-sharing agreement between the Liberals and the Conservatives following .

Fronteras Invisibles Invisible borders which mark out territory between different criminal groups in Medellín.

La Violencia A civil war from 1948 to 1958 between the Liberals and Conservatives which ended in 1958 with a power-sharing agreement called the National Front.

Los Extraditables A group in the 1980s which campaigned that the Extradition Treaty with the US was harming the national sovereignty of Colombia.

Medellín Como Vamos A program which evaluates changes in the quality of life in different socio-economic areas of Medellín.

MetroCable A form of public transport using ski-lift technology.

Mona Local term for a blonde female.

xi

Narco-para Groups protecting the economic interests of cocaine production facilities that are also affiliated with right-wing paramilitaries.

Oficina Office. A term usually used to refer to a criminal organisation that connects customers with trained assassins and other criminal services.

Oficina de Envigado A network of criminal actors in Medellín, sometimes referred to as ‘la oficina.’

Pacto de Fusil Agreement between the Oficina de Envigado and Los Urabeños in July 2013.

Pactos de Conviviencia Peaceful coexistence pacts formed in some neighbourhoods with the bandas in the late 1990s.

Paisa A person born and raised in or around Antioquia.

Parque Bibloteca Library Park.

Plata o Plomo Silver or bullet. A generous bribe (plata) offered to public officials. If these public officials were not cooperative they were met with plomo, a rain of bullets.

Policia de Cuadrantes Policing by neighbourhood quadrants.

Sicario Hired assassin, usually working as part of a banda.

Urbanismo Social Social Urbanism. Initiatives that invested in infrastructure and other projects for the poorer areas of Medellín.

xii

Vacuna Payment made to non-state armed groups by community members for security

xiii

List of Figures

Figure 1: Map of Colombia...... 3

Figure 2: Homicide Rates: Medellín, Colombia 1990-2014 ...... 4

Figure 3: Targeting Urban Violence Prevention and Reduction Interventions ...... 29

Figure 4: Map of Medellín, Colombia ...... 58

Figure 5: Homicide Polygons in Medellín ...... 135

xiv

xv

Chapter One: Introduction

The world is currently undergoing the largest wave of urban growth in history. In 1950, there were 80 cities with populations over a million, but in 2015 there were over 500. By 2045, it is predicted that more than three-quarters of the world’s population will live in cities. This urban growth is of great concern to policy makers in developing nations, as most of it is occurring in the slums of these nations. These policy makers have acknowledged that while urban growth can bring improved access to employment and services for the poor, it can also bring challenges, such as escalating levels of conflict and violence. In the last 15 years, scholars from a wide range of fields have predicted that these growing urban environments could be the future sites of national insecurity and urban insurgencies (Patel & Burkle 2012; Kilcullen 2013; Moser & McIlwaine 2014; Winton 2014; Kalyvas 2015; De Boer & Bosetti 2015; Muggah 2012).

In the academic debate on urban violence there is strong consensus on the effect high levels of violence can have. For example, the effects can be economic, such as an increased expenditure on law enforcement, taking away spending from social services and dissuading investors, with the Mexican government estimating that in 2007 alone, violence caused a loss of nearly $9.5 billion to the country’s economy (World Bank 2011a; UNODC 2007). The adverse effects of urban violence can also be social with neighbourhoods divided and stigmatised (World Bank 2011a).

There is a large amount of research and analysis from urban environments in high-income countries on policy approaches to reduce and prevent violence with the most influential approaches being social disorganisation, broken windows and social capital. For example, aspects of social disorganisation theory have been tested in a variety of high-income urban environments, including Brisbane, Australia (Mazerolle et al. 010); Freiberg, Germany (Oberwittler 2004); Indianapolis, Rochester, Tampa-St. Petersburg and St. Louis, United States (US) (Veléz 2001); and Stockholm, Sweden (Sampson & Wikström 2008). The zero-tolerance policy approaches used in New York City in the US in the 1990s were based on a broken windows approach and were largely credited for the decrease in violent crimes and property crimes (Kelling & Sousa 2001). More recently, following the London riots in the United Kingdom in 2001, reports by ‘community cohesion review teams’ attributed the violence to an ‘absence of adequate interconnectivity and direct engagement of community in defining their

1 own security and safety priorities to prevent this violence’ (Community Cohesion Review Team 2001).

While there is acknowledgement of the effects of violence in the urban environments of high- countries environments, urban environments in middle- and low-income countries have not received this same attention. For example, it was only recently, in 2011, that the World Bank published a first global study, Violence in the City: Understanding and Supporting Community Responses to Urban Violence, (World Bank 2011a) focused on urban violence in middle- and low-income countries. Another study which provided an important overview of the state of research on emerging practices and approaches to reduce urban violence in high-middle-and- low-income countries was only published in 2012 by Canada’s International Development Research Centre (IDRC) (Muggah 2012). In addition to this urban violence research which has focused on the urban environments of high-income countries, a large amount of urban violence research has used frameworks which involve the relationship between the state, poverty and fragility to explain urban violence. For example, the World Bank’s 2011 World Development Report (World Bank 2011b) considered urban violence in relation to national state institutions and in the contexts of civil war, using approaches such as the grievance perspective and greed hypothesis.

This limited analysis and testing of approaches to reducing urban violence in urban environments of middle- and low-middle- and low-income countries is of concern to policy makers in these countries. These urban environments, particularly Latin American ones, are already experiencing high levels of urban growth and high levels of violence. For example, from 2000 to 2010, homicide rates in the region of Latin America increased by 12 percent, while most regions in the world experienced an almost 50 percent decrease (UNDP 2013). By 2015, Latin America was the world’s most violent region with 9 percent of the world’s population but 33 percent of global homicides and contained 47 of the 50 most dangerous cities in the world (Muggah & Giannini 2015).

Reducing these levels of violence and preventing future outbreaks is of utmost concern to policy makers in the region. This thesis adds to the growing literature on urban violence in middle- and low-middle- and low-income countries by focusing on the Latin American city of Medellín, Colombia and exploring the questions of: What explains the emergence of urban violence? Who are the perpetrators of violence? What explains the reduction of urban violence?

2

These are questions that are of great importance to policy makers in Latin America and are the leading questions for this thesis.

Medellín, Colombia as Case Study Figure 1: Map of Colombia

Image Removed due to Copyright

Source: Global Affairs Canada 2013

For this research, Medellín, Colombia has been chosen as a case study to understand the contemporary phenomenon of violence and its reduction in an urban environment in a developing country. Colombia has a population of just over 48 million and is located in the north-western corner of South America, alongside the Caribbean Sea between Panama and Venezuela and bordering the Pacific Ocean between Panama and Ecuador and sharing a border with Brazil and Peru to the southeast (World Bank 2016a). Colombia has a Human Development Index (HDI) of 0.720 making it a high human development country. However, their HDI is below the average for countries in the high human development group and below the average for countries in Latin America and the Caribbean (UNDP 2015). Furthermore, Colombia is the eleventh most unequal country in the world and the second most unequal country in Latin America (after Honduras), with a very high Gini coefficient of 0.535 (World Bank 2016a). When these levels of inequality are ‘discounted’ from Colombia’s HDI the country has an inequality-adjusted HDI of 0.542, putting it in the middle to low human development category, below Zambia and Cambodia (UNDP 2015).

Medellín is the second largest city in Colombia with a current population of two million and is a particularly interesting case for examining the emergence and reduction of violence in an urban environment for three reasons. First, the city has experienced such extremely high levels of violence. When Medellín was portrayed as the ‘most dangerous city in the world’ with 381 homicides per 100,000 inhabitants, its homicide rate was almost 40 times higher than the United Nations definition of epidemic violence. At the time, this level of violence was only exceeded by Beirut in Lebanon, where civil war was raging. Second, Medellín represents a complex case involving multiple perpetrators of violence including drug cartels, gangs, youth assassins, urban guerrillas, community militia and paramilitaries. Third, no other city had

3 experienced such a dramatic reduction in homicides in less than a generation. There was a reduction of over 90 percent in homicides from 1991 to 2015 in Medellín (Figure 2).

Figure 2: Homicide Rates: Medellín, Colombia 1990-2014

Image Removed due to Copyright

Source: SISC 2016

The reduction in homicides led to the city being referred to as the ‘Medellín Miracle’ and as a benchmark for Latin American policy makers seeking to understand and implement successful urban violence reduction and prevention policies. International organisations, such as UN- Habitat, Inter-American Development Bank and the World Bank have praised the policy approaches used in the city for this reduction of homicides (UN-Habitat 2007; Inter-American Development Bank 2009; World Bank 2011a).

Aims of Research

There are four specific aims of this research. The first aim is to identify what explains the emergence of violence in Medellín. Thus, the first data chapter of this thesis examines key events and issues relating to the emergence of high levels of violence in Medellín. The second aim is to explain the decline of violence in Medellín. This will be achieved by analysing the policy approaches used to reduce and prevent violence. Thus, the first and second data chapters examine the assumptions policy makers made about why violence emerged and the policy solutions to reduce this violence. The third data chapter examines the period when Medellín’s homicide rate declined from appallingly high levels to moderate ones and the policy approaches which were implemented to reduce the violence and prevent future outbreaks. The third aim of this thesis is to further understand perpetrators of violence in an urban environment and how they use violence (homicides). Thus, the three data chapters pay attention to the characteristics, methods and motivations of the perpetrators of violence in Medellín. The final aim of this thesis is to make a contribution to literature on urban violence, particularly in middle-and low-income countries. This will be achieved by examining key events and issues relating to the emergence

4 and decline of violence in Medellín from the 1980s to November 2015. This examination will provide insights which may have value for policy makers in the urban environments of middle- and low-income countries.

Research Questions

This thesis seeks to explain the rise and the reduction of violence in Medellín, a large city in Latin America. To accomplish this, this thesis is guided by the following major question: ‘What explains the rise and decline of urban violence in Medellín?’ This can be broken down into a set of sub-questions:

1. What explains the emergence of violence in Medellín? 2. What factors explain the decline of violence in Medellín? 3. Who are the perpetrators of violence? 4. How do these perpetrators use violence? 5. What implications does the Medellín experience have for Colombia and for addressing urban violence more generally?

These questions guide this research. In making reference to violence, and the reduction of violence, in urban environments the thesis uses homicides as the indicator as they are the most widely used measure for accessing and assessing levels of violence in urban environments (World Bank 2011a).

Importance of this Research

The importance of this research is threefold. First, this research on urban violence in the context of Medellín tests both current theories and practices. As previously stated, the majority of research on urban violence has focused on cities in high-income countries. By focusing on a Latin American city, this research will make a significant contribution to understanding effective policy approaches in middle- and low-income countries. Second, by analysing the perpetrators of violence this research will lead to greater understanding of their characteristics, motivations and methods, matters that have been largely overlooked in existing studies. Third, this research is important because of the contribution it makes to the methodological literature on researching violence in cities of middle- and low-income countries. Policy-focused research on urban violence often uses objective indicators that are amenable to portraying the incidence of urban violence, for example homicide rates in urban environments (World Bank 2011a).

5

However, the use of qualitative research enables a deeper understanding of how citizens perceive violence and the effect it has on their daily lives. The qualitative data collected for this thesis provide deep and rich understanding of the complexities of urban violence. The insights in this thesis can assist policy makers in designing policies to reduce and prevent violence. Thus, while using statistical indicators, such as homicide rates, this thesis relies heavily on qualitative data to make an important contribution to the academic literature and policy-focused analysis on urban violence. Given the reliance on qualitative data in this thesis, a brief discussion on the data collection methods is provided next.

Fieldwork in Medellín, Colombia

After conducting a review of the literature on urban violence and sketching out my research plans I knew it was important to conduct fieldwork in Medellín. What I wished to achieve was only possible with fieldwork for primary data collection. Thus, I engaged in two periods of fieldwork; May 2014-August 2014 and April 2015-August 2015. The purpose of this fieldwork was twofold. First, I needed to conduct semi-structured interviews with actors who had knowledge about the emergence of violence, the perpetrators of violence and the reduction of violence. The second purpose of this fieldwork was to engage in participant observation and obtain a deeper understanding of how citizens perceive violence and the effect this violence has had on their daily lives.

As is further discussed in the methodology chapter of this thesis, during the two fieldwork periods, semi-structured interviews were conducted with an array of actors including government officials, former government officials, academics, business leaders, and members of community and civil society organisations. These actors all had valuable knowledge about the emergence and decline of violence as well as the different perpetrators of violence in Medellín. These local actors provided their perspectives on how levels of violence are perceived by different groups and the effect this violence has on their daily lives. The questions asked to these participants focused on themes that had emerged from the literature concerning explanations for the emergence of violence in Medellín and interventions that reduce violence. The interview questions were open-ended and designed to ensure the interviews were more of a conversation with the participants rather than following a set script (Liamputtong & Ezzy 2005, 55). They also allowed a number of topic-related questions about the patterns of violence in Medellín and suggested further questions to be asked, allowing for wide-ranging discussion

6 on the research topic (Burnham et al. 2008, 118). The participants were provided with the questions and the research objectives before I conducted the interview with them. The insights provided by these actors make an essential contribution to the studies of urban violence, as it is the perspective of these local actors that allow a deeper and richer understanding of the complexities of urban violence.

Another theme discussed in the methodology chapter is the difficulty of collecting data in a volatile environment, such as one with high levels of violence and conflict. The methodology chapter outlines how the nature of the research topic and my personal identity as an outsider, necessitated that I was creative and flexible, and adopted a wide range of data collection methods to achieve the objectives of this research and ensure that the data presented provide an accurate representation of the realities of violence in an urban environment.

The observations and insights from these fieldwork periods are referred to throughout this thesis and provide an important contribution to the study of urban violence as they present a deep and rich understanding of the complexities of violence in a Latin American urban environment.

Structure of the Research This thesis is structured into eight chapters, the contents of which are briefly described below.

Chapter Two reviews the literature that specifically focuses on urban violence and policies to reduce or prevent it. The chapter examines the literature on the different approaches scholars have employed to explain high levels of violence in urban environments. There is also a discussion on gangs, which have been identified in the literature as the main perpetrators of violence in Latin American urban environments.

Chapter Three provides a detailed description of the methodology I have used for data collection for this research. The chapter outlines key components of the methodology and provides an explanation and justification for my particular approach to the research. These components include case study analysis, gathering information from interviews and participant observation during two periods of fieldwork in Medellín, and document and archival analysis. A discussion on the risks involved with researching in areas with high levels of violence and conflict.

7

Chapter Four is the first data chapter and provides a historical narrative on the growth of urban violence in Medellín in the 1980s. There is a discussion of the themes and concepts that might help explain the emergence of these high levels of violence, focusing on the effects of the waves of migration to the city from the rural areas that began in the 1950s. The chapter also delineates the involvement of different perpetrators of violence, focusing on their characteristics, methods and motivations. This chapter shows the complexity of the situation and that no single event, actor or situation throughout the 1980s can be considered in isolation when analysing the emergence of the high levels of violence in the city.

Chapter Five, the second data chapter provides an overview of the rising levels of violence in Medellín in the 1990s. There is a discussion of the different government policies to reduce the violence and prevent future outbreaks. There is also a discussion on the different non-state armed actors who were active in the city, such as the militia, criminal gangs (referred to as bandas) and right-wing paramilitaries, focusing on how they used violence to obtain and/or maintain control of territories, their motivations and their characteristics.

Chapter Six is the third data chapter and provides an overview of the decline of levels of violence in the city in the 2000s up to November 2015. In this period, there was a reduction from the extremely high levels of homicides in the 1990s to moderate ones so that by 2015 Medellín had ceased to be one of the ‘50 Most Dangerous Cities in the World’. The purpose of this chapter is to provide a discussion on the policy approaches used by the local government to reduce the violence and prevent future outbreaks. There is also a discussion on the non-state armed groups, the militia, the criminal bandas and the paramilitaries, focusing on their use of violence to obtain and/or maintain territorial control, their characteristics and their motivations. The interactions between these groups and with the state are of particular concern.

Chapter Seven analyses the data presented in chapters Four to Six using the analytical frameworks and concepts presented in chapter Two. The aim of this chapter is to explain what factors led to the 90 percent reduction of homicides Medellín experienced from 1991 to November 2015. This chapter achieves this by first providing an analysis of the policy approaches to reduce and prevent the violence. The chapter also analyses the perpetrators of violence focusing on their use of violence, their motivations and if a change in how they use violence can in fact explain the reduction of violence.

8

The final chapter, the conclusion, reviews the major themes of the thesis and events related to the reduction of violence in Medellín, Colombia. The conclusions presented provide insights that are useful for urban violence reduction and prevention policy makers in middle-and-low- income countries particularly Latin American ones, that are struggling to reduce violence in their ever-expanding cities. A key finding of this thesis is the need for Latin American urban violence reduction and prevention policy makers to widen the scope of assumptions they make about perpetrators of violence. These policy makers should take into consideration the processes of non-state armed group formation and maintenance, the dynamics of non-state inter-group relations and the relations between state and non-state groups when designing and implementing urban violence reduction and prevention policies.

9

Chapter Two: Literature Review

Introduction

Urban violence has attracted considerable attention in academic circles, particularly in the last 30 years. During this time, scholars have constructed theoretical frameworks drawing from both the conflict and development literature to help explain the emergence of violence in urban environments as well as the best approaches to reduce that violence and prevent any future outbreaks. The aim of this chapter is to review this literature on urban violence and policies to reduce or prevent it.

The chapter is divided into four thematic sections that address the key issues. The first section discusses the concepts of violence, conflict and urban violence highlighting the differences between them. The second section of this chapter reviews the literature on the different approaches scholars have employed to explain high levels of violence in particular urban environments. The third section provides a discussion on gangs, which have been identified as the main perpetrators of violence in Latin American urban environments. This discussion involves comparing gangs with other non-state armed groups that use violence. Lastly, there is a discussion of the most influential approaches for reducing or preventing urban violence.

Urban Violence

In order to understand levels of violence in urban environments, it is important to address the concepts of violence and conflict. This is because they are concepts which are often used in urban violence literature and because some urban environments can be characterised as having high levels of violence and conflict. First, we will consider the concept of violence. This can come in many forms with most generic definitions of violence presenting it as the use of physical force, causing hurt to others in order to impose one’s wishes (Keane 1996). Broader definitions extend beyond this physical act to include psychological hurt, material deprivation and symbolic disadvantage (Galtung 1985 & 1991; Schroder & Schmidt 2001). Violence ‘touches’ on levels of what are acceptable and unacceptable types of behaviour, with these levels differing from one society to another and changing over time (WHO 2014, 4). Currently, the World Health Organization (WHO) provides the most used definition of violence, as ‘the intentional use of physical force or power, threatened or actual, against oneself, another person, or against a group or community, that either results in or has a high likelihood of resulting in

10 injury, death, psychological harm, maldevelopment or deprivation’ (WHO 2014, 4). This definition encompasses a distinction between the different direct acts of violence, such as self- directed violence (self-abuse and suicide), interpersonal violence (violence between individuals) and collective violence (violence between large groups of individuals).

The second concept which is often used in urban violence literature is conflict. Conflict is a concept closely associated with violence but carrying a slightly different meaning. Conflict is generally defined as a situation where individuals and groups have incongruent interests, which may be contradictory and potentially mutually exclusive (Moser & Rodgers 2012). While conflict and violence both involve power struggles, conflict can be resolved peacefully through negotiation without using physical or mental harm, whereas violence by its very nature does inflict harm. Political scientists and international relations scholars are particularly concerned with the use of violence associated with armed and civil conflict.

In the last 30 years, the number of armed and civil conflicts has declined. The reason for this decline is because since the end of the Cold War fragmented states are no longer supported by their former superpower sponsors. Armed conflicts are now mainly found in rural areas and considered as limited in their scope and scale (Beall et al. 2013). However, there are still armed conflicts with scholars such as Kilcullen (2013) and Muggah (2014a) predicting that urban environments could be the future sites of national insecurity, civil conflict and urban insurgencies, which may ‘pose security threats on a scale hitherto not encountered’ (Norten 2003, 105). This threat is already being noted in Latin America, a region which is the location of the 47 of the 50 most dangerous cities in the world despite the region not being affected by conventional warfare (Muggah & Giannini 2015).

To describe the type of conflict and violence seen in urban environments, development and conflict literature has started using the social category, ‘fragile city’ (see Beall et al. 2011; Muggah 2014a) to describe environments where ‘the social contract binding municipal governments to their citizens has crumbled and anarchy rules’ (Muggah 2014a, 4). This social category emerges in the work of Zartman (1995) and colleagues (Rotberg 2004; Stepputat & Engberg-Pedersen 2008; Kaplan 2009; Stewart & Brown 2010; Ghani & Lockhart 2009) and their studies on collapsed, failing and failed states. In these fragile urban environments, political authorities may share or cede control to warlords, informal militia, criminal gangs or vigilante groups and authorities may deploy counter-violence through state organisations, paramilitary or non-state proxies, resulting in formal systems of municipal authority and

11 service delivery possibly ceasing altogether (Muggah 2014a, 5). While the anxieties over ‘fragile cities’ have some similarities with those associated with failed and fragile states, not all fragile cities are located in fragile states or are geographically confined to fragile and conflict-affected states (Muggah 2014a, 4). Furthermore, in contrast to the levels of violence encountered in Pre-Westphalian Western Europe, which was seen as a crucial era in the state building process, ‘fragile cities’ often undermine this state-building process because the state is sharing with or ceding control to other groups (Tilly 1985; Muggah 2014a).

Academic debate on urban violence is still in its infancy. This has given rise to the view that many of the basic assumptions underpinning urban violence policy and academic assessments are ‘shaky and untested’ (Muggah 2012, 10). One basic problem is that there is no clear definition of urban violence. Definitions and characterisations are contested and often overlap with those of conflict and crime (Moser 2004). The manifestations of urban violence include a range of acts from homicide to other forms of victimisation, such as crime against property, burglary, assaults, rape and organised crime linked to drugs or political activities (Moser & McIlwaine 2004, 7; Muggah 2012, 19-20). Such types of urban violence can occur in different contexts. It can be state-based and occur during situations of war (such as in Baghdad in Iraq or Kabul in Afghanistan) or non-state based (such as gang violence in Guatemala City in Guatemala) or a combination of both such as, large-scale gang violence sometimes drawing in the state to combat it (such as in Brazil) (Reid-Henry & Sending 2014, 429). However, a common theme in the literature on urban violence is a focus on the interconnected forms of insecurity which encompass physical and psychological harm against persons in urban environments.

Another problem with studying urban violence is the difficulty with measuring levels of violence in urban environments. The most widely used indicator for measuring levels of violence in an urban environment is homicide rates, generally defined as the ‘intentional killing of a person’ as, given its severity, it is more likely than other violent crimes to be recorded (World Bank 2011a, 13). However, these statistics are notoriously unreliable due to under- reporting, difficulties in interpretation and lack of data reliability (World Bank 2011a, 13). Given these limitations, there is an increased use of complementary qualitative methodologies, which give a voice to people’s perceptions of violence, by looking at levels of insecurity, fear and victimisation (Moser 2004, 7; World Bank 2011a, 13). Another limitation is encountered with the quantitative attempts at measuring the effects of urban violence. This is because the direct costs of urban violence are normally measured through economic indicators, such as the

12 cost of the impact of the crimes on individuals and society. While many of the indirect costs of urban violence are intangible, they still have dramatic effects on people’s wellbeing in terms of their livelihood, eroding both financial and social capital (Moser 2004, 8- 9).

This thesis will use homicide rates, when examining the changing levels of violence in Medellín. However, given the limitations encountered with measuring violence in homicide rates this thesis will rely heavily on qualitative data to provide a deep and rich understanding of how citizens perceive violence and the effect this violence has on their daily lives. The use of this qualitative data will make an important contribution to the academic literature and policy-focused analysis on urban violence. The following section reviews the literature on the different approaches scholars have employed to explain the emergence of high levels of violence in urban environments.

Emergence of Violence in Urban Environments

Given the recent entrance of urban violence to the academic debate, frameworks used to explain civil conflict are being employed to explain the emergence of violence and conflict in urban environments. Currently, a large amount of the literature uses frameworks involving the relationships between state, poverty and fragility (Muggah 2012, 8). This is illustrated with the 2011 World Bank World Development Report (World Bank 2011b) considering urban violence only in relation to national state institutions and in contexts of civil war. Within the development and conflict literature relating to urban violence, three theories appear more than others. These are the grievance perspective, greed theory and structural violence. On the one hand, the grievance perspective and greed hypothesis derive from examination of the significance of economic versus socio-political drivers of civil war (Berdal & Malone 2000; Collier & Hoeffler 2012). On the other hand, structural violence looks at the difference between potential and actual life experiences, referring to non-behavioural or indirect violence (Galtung 1969). While these theories were not specifically constructed for application to exclusively urban environments, they are nonetheless currently the most used frameworks.

The grievance perspective and greed hypothesis

The grievance perspective has been used for over 70 years and previously dominated the discourse on conflict. It had its origins in relative deprivation theory, assuming that if an individual is prevented from achieving a certain goal, it is followed by a tension which can only be released by aggression towards the cause of the frustration or another target. In

13 explaining civil conflicts, grievance perspective advocates argue that when groups or regions are explicitly or implicitly neglected or marginalised by the state, they may take up arms to redress this neglect (Sambanis 2002, 223; Gurr 1970). The response of the state, to send in the military to eliminate the armed opposition, can lead to increased levels of grievance through extra-judicial or corruption. Feelings of grievance grow and levels of violence escalate.

Grievance theory was challenged by Paul Collier and colleagues in the 1990s with the greed perspective (Collier & Hoeffler 1998; Collier 2000; Collier & Hoeffler 2004). The greed perspective has its roots in economic theory and assumes that individuals make logical and rational decisions that are in their own self-interest and which provide them with the greatest benefits. In explaining why civil wars start and continue the greed hypothesis argues that individuals find they can profit by gaining control of economically valuable resources (such as diamonds or coca) and take advantage of the associated breakdown of law and order (Collier & Hoeffler 1998; Collier 2000; Collier & Hoeffler 2004). Given the rational nature of this approach, it has been particularly appealing to policy makers, having a direct impact on initiatives in conflict zones by international agencies, such as the UN and the World Bank (Berdal 2005). However, the ‘probabilistic nature’ of this approach has been extensively challenged (Ballentine & Sherman 2003, 5; Ballentine & Nitzchke 2005; Jacoby 2007). For example, Ballentine and Sherman (2003, 5) argued that economic calculations only became relevant when they were ‘intertwined with factors like government mismanagement, exclusionary and repressive political systems, weak states, security dilemmas and inter-ethnic disputes.’ More recently, Ballentine and Nitzchke (2005) criticised the approach for not taking into account the roles of politics, identity and ideology in intra-state wars.

The major shortcoming of the ‘greed versus grievance’ debate also concerns the lack of clarity over the meaning of these terms. This is because while they commonly designate economic and political motivations, factors like inequality could potentially fall within both of these areas (Ballentine & Sherman 2003). This shortcoming was acknowledged by Collier and Sambanis (2005) when they conceded there was a need to move beyond the ‘greed versus grievance’ debate in favour of a more complex model that considers greed and grievance as ‘inextricably fused motives’ for civil conflict. These ‘inextricably fused motives’ were exemplified by Bellows and Miguel (2006) in their study on the conflict in Sierra Leone. This study emphasised that while profits from diamond smuggling helped to encourage armed forces to extend the conflict, the failure of the state to provide economic growth, employment and public services were forms of grievances, which also contributed to the conflict. Therefore, from the

14 debate on ‘greed versus grievance’ the newly emerging model emphasises the need to examine the particular conditions within individual developing societies. This model moves away from exclusively focusing on individuals and rational choice and takes into consideration the role of the state and economic and political differences between groups, or horizontal inequalities (Steward & Brown 2007).

There is not a single approach in the literature to explaining increasing levels of violence in urban environments. While the grievance perspective and greed hypothesis were not specifically constructed for application in urban environments, they do provide useful tools for analysing the emergence of high levels of violence in an urban environment, as they allow the diversity of the manifestations and characterisations of urban violence to be captured.

Structural violence

The second major theoretical influence on explanations of urban violence derives from the concept of structural violence, originally introduced by Johan Galtung in 1969. This type of violence distinguishes between personal and visible acts of violence (direct violence) and potential and actual life experiences (indirect or structural violence). Galtung (1969) argued that structural violence is embedded in the wider social structures of society and, therefore, limits or proscribes an individual’s life experiences and can lead to conflict and direct violence. This approach is important as it extends the notion of violence beyond situations of physical violence to include more implicit forms such as exploitation, exclusion, inequality and injustice. Jacoby (2007, 29) illustrated Galtung’s (1969) approach with the example of early death from tuberculosis. At the beginning of the nineteenth century, this would not be classed as structural violence as treatment did not exist. However, if this happened in recent times, it would be described as structural violence because of the inequitable distribution of the world resources that prevented treatment. More recently Galtung (1990, 291) added the concept of cultural violence which encompassed religion, ideology, art, empirical science as any aspect of culture that can be used to legitimise violence in its direct or structural form.

Since the concept of structural violence was originally proposed by Galtung, it has been heavily criticised. For example, Röling (1970, 206) argued that it did not take into account the different potential realisations of human beings on normative grounds and that ‘not every human potential deserved to be realised.’ Eide (1971, 7) claimed that this approach could be too ‘subjective’, meaning that ‘violence could simply be the cause of what the user of the term does not like.’ Boulding (1977, 83) criticised Galtung for focusing on a ‘concentration on 15 redistribution over production, an obsession with equality and support for liberationism’ and that structural violence is ‘too broad to be analytically useful.’ More recently, Becker (1996) stressed how the existence and maintenance of every social structure depends on some form of ‘structural inequality’ and this inequality does not always lead to conflict or direct violence. Galtung (1975, 371) did, however, assert the need for the concept to be broad to ‘ensure a social net is constructed whereby birds of many feathers can be caught and kept.’ The approach while broad does allow an array of social and institutional failings to be captured and the need to work with these failings for any reduction in conflict and violence.

The concept of structural violence was not constructed exclusively for examining the emergence of high levels of violence in urban environments. However, using this perspective extends the understandings of violence from direct ones, such as homicide, to include indirect ones, such as exploitation, exclusion, inequality and injustice. This perspective also allows some analysis treatment of the diversity of the manifestations and characterisations of urban violence to be captured.

Different Causes or Circumstances to Explain the Emergence of Urban Violence

There is not a single approach to explaining civil conflict and no one theory can explain all civil conflicts. In the twenty-first century, while there are still many international armed conflicts, the rising levels of violence in urban environments have become of increasing concern to policy makers. These rising levels have led scholars from different fields including criminologists, social scientists, urban geographers, demographers and epidemiologists to explore what leads a city to spiral into perpetual crisis (Muggah 2015). Within this scholarly discussion on explaining the emergence of urban violence, there is no consensus and currently no single event or situation can explain the emergence of high levels of violence in particular urban environments. However, there is growing agreement that the different causes or circumstances which explain the emergence of this violence are multi-layered and inter-related. The following section will identify and examine the specific circumstances and causes that have been identified in the literature.

Population increase and density

The first specific circumstance in the literature to explain rising levels of violence in urban environments is population increase and density. The World Bank had previously noted this

16 relationship through economic modelling of crime rates that suggested, under certain conditions, rapid urbanisation rates are associated with higher homicide rates than urban environments which do not experience rapid urbanisation (Fajnzylber et al. 1998, 32). As an example, Karachi in Pakistan experienced rapid growth from the 1950s and is currently one of the most violent cities in the world in terms of its level of homicides, with only 8 percent of the national population but with 42 percent of all reported homicides between 1988 and 2010. Dhaka in Bangladesh, Kinshasa in the Democratic Republic of Congo and Lagos in Nigeria also followed a similar pattern (Shapiro et al. 2012; Gazdar & Mallah 2013). It is widely believed that the lack of ability and preparation of national and subnational governments to absorb rapid urban population growth is a contributory factor to a rise in organised forms of urban violence, as has been noted for Colombian and Brazilian cities (World Bank 2011a, 46; Muggah 2012, 2).

While increase in population and its density are important phenomena for explaining the emergence of violence, unregulated urban growth appears to have more of a relationship than the size and population densities of cities. Thus, Seoul in South Korea, Shanghai in China and Tokyo in Japan are among the largest cities in the world but have maintained high levels of citizen safety even when growing rapidly (Muggah 2015). Previously, urban density was considered a determining factor that shapes a city’s vulnerability to different forms of urban violence. This idea was deeply rooted, extending back to Karl Marx, Max Weber and Emile Durkheim (Muggah 2012, 43). More recently, in the 1930s, the sociologist Louis Wirth in the pioneering work ‘Urbanism as a way of life’ looked at how cities produced their own modes of behaviour and described urban life as one characterised by ‘personal disorganisation, mental breakdown, suicide, delinquency, crime, corruption and disorder’ (Wirth 1938, 23). High rates of population density in urban areas can also lead to competition for limited resources, greater stresses and increased conflicts and is therefore ‘thought to be associated with crime’ (Naudé et al. 2006, 73). This finding of high urban population density leading to the emergence of violence has more recently been confirmed in 2007, when UN-Habitat observed how the prevalence of overcrowding in inadequate dwellings has been linked to increases in negative social behaviours, such as domestic violence and child abuse. However, there are also conflicting assessments of the causal association of ‘densely packed slums’ and the increased incidence of urban violence. For instance, Rodgers (2010, 1) observed how ‘while large numbers, density and heterogeneity can be considered universal features of cities, it is much less obvious that they lead to urban violence.’ More recently, the United Nations Office on

17

Drugs and Crime (UNODC) argued that while population density is a factor in influencing the incidence of homicide and crime, and population density shows a general correlation with homicide rates, other factors, such as the ones discussed below, may nonetheless result in unexpectedly high homicide rates in less densely populated areas (UNDOC 2012). This was backed with research by urban geographers, Buhaug and Urdal (2013), who concluded there was little support for the notion that high and increasing population pressure leads to higher risk or frequency of social disorder.

Weak or ineffective state institutions

A second specific circumstance indicated in the literature to explain the emergence of violence in urban environments is weak or ineffective state institutions. Governance is particularly important in urban environments, as it can promote interaction between local residents and public institutions. If this local governance is characterised as good governance, then it can provide a more capable, accountable, effective and responsive public sector including control of violence (Muggah 2012, 65). However, if state institutions such as police, judiciary, penal system and other forms of legal enforcement are unable or unwilling to regulate and ultimately are unable to effectively claim the legitimate use of force, then violence may thrive in urban environments (Muggah 2012, 49). If these state institutions are weak, they can be considered ‘illegitimate’ by citizens and other groups may emerge to fill the ‘institutional power vacuum’ often resulting in state institutions sharing control with these informal institutions (Winton 2004, 169; Muggah 2012, 49; Muggah 2014a, 4). The capacity of state institutions has been an important consideration in the greed and grievance approaches to urban violence as it emphasises the failure of the state to address symptoms related to greed or local grievances, and the potential for violence can increase as spaces become available for non-state armed groups (Muggah 2012, 49). The lack of state capacity is noted particularly in some newly democratic states which have not been able to successfully manage to reform the police and judiciary or where there has been no systematic dismantling of past institutional structures of terror and oppression (Pereira 2001; Rodgers 2003). For instance, the World Bank (2011a, 72) Violence in the City Report from five cities in Kenya, South Africa, Haiti, Brazil and Timor- Leste, found that citizens in all countries saw the police as possessing one or a combination of the following characteristics: repressive, inefficient, corrupt and biased.

The inability of the state to provide for the legitimate control of violence in urban environments can also lead to the ‘normalisation of violence’ and the ‘cultural construction of violence.’ This

18 violence occurs when symbols, laws, actors, values and attitudes (all sanctioned and codified by the state) either permit or condone the use of violence to resolve tensions and grievances (Muggah 2012, 49). One example of this is in South Africa when during the apartheid system, violence became ‘normative rather than deviant’ with the police playing a fundamental role in enforcing the repressive policies, often relying on ‘draconian security legislation as well as illegal tactics of torture and assassination’ (Altbeker 2009, 260). This use of sanctioned state violence resulted in an entrenched culture of violence, with everyday conflicts in households and schools often resolved through violent means (Simpson 1993; Cock 2000).

Inequality and poverty

A third specific circumstance in explaining levels of urban violence is levels of inequality and poverty. This is because as some cities have experienced rapid growth, levels of urban poverty and inequality have also risen (Baker 2004; Baker 2008). The relationship between poverty or inequality and urban violence has been an important academic debate with poverty long considered a ‘key determinant’ of urban violence (Neumayer 2005). Inequalities can be linked to vertical inequality structures, such as income inequality, with statistical modelling showing ways in which this inequality is one of the ‘most important variables’ in explaining increasing levels of violence (Fajnzylber et al. 2000; Graham & Chaparro 2011). While at a national level, these inequalities are thought to promote criminal violence, there is currently no study confirming this hypothesis in urban settings (see Faijnzylber et al. 2002; Graham & Chaparro 2011). Furthermore, Baumer and Wolff (2014) stressed that while homicides are generally correlated with levels of income inequality across countries, this relationship is much less obvious when examining trends over time.

Economic and social differences between groups, such as inequality in access to basic social services or horizontal inequalities, have also been identified in the literature as leading to rising levels of urban violence. For example, Moser and Holland’s (1997) study on urban poverty and violence in Jamaica revealed that a lack of access to employment was a significant trigger for violence. Other studies have argued how the unequal distribution of resources in urban contexts has a strong impact on violence levels (see Fajnzylber et al. 2000; Vanderschueren 1996; Moser & McIlwaine 2004). Another inequality concerns the access to justice and security services. For example, from studies in Nairobi in Kenya, Gimode (2001) found that the crimes of the rich, such as corruption and economic criminalisation, are far less visible than those of the poor, and the wealthy are far more able to manipulate or ‘buy’ justice whereas those stealing

19 to guarantee their day’s meal can be subjected to lynching. In some urban environments, wealthier neighbourhoods are characterised by state-of-the-art technology and private guards, whereas poorer neighbourhoods are often overtaken by informal militias or vigilante groups (World Bank 2011a, 75). This is noted with Rodgers’ (2004) studies in Managua in Nicaragua, where he found elites often use private security services and live in privately protected spaces whereas the poor neighbourhoods do not have access to these private security services and are often denied entry to these protected spaces, such as public shopping malls. Winton (2004) emphasised how these inequalities are comprised of an important form of ‘structural’ violence, as they may limit or proscribe an individual’s life experiences, leading to conflict and ultimately violence, thereby taking into consideration Galtung’s (1969) structural violence.

The discussion on the different frameworks and causes or circumstances to explain the emergence of violence in an urban environment has emphasised the need to take a multifaceted approach drawing from the conflict and developmental literature. Some of the causes or circumstances focus on unregulated urban growth and urban density, others on ineffective or weak state institutions as causal factors for urban violence while others find explanations in horizontal and vertical structures of inequality. Throughout the academic literature which focuses on explaining the emergence of violence in Medellín these multifaceted approaches have also been used. For example, Restrepo & Velez (1998) draw from the greed hypothesis by arguing that the emergence of violence in the city is largely linked to the growth of narco- trafficking. The Medellín-based research, Giraldo (2008) argues the emergence of violence is related to the weak or lack of state institutions in the poorer neighbourhoods. Another Medellín- based researcher, Jaramillo (2011), places importance on the severe inequalities in the city in explaining the emergence of violence. More recently, in 2016, the Medellín-based researcher, Davila (2016) concluded that the explanations for the rise of violence in Medellín not only come from multiple disciplines but also that there are various explanations for it. This thesis also acknowledges the need to take a multifaceted approach in explaining the emergence of violence in Medellín. It is only by drawing from these multi-layered and inter-related approaches that a holistic explanation of the emergence of violence in Medellín can be provided.

20

Perpetrators of Violence in Urban Environments

This section of the literature review discusses the main perpetrators of violence in Latin American urban environments which have been identified in the literature as gangs. This discussion involves comparing gangs with other non-state armed groups that use violence.

While gangs have featured as a social phenomenon for generations, their regional growth and transnational influence in the last 20 years is ‘without precedent’ (Muggah 2012, 47). Given this regional growth and transnational influence, some scholars have warned of an impending ‘gang insurgency’ in urban environments because of the threat these groups pose to both security and socio-economic development (Manwaring 2005; Sullivan & Bunker 2011). Thus, a recent development in the literature is that gangs should be included in the non-state armed group framework because of the threat to the rule of law these groups pose (Rodgers & Muggah 2009, 2). Gangs have particular characteristics. Some of these include their street-oriented lifestyle, being comprised of youths who normally engage in some type of illegal activity and the illegal activities often help to define the identity of the group (Klein & Maxson 2006). However, practical and theoretical understandings of these actors are limited because of the scarcity of reliable information on them. This is due to under-reporting, low levels of data collection and political interference, thereby limiting both the practical and theoretical understandings of these actors (Cockayne & Lupel 2009; Muggah 2012; De Boer & Bosetti 2015).

This is no unified definition of a non-state armed group. There is, however, a consensus in the conflict and insurgency literature that they generally have some sort of structure, membership, ‘use arms to achieve their objectives and are not under state control’ (International Council on Human Rights Policy 1999, 5). They can take a variety of forms, varying in their purpose, composition, membership, organisation, longevity, activities and use of small arms. They can have a military and political line of command, such as the Revolutionary Armed Forces of Colombia (Fuerzas Armadas Revolucionarios de Colombia - FARC), or can be spontaneously gathered groups engaging in combat or sporadic acts of collective violence without a command structure or autonomous individuals acting alongside other groups (Bassiouni 2008, 717). Given the diversity in the variety of non-state armed groups, they encompass a wide range spanning militias, rebel groups, insurgents, terrorists and criminal organisations.

21

While the label that is given to a non-state armed group by the government and media is important to obtain an understanding of their manifestations or motivations, often the label that is used to describe a certain group, does not always assist in obtaining these. For example, a label can indicate the motivation of a group, with politically motivated non-state armed groups, having ‘positive labels,’ such as revolutionaries, liberation movements, and militias. The use of these positive labels can sometimes be attempts to indicate the group has some legitimacy in taking up arms, such as the Liberation Tigers of Tamil Eelam in Sri Lanka (Hazen 2010a, 374). Alternatively, ‘negative labels’ such as terrorists, rebels, insurgents, criminals, gangs and warlords can depict the illegitimacy and illegality of the group (Hazen 2010a, 374). Criminal non-state armed groups often do not have a label which indicates their motivation, such as the Sinaloa Cartel or the Sicilian Cosa Nostra. Some non-state armed groups can also be labelled differently by external actors or over time, resulting in confusion as to whether the group itself has changed or if external parties have changed the label of the group to possibly influence public policies (Hazen 2010b, 258). One example of non-state armed groups being labelled differently is found in the Philippines when the Moro Islamic Liberation Front (the MILF) were linked to ‘terrorist attacks’ and accused of harbouring terrorists from South-East Asia by the government, despite the group denying such labels (Bhatia 2005). Finally, some non-state armed groups often combine political and criminal activities. As an example, the United Self- Defence Forces of Colombia (Autodefensas Unidad de Colombia - AUC) began as a right-wing paramilitary group opposing left-wing groups, but, they have increasingly become involved in criminal activities (World Bank 2011a, 25; De Boer & Bosetti 2015, 13).

Given the difficulties of labelling non-state armed groups, scholars have argued it is more important to analyse the particular characteristics of a non-state armed group, rather than classify them into these rigid, actor-based groupings. For example, Vinci (2006) has suggested the use of the ‘the lowest common denominator’ of how groups choose to mobilise, based on three key elements: membership (e.g. recruitment), logistics (e.g. weapons and food) and direction (e.g. command, control and communication). By contrast, Williams (2008) categorised groups based on how they fare across nine dimensions: motivation, purpose, strength, scope, funding, organisational structure, role of violence, relationship to the state and the function they play in society. Hazen (2010b) positioned non-state armed groups along a continuum that captures the respective group’s relationship to the government, its level of organisation and its capacity for wide-scale violence. In this continuum, Hazen (2010b) argued there are groups which neither support nor oppose the government, therefore challenging any

22 previous divisions of non-state armed groups acting within either the greed or grievance framework. Hazen called these groups ‘neutral groups’ and while they do not seek to overthrow the government, they do pose a challenge to law and order as they can infiltrate the government and influence policy through corrupt government officials (Hazen 2010b, 259). More recently, organisation theory has also been used to analyse the characteristics of a non-state armed group, with Ugarte and Turner (2011) asking whether the Philippines’ terrorist or bandit group, the Abu Sayyaf, demonstrated the characteristics of a formal organisation. By analysing the Abu Sayyaf’s goals and co-ordination between members, they concluded that groups which were labelled by the government and the media as the Abu Sayyaf did not display these characteristics. Flexible networks often based on kinship were the defining features of the phenomenon referred to as the Abu Sayyaf.

There are also limitations with the use of the label ‘gangs.’ For example, within the academic discussion on gangs, some scholars emphasise the need to distinguish between groups of individuals committing less well-organised and well-planned crimes without any clearly defined purpose, or street gangs and gangs linked to organised crime (Klein 2005; Hauck & Peterke 2010). The latter often have more of a formal organisation structure, with the leader operating the gang like a business. Thus, they normally have a market orientation, operate over a large geographic area and have a centralised leadership (Sullivan & Bunker 2007). However, often the line between supposedly different types of gangs are often not so clear, with Hauck and Peterke (2010, 413) arguing the separation between street gangs and gangs linked to organised crime is ‘artificial and of little use.’

One example of the limitations of using the labels ‘gangs’ or ‘gangs linked to organised crime’ has been noted in the Northern Triangle (Guatemala, El Salvador and Honduras). In 2009, Rodgers and Muggah (2009) found that clikas (local cells) are generally affiliated with the larger gangs, such as the Mara Dieciocho or the Mara Salvatrucha, with clikas from different neighbourhoods often joining together to fight other groups claiming allegiance to the opposing larger gang or mara. They also found that neither the Mara Dieciocho nor the Mara Salvatrucha answer to a single chain of command, with their umbrella nature more symbolic of historical origin than of any real organisational unity, be it of leadership or action. Furthermore, it was previously understood that maras played a small, if any, role in transnational cocaine trafficking. However, in 2015, Rodgers and Baird (2015, 5) claimed there was ‘clear evidence the maras have become more and more involved in drug trafficking.’ More recent research argued ‘the Mara Salvatrucha do not have a direct relationship with the

23

Mexican drug cartels, although many would like to’ (Farah 2016, 1). This example highlights the difficulties of using labels such as ‘gangs’ or ‘gangs linked to organised crime.’

Not only is there a lack of consensus over how to differentiate between street gangs and gangs linked to organised crime but also organised crime is a concept contested throughout the literature. This is noted with the organised crime expert Klaus Von Lampe (2015) listing no less than 180 definitions of the concept. Generally, organised crime refers to sophisticated criminal activities embedded in one form or another in complex illicit markets, which can range from trade in arms, contraband cigarettes, stolen cars, drugs and human trafficking to the extortion of multinational corporations, gambling and prostitution (Fijnaut 2000; Lyman & Potter 2007). Abadinsky (2010) has proposed a useful list of common attributes often found in organised crime entities as groups having a lack of political goals, limited or exclusive membership, constitution as a unique subculture, self-perpetuation, willingness to use illegal violence and bribery, demonstrated specialisation/division of labour, monopolism, strict rules and regulations, and a hierarchical structure. Other studies have challenged this definition that includes a ‘hierarchical structure’ and have instead pointed out how some groups are disorganised and are better characterised as small and loose networks (Williams 2001, 65; UNODC 2002, 30).

The most used definition not of organised crime but of an organised criminal group comes from the United Nations Convention against Transnational (2000). The Convention defines such a group, as a ‘structured group of three or more persons, existing for a period of time and acting in concert with the aim of committing one of more serious crimes or offences… in order to obtain, directly or indirectly, a financial or other material benefit.’ The transnational dimension is then applied when operations extend beyond one nation’s borders. This definition, while vague, recognises that not all groups follow hierarchal structures and also focuses on ‘enabling activities’ that are often characteristics of organised crime groups. These activities are generally seen as violence and/or the threat of it, corruption and money laundering with the ‘core activities’ trafficking in arms, drugs and humans (Mair 2005; Van Dijk et al. 2007). Another issue with this definition is that some organised crime groups are comprised of hundreds if not thousands of members and affiliates. For example, the Russian-based organised crime group, the Solntsevskaya Bratva, involved in the trafficking of heroin and human trafficking, is estimated to have over 9000 members (Fortune 2014).

24

The activities of organised crime groups have recently challenged insurgency as the main threat to the state’s monopoly of violence (Kalyvas 2015, 1522 & 1534). This challenge is particularly noted in Latin America as one of their enabling activities of these groups, their use of violence, had been noted by the UNODC in 2007 as the explanation for the rising homicide rates in the region. More recently, in 2014, the UNODC estimated that 30 percent of all homicides in the area were linked to these groups using violence to obtain and/or maintain territorial control. This new type of insurgency has led civil war and development scholars to integrate organised crime into their analyses (see Duffield 2001; Miraglia et al. 2012; Kaldor 2013; Locke 2012; Keen 2012; Kalyvas 2015). This scholarly work provides useful tools to further understand not only how the perpetrators use violence but also the type of challenge they pose to the state. Thus, the following section provides a discussion of how gangs use violence and their relationship with the state.

Use of violence

Gangs can be classified under the broad concept of non-state armed groups. However, the main difference between these actors and other non-state armed groups appears to be their contrasting use of violence. On the one hand, non-state armed groups typically use systematic and purposeful violence in their long-term goal of overthrowing the government, seceding, or in some other way changing the governing structure of the state (Hazen 2010a, 381). In some cases, non-state armed groups have control over territories or significantly influence activities within a territory and the population living there as with the MILF in the Philippines and the Liberation Tigers of Tamil Eelam in Sri Lanka (Geneva Centre for Democratic Control of Armed Forces 2015, 9). In other cases, non-state armed groups move through areas and have no attachment to a territory and often target innocent civilians. For example, the Lord’s Resistance Army which has operated in Uganda, the Democratic Republic of the Congo, the Central African Republic and South Sudan used extreme violence, such as rape, sexual slavery, killing and maiming (United Nations 2014). On the other hand, gangs appear to be motivated by their own economic interests, such as profits obtained from their involvement in organised crime. Gangs also tend to be territorially based and often protect their ‘turf’ and the people living there (Rodgers 2002; Hagedorn 2007; Rodgers & Jensen 2008; Hazen 2010a). Thus, they use violence, often homicidal violence, to obtain and/or maintain control over territories to advance or maintain their illicit business opportunities. This is in contrast to non-state armed groups who obtain control of territories in their efforts to overthrow the government, secede, or in some other way change the governing structure of the state (Hazen 2010a, 381).

25

To explain levels of violence in territories where gangs operate, Schneider (2013) analysed violence that occurred between the criminal group the Hell’s Angels and its rivals over dominance in Quebec’s cocaine market in Canada in the 1990s and drew the distinction between stable markets and unstable markets. Schneider (2013) argued how on the one hand, stable markets tend to have lesser levels of violence as they tend to operate according to entrenched rules. On the other hand, unstable markets tend to be more prone to violence as roles are not clearly identified and lines of authority and hierarchy are not clearly established.

Recent research that has analysed how gangs use violence to obtain and/or maintain control over territory has suggested that gangs members may prefer using corruption and intimidation, rather than direct violence, such as homicidal violence (De Boer & Bosetti 2015; Cockayne 2013). The reason for using violence in this manner to obtain and/or maintain territorial control is so they can refrain from drawing the attention of the authorities. In taking a rational approach to how they use violence, members can concentrate on their long-term goals of maximising their profits from their criminal activities. This rationality may not always be evident in practice as it is tempered by the high degrees of uncertainty involved in criminal activities, judgement error, and conflicts between the short- and long-term interests of these criminal actors (Brezina et al. 2002). Challenges to the state and ungoverned spaces

There are similarities between non-state armed groups and gangs in how they operate in ungoverned spaces. The term ‘ungoverned spaces’ was first used by the former US Secretary of State, George Shultz, to capture the lack of effective state presence in particular countries or cities. The US government was concerned that ungoverned spaces would provide a ‘safe haven’ to terrorist groups, such as Al Qaeda and other armed groups, which could capitalise on the lack of government presence (Lamb 2008). The term, ungoverned spaces, refers more to a lack of ‘effective state governance’ rather than a lack of governance in total (Hazen 2010a, 379). This lack of effective state governance can include states or areas characterised by poorly controlled borders or airspace, a lack of government authority beyond the capital, little or no provision of security by the state, limited protection of citizens from violence, weak political systems, deficiency in the rule of law and the inability to establish a legitimate monopoly of violence (Rabasa et al. 2007; Tedesco 2007). In these geographical areas, informal structures usually exist to fill any ‘vacuum’ left by the government (François & Sud 2006). These

26 structures will differ from each context in terms of the type of control they impose, their use of violence and the extent to which they provide social services, depending on the goals of the armed group and what they may need to achieve these goals. Research on gangs shows how they operate in urban ungoverned spaces. For example, Koonings and Kruijt (1999) stressed how state withdrawal from poor communities in Latin American cities left ‘governance voids’ which gangs ultimately filled (O’Donnell 1993). In these areas, community members may turn to these actors to provide certain services, such as employment and security, which are not made available by the state. For example, from studies in urban Nicaragua, Rodgers (2002) demonstrated how gangs consolidated political, social and economic authority in particular neighbourhoods. In situations of entrenched institutional and social marginalisation, the gang can become a key institution, part of how residents endow their world with form and meaning (Rodgers & Jones 2009). The under-governed area contributes to the creation of the gangs and even to a perceived reliance on them, although the community can sometimes fear rather than support these actors. The role of these gangs may become normalised because of a perceived inability to change the situation and community acceptance of the authority of the gang (Hazen 2010b, 381-383). Therefore, even though gangs lack a political agenda, their activities have clear political implications as they can establish a separate, localised order, possibly evolving into institutions that resemble those of governments (Johnson 2013). This normalisation can have a detrimental effect on the prospects for the return of the government to these areas as it may diminish the legitimacy of government with Rodgers & Baird (2015, 13) describing this phenomenon as a type of ‘localised form of sovereign power.’ Therefore, even though gangs do not seek to overthrow the state, similar to non-state armed groups, they do pose a challenge to the state.

An example of gangs operating in ungoverned spaces and posing a challenge to the state has been noted in Rio de Janeiro, Brazil. Gangs in this urban environment have been linked to international organised crime groups since at least 1982 (Gay 2010, 33). These actors have provided not only ‘law and order’ and ‘protection’ but also ‘social welfare services’ for slum residents, such as financing child day-care facilities, giving handouts for cooking gas, organising parties and community social events, acts perceived by the community as a ‘public good and act of benevolence’ (Perlman 2009, 63; Gay 2010, 33; Wolff 2015, 31). These gangs took advantage of the ‘widespread distrust of the police’ and the abandonment of the favelas (shanty towns or slums) by state welfare services and became part of the community, rather than separate from it (Leeds 1996, 59-61). Wolff (2015) argued that these gangs were among

27 the actors most capable of providing conflict resolution services for residents, allocating scarce resources and enforcing contracts, as the state did not control these areas. Wolff (2015) also argued that these actors have a rational incentive to provide such services, as they create a sense of loyalty, which decreases the need to use force to maintain territorial control. This territorial control is needed for their enabling and core activities linked to organised crime and on which their livelihoods depend. These conclusions resemble Kalyvas’s (2006) seminal work ‘Theory of Irregular Warfare’ in which he argued that criminal groups will invest in improving their relationship with their host communities in order to reduce the cost of control, which by force alone may be exorbitant. Kalyvas (2006) called this the ‘logic of territorial control.’

Another issue encountered with ‘ungoverned spaces’ is that ‘ungoverned’ can often legitimate repressive approaches by the state, such as militarised ones, to return governance to the areas. For example, in Rio de Janeiro, Brazil, in 2007, over a thousand military police invaded a selection of the poorer favelas to expel the local gang, the Comando Vermelho (Gay 2009). Similar state responses to ‘ungoverned spaces’ have also been used in the Northern Triangle with the repressive anti-gang ‘mano dura’ policy approaches (Rodgers 2009). These repressive state approaches however do not always provide the state with a return of good governance to the area and can instead contribute to greater insecurity. For example, in the Northern Triangle, there have been reports of ‘tit-for-tat violence’ by gangs in response to the repressive approaches by the state. This reaction was noted particularly in August 2003 when gang members attacked a bus in the Honduran city of San Pedro Sula, killing fourteen and wounding eighteen and left a note for the Honduran president ordering him to withdraw the new repressive policy. Over the following 12 months, more than ten decapitated corpses were left in various Honduran cities with similar messages for the President (Rodgers 2009).

The discussion has employed elements of the civil wars and insurgency literature to compare gangs with other non-state armed groups that use violence. In Medellín, the residents have experienced various types of non-state armed groups including militia, paramilitaries and gangs linked to organised crime. The comparison of gangs to non-state armed groups provides tools to analyse not only the nature of these entities and their use of violence in Medellín but also the challenge they pose to the state.

Approaches to Reduce and Prevent Violence in Urban Environments

28

This final section reviews the current theoretical and policy approaches to address levels of urban violence. There is a substantial literature on responses to violence in urban environments and there is a wide range of approaches designed to contain, prevent and reduce this violence (Muggah 2012, 53). Currently, there is no simple transferable intervention policy with policy makers normally adopting a combination of policies rather than a single one. Figure 3 shows how these interventions normally take a soft (prevention) or hard (reduction) approach.

Figure 3: Targeting Urban Violence Prevention and Reduction Interventions

Image Removed due to Copyright

Source: Muggah 2012, 54

Figure 3 shows how either ‘formal’ public institutions or ‘informal’ civil society actors can undertake interventions, although, many can involve collaborations across the two types of actor (Muggah 2012, 55). One the one hand, soft approaches are often violent preventative and normally focus on social and economic factors such as urban renewal for security or promoting social capital and cohesion. On the other hand, hard approaches are often about violence reduction and focus on controlling the violence, such as through policing interventions. However, one shared issue of the violence reduction and prevention approaches is that the evaluation of their effectiveness is still not clear. For example, the 2011 World Bank Violence in the City report commented that urban violence prevention ‘is an area that the development community still have little knowledge of the types of programmes that can be effective’ (World Bank 2011a, 86).

In measuring the success of a particular approach, policy makers often rely on the reduction of homicide rates. This is because these indicators are the most widely used measure for measuring and assessing levels of violence in urban environments and thus become the prime indicator for charting the reduction of violence (World Bank 2011a). However, by only relying on this objective indicator, policy makers may fail to capture any changes in subjective indicators, such as perceptions of safety and security (McCord 2003; Muggah 2012).

29

The following section provides a discussion of the most influential approaches that currently influence urban violence reduction and prevention policies. These are social disorganisation, broken windows and social capital.

Social disorganisation

The first of the three most influential approaches that deals with urban violence reduction policies is the sociological theory called social disorganisation. This approach, developed from research in North American urban environments, assumes that perpetrators engage in violence because of the disruption of community social organisation which can be linked to the economic disadvantage, ethnic and racial heterogeneity and residential instability (high turnover of residents) in a community (Shaw & McKay 1942). From a policy perspective, the ‘social disorganisation’ can be addressed through slum upgrading, which focuses on modernising essential infrastructure, promoting public spaces, formalising land title and tenure, and improving housing. With an estimated 863 million people living in slums in 2014 with predictions this will rise to 1.4 billion in 2020, this has been a popular approach for urban planners (UN-Habitat 2014; Muggah 2012, 62; UN-Habitat 2011). Some even claim that mega slums constitute the new frontier of armed violence and sources of insecurity (Rapley 2006).

Much slum upgrading has focused on poverty reduction as it can address a range of deprivations, such as improving access to formal sector employment opportunities and improved secure housing, social protection mechanisms and better access to basic services (Baker & Schuler 2004). For instance, poverty reduction initiatives in Dhaka in Bangladesh, Nairobi in Kenya and Port Moresby in Papua New Guinea sought to upgrade essential infrastructure, promote open public spaces, formalise land title, and tenure and improve housing, with many initiatives complemented with targeted interventions to promote income generation through job training, micro-finance schemes and support for local enterprises (Muggah 2012, 62). Some initiatives have, however, sought to engage directly with urban safety. As an example, slum upgrading in the Latin American cities of Rio de Janeiro in Brazil, Caracas in Venezuela and Ciudad Juarez in Mexico, aimed to improve livelihoods while also directly promoting enhanced safety and security. These interventions emphasised a combination of investment, service delivery, physical infrastructure and training, and also sought to bring together the formal and informal areas of the city and promote exchanges between them (Muggah 2012, 63).

30

The outcomes of slum upgrading in reducing or preventing violence are highly contested. For example, in intervention neighbourhoods, while quality of life and safety indices may have improved, the social and economic integration of these settlements into the mainstream of city life has sometimes been less than impressive. In the Favela-Barrio project in Rio de Janeiro in Brazil, one of the core objectives was to reconnect the slums with the city. While violence rates decreased, some communities have allegedly been disempowered by international development agencies reinforcing top-down development practices, rather than engaging with the communities at a grassroots level (Samper 2011). Jütersonke et al. (2009, 168) have even suggested that evidence supporting claims that large slums are ‘places of armed violence or insecurity is thin.’ Rather than being places of disorganised violence and insecurity, these areas may be controlled by gangs who may provide services to the community which they are lacking from the state, such as security services. Therefore, unless these underlying gang structures are addressed, the effects of slum upgrading can be criticised as ‘purely cosmetic’ (Muggah 2012, 63). This criticism was more recently confirmed by research from the US Agency for International Development (USAID) that stressed ‘it is clear that criminal organisations are capable of relocating or otherwise responding to [these types of] targeted interventions’ (USAID 2016, 27). Finally, critics of the social disorganisation approach, see the relationship between neighbourhood disorder and crime to be ‘spurious.’ For example, qualitative and quantitative data collected from neighbourhoods in Chicago in the US by Sampson and Rudenbush (2004) suggested that reducing actual levels of disorder will not change the perceptions of disorder and bias that residents outside these communities have towards particular neighbourhoods.

Broken windows

The second of the three most influential approaches that deal with urban violence and reduction policies is the ‘broken window’ theory of crime. This approach, developed from research in North America, suggests that public disorder can encourage violent behaviour as it signals to would-be perpetrators there is little consequence to crime (Kelling & Coles 1997). Physical deterioration such as broken windows, garbage in the streets, abandoned buildings and a lack of street lighting can create an environment that promotes crime and delinquency and thereby has an impact on citizens’ perceptions of feeling safe (International Centre for Prevention of Crime 2008). The physical disorder can be addressed through policing interventions and improvements to the built environment.

31

Policing interventions involve the re-assertion of state authority through efforts to reinstall social networks in neglected areas. This is because the police can play a role in focusing on disorder and less serious crime and thus help to reduce the fears residents have. The promotion of informal social control will then help residents to take control of the neighbourhood and prevent more serious crime (Centre for Evidence-Based Crime Policy 2016). These types of interventions are frequently supported by the public because of the visibility they bring (Muggah 2012). Famous examples of this approach include the ‘Zero Tolerance’ policies in New York City in the US and the ‘Iron Fist’ (mano dura) policies in San Salvador in El Salvador. However, there are challenges of rebuilding trust between policing institutions and poorer segments of the population who may have been previously targeted by police, and they may view the police as ‘inefficient and irresponsible’ (Winton 2004, 86). There are also further questions of how to manage wider issues of new and emerging forms of criminality and promote improved socio-economic livelihoods in newly ‘cleared areas’ (Muggah & Mulli 2012). Repressive approaches may also initially reduce criminality, but can be ‘temporary and tenuous’ and may only encourage individuals to become more organised and violent (Aguilar & Miranda 2006, 42). One example of this is noted from Jütersonke et al.’s (2009) research on the mano dura policies. They found gang members were ordered by bosses to become less conspicuous and to deploy less obvious signs and symbols to avoid detection because of the repressive state approaches.

The physical disorder can also be addressed through physical repairs to an area. One example is the Safe Streets campaign in various cities in the US which:

Reclaimed neighbourhoods, business districts and schools from extreme street violence, gang recruitment of youth and street-level drug dealing by organising events such as block watches, neighbourhood patrols, youth leadership teams or graffiti removal teams (Safe Streets 2015).

However, improving a built environment does not address the underlying causes of violence and only focuses on reducing the opportunity for potential offenders to commit a crime (Kruger et al. 2001). Improving the built environment may also have ‘limited payoffs’ in neighbourhoods inhabited by large numbers of ethnic minorities and poor people as the perceptions of disorder in a particular environment extend far beyond the presence of run-down infrastructure as was noted by qualitative and quantitative data collected by Sampson and Rudenbush (2004) from neighbourhoods in Chicago.

32

Social capital

The third of the three most influential approaches that deals with urban violence and reduction policies is the bottom-up approach of social capital. This approach focuses on the networks, norms and trust that facilitate both coordination and cooperation for mutual benefit (Putnam 1993). The concept of social capital mainly appeared in the work of French sociologist Pierre Bourdieu in the 1970s and 1980s and was further developed by James Coleman in 1990 and then greatly popularised in the work of Robert Putnam on Italy and the US. Putnam (2001, 19) concluded that ‘spontaneous cooperation’ is facilitated by social capital, or the connections among individuals through social networks and the norms of reciprocity and trustworthiness. Such networks are comprised of relationships in which people are ‘doing with’ others, rather than ‘doing for’ others. Social capital is regarded as a critical factor shaping collective action and fostering working trust with shared expectations for action or collective efficacy. Various studies have previously concluded that social capital is critical for addressing issues of social disorganisation and as such can reduce and prevent violence (Sampson et al. 1997; Sampson et al. 1999; Moser & McIlwaine 2006; World Bank 2011a).

Social capital can be enhanced by the state through participatory citizen planning and exchanges with the public sector, such as participatory budgeting. Informal institutions such as community organisations can also play a role in creating trust in a community and building on the community’s own identification of its needs (World Bank 2011a). For over a decade, major organisations such as the UNDP and World Bank have been redirecting some of their investment and development programming towards promoting social capital and cohesion, including in societies emerging from armed conflict or at risk of experiencing outbreaks of violence (Muggah 2012, 59). More recently, the Rockefeller Foundation launched a US$100 million initiative to support 100 ‘resilient cities’ in confronting the shocks and stresses in the twenty-first century. This initiative focuses on the capacity of individuals, communities, institutions, businesses and systems within a city to survive, adapt and grow, no matter what kinds of chronic stresses and acute shocks they may experience (Resilient Cities 2016).

Social capital does appear to be relevant in urban environments with high levels of violence. This is because in these environments acts of violence may signal to residents that mechanisms of social control are not working, and to would-be perpetrators that opportunities for crime and violence exist (Foster 1995). Urban environments with high levels of violence often also have high levels of trauma. Trauma is understood as the response to an extraordinary event that can

33 overwhelm an individual’s coping resources, which can adversely affect both the individual victims and social relationships and networks (Pouligny 2010). This is because trauma can lead people to change their habits and/or create new social patterns such as victims re-enacting the trauma by perpetrating violence against others. This can then result in the breakdown of trust leading to reduced participation in community life particularly in community collective action, and this works against the establishment or maintenance of collective efficacy (World Bank 2011a, 34). It is, therefore, relevant to look at coping mechanisms in these environments for dealing with high levels of violence, with recent research from World Bank (2011a) indicating that these mechanisms occur mostly on the individual level, such as keeping silent, buying guns or relying on extra-legal security groups for protection. As an example, in Cato Manor in South Africa, Meth (2004) described how local community members adopted highly informal mechanisms of ‘revenge violence’ and vigilante crime to deal with criminals and robbers. These mechanisms, undermined long-term violence prevention by generating a type of perverse social capital (Rubio 1997).

The criticism of the social capital approach is that strong social ties can actually hinder collective action. For example, Wilson (1996) noted from research in the US how strong social ties within a group created a dense network of connections, but that these ties could end up isolating residents from the broader society. Collective efficacy can also result in communities bonding together to undertake actions that may be discriminatory against minority groups, such as with the emergence of vigilante or militia groups which bring people together for the benefit of certain individuals over others in a community (World Bank 2011a, 33). From research in Colombia, Rubio (1997) mentioned ‘perverse social capital’ which is the use of strong ties and networks in communities to facilitate drug trafficking networks that have negative repercussions across those communities and beyond. Dowdney (2003) noted in research from Rio de Janeiro, that emerging norms condoned a drug culture glorifying power, machismo and easy money. The networks and norms can become institutionalised and oppose the values and claims of the state and public authorities. Finally, some urban communities have managed to exert social control over violence even without having strong ties among neighbours (Sampson et al. 1999). For example, from data collected in the US, Bellair (1997) suggested that the type of interaction between neighbours matters more than the frequency of interaction and that less frequent interaction among neighbours, were predictive of lower crime rates.

34

Context-specific approaches

The above approaches provide a useful starting point for designing urban violence reduction and prevention policies. However, a recent report by USAID on What Works in Reducing Community Violence stressed that these approaches should only be implemented with ‘a careful process of consultation and adaptation with the participation of local stakeholders’ (USAID 2016, 32). This acknowledges that the most influential approaches to urban violence reduction and prevention have all been developed from either a North American or Western European perspective. The approaches have also been subjected to considerable testing and analysis in the urban environments of high-income countries such as Brisbane, Australia; Freiberg, Germany; Stockholm, Sweden; London, United Kingdom; and New York City, Indianapolis and St. Louis in the US (Mazerolle et al. 2010; Oberwittler 2004; Veléz 2001; Sampson and Wikström 2008; Kelling & Sousa 2001; Community Cohesion Review Team 2001). Their effectiveness in urban environments of middle- and low-income countries, such as those in Latin American, has not received this same testing and analysis (Muggah 2012). Furthermore, the urban environments in high-income countries have substantively different socioeconomic conditions, institutional structures and manifestations of violence very different to developing countries. The urban environments of middle-and low-income countries are also contexts where there is an assumption of more trusting relations in the broader institutional framework of society and the state.

The need for context-specific approaches confirms previous suggestions from development and conflict scholars such as Moser & McIlwaine (2014), Muggah (2015) and Kalyvas (2015). These scholars argue that urban violence reduction and prevention approaches need to be designed with nuanced understanding of the historical, geographical, spatial and structural complexities of specific urban contexts as well as a more realistic assessment of what can and cannot be done to reduce, better manage and contest violence. The recent civil war and development literature emphasises the need for more pragmatic and innovative approaches, and systemic solutions (Moser & McIlwaine 2014; Kalyvas 2015; Muggah 2015). For example, Moser and McIlwaine (2014, 338-339) argue for pragmatic approaches and while these may be ‘politically challenging and socially provocative’, they should seek to manage rather than eradicate violence. Muggah (2015) stressed the need for innovative solutions such as improving communication within cities, with dialogues between vulnerable communities and local governments, and between cities, by bringing together fragile cities with healthier and

35 wealthier ones. Alternatively, Kalyvas (2015) argued the need for ‘drastic changes’ in the international context, such as levels of drug consumption and their illegality in the US to address the systemic problems of violence in Latin American cities.

Towards an Analytical Framework

As set out in Chapter One, this thesis seeks to explain the reduction of the most serious type of violence, homicides, from appallingly high levels to moderate ones in Medellín. To explain the reduction, this thesis draws from development and conflict literature, focusing on three key themes. The first theme this thesis employs draws from the conflict and development literature on the ‘greed versus grievance’ debate and structural violence to explain the emergence of urban violence. These approaches while normally used to explain societal conflicts, such as civil war or secession, will guide the discussion on the emergence of the high levels of conflict and violence in Medellín. This discussion will also take into account the specific circumstances identified in the literature that assist in explaining the emergence of high levels of violence and conflict in urban environments. These are population increase and density, weak and ineffective state institutions, and inequality and poverty. This thesis will take a multifaceted approach in explaining the emergence of violence in Medellín drawing from these different frameworks and causes or circumstances. It is only by drawing from these multi-layered and inter-related approaches that a holistic explanation of the emergence of violence in Medellín can be provided.

The second theme of this thesis is employing concepts used in conflict and insurgency literature to analyse the different non-state armed groups in Medellín. This analysis will allow some understanding of their use of violence and the challenge these actors pose to the state. This analysis provides understanding of the nature of non-state armed groups in Medellín including their motivations, structures, interactions and how they use violence.

The third theme of this thesis is employing the most influential policy approaches for urban violence reduction and prevention - social disorganisation and broken windows, and social capital - to guide the discussion on the reduction of violence in Medellín. By employing these policy approaches this thesis makes an important contribution to the urban violence literature as there has been limited testing and analysis of these approaches in the urban environments of middle-and low-income countries. The methodology of this thesis, discussed in the following chapter, makes an important contribution to this literature, as the case study approach provides

36 a deep and rich understanding of the complexities of urban violence in a specific urban environment. The reliance on qualitative data also provides important insights into how an array of local citizens perceive levels of violence and the effect this violence has on their daily lives.

37

Chapter Three: Methodology

Introduction The purpose of this chapter is to provide a detailed description of the methodology I have used for data collection for this research. This chapter outlines key components of the methodology and provides an explanation and justification for my particular approach to the research. These components include case study analysis, gathering information from interviews and participant observation during two periods of fieldwork in Medellín, and document and archival analysis. There is also a discussion on the risks involved with researching in areas with high levels of violence and conflict.

Case Study

This research employs a case study approach to understand the contemporary phenomenon of violence and its reduction in Medellín. Yin’s (2003) work on case study research provides the most comprehensive exploration of the types of case study approaches that can be used in research work, as well as detailed analysis of the limitations of each approach. I used Yin’s (2003) contribution to assist and guide my understanding and use of single-case study research with regard to patterns of violence in Medellín. A case study is the most appropriate approach when ‘how’ and ‘why’ questions are posed, when the investigator has little control over events and when the focus is on a contemporary phenomenon within some real-life context (Yin 2003, 7). This fitted the situation in Medellín. Case studies also provide an opportunity to expand on previous theories (Yin 2003). Thus, this thesis presents an opportunity to make generalisations on the emergence and decline of violence in urban environments and of the perpetrators of that violence.

Within this case study method, Yin (2003) has identified four specific research designs; single- case (holistic), single-case (embedded), multiple-case (holistic) and multiple-case (embedded). The key difference between these types is that a holistic case study design concentrates on a single unit of analysis in either a single-case study or multiple-case studies, whereas the embedded approach utilises multiple units of analysis within a single case study or sometimes multiple case studies. I chose to apply a single-case (holistic) design, focused on levels of violence in Medellín. The key reason for this, is because a ‘single-case can demonstrate or represent an extreme of a unique case and its circumstances’ (Yin 2003, 42). This single-case

38 approach provides me the opportunity to observe, analyse and understand the radically changing levels of violence in Medellín through a historical framework.

While there are strengths to this research method, there are also weaknesses. As Yin explains, ‘the entire nature of the case study may shift, unbeknownst to the researcher, during the course of the study’ (2003, 45). For example, if the high levels of violence had returned to Medellín, this could have altered the overall focus of my research. However, this weakness can be transformed into a strength for the investigation in its capacity to provide a flexible approach to a phenomenon over which the researcher has no control. This flexibility is an inherent element of the research design using a single-case approach. To avoid the potential problem of the nature of the case study shifting during the research period, the research design was constructed with a clear and concise definition of the scope and limitations of the project (Yin 2003). This is why I chose to organise my data in the following way: The emergence of violence from the late 1980s, the peak of the violence in the 1990s and the reduction of violence from the 2000s. Furthermore, to limit the scope of my thesis, I stopped collecting data in November 2015. Using a historical single-case study approach encompasses both the emergence and reduction of violence in Medellín and will assist my research in seeking to answer the overarching question of ‘what explains the reduction of urban violence?’

Another weakness of the case study method is the inability of the case study to provide general conclusions. The findings of this research should also not be generalised to all urban environments in middle- and low-income countries as policy makers should take into consideration the historical, geographical, spatial and structural complexities of an environment when implementing urban violence reduction and prevention policies. Thus, while this thesis does not provide a blueprint for urban violence reduction and prevention, policy makers in urban environments, especially Latin American ones, it does give insight into the rise and fall of violence in a particular context that may have value for policy makers in other jurisdictions.

Selection and Justification of Case Study

As identified in the previous chapter, there is extensive conflict and development literature to explain levels of violence and conflict at national levels. A large amount of literature also focuses on the threat non-state armed groups pose to the state and society. Only recently has this literature discussed violence and conflict in urban environments and the threat urban

39 perpetrators of violence pose to the state. Thus, many of the assumptions underpinning the emergence of violence in urban environments are ‘shaky and untested’ (Muggah 2012, 10). International and national policy makers within this area also have limited knowledge as to the types of approaches that can be effective in the reduction or prevention of high levels of violence in urban environments (World Bank 2011a, 86). Therefore, the case study of Medellín provides a compelling example for testing both the emerging themes in the literature and current practices.

Medellín is a compelling case study for examining the emergence and reduction of violence in an urban environment for three reasons. First, no other city in the world has experienced such critical and high levels of violence. When Medellín was portrayed as the ‘most dangerous city in the world’ with 381 homicides per 100,000 inhabitants, its homicide rate was almost 40 times higher than the United Nations definition of epidemic violence. At the time, this level of violence was only exceeded by Beirut in Lebanon, where civil war was raging. Second, Medellín represents an extreme case of multiple perpetrators of violence including drug cartels, gangs, youth assassins, urban guerrillas, community militia and paramilitaries. Third, no other city had experienced such a dramatic reduction in homicides in less than a generation. The reduction in violence has gained international attention with the city becoming known as the ‘Medellín Miracle’ and as a benchmark for Latin American policy makers seeking to understand and implement successful urban violence reduction and prevention policies (Maclean 2014; The Economist 2014; Muggah 2012). International organisations, such as UN- Habitat, Inter-American Development Bank and the World Bank have praised the policy approaches used in the city for this reduction of homicides (UN-Habitat 2007; Inter-American Development Bank 2009; World Bank 2011a).

Measuring Levels of Violence in Urban Environments

As indicated in the previous chapter, levels of violence in urban environments are difficult to measure. The most widely used indicator is the homicide rate, which is generally defined as the ‘intentional killing of a person’. This is because given homicide’s severity it is more likely than other violent crimes to be recorded (World Bank 2011a, 13). However, there are limitations to relying exclusively on homicide rates to explain and describe levels of violence in urban environments. This can be because of under-reporting, difficulties in interpretation and that the rates fail to capture changes in public perceptions of safety, security, fear and

40 victimisation (McCord 2003; Moser 2004; World Bank 2011a; Muggah 2012). Given these limitations, there is an increased use of complementary qualitative methodologies, which give a voice to people’s perception of violence, by looking at levels of insecurity, fear and victimisation (Moser 2004, 7; World Bank 2011a, 13).

This thesis therefore relies on both quantitative and qualitative data to explain the emergence and reduction of violence in Medellín. The data on homicides rates in Medellín are largely taken from the Medellín-based government department, Information System for Security and Cohabitation (Sistema de Información para la Seguridad y la Convivencia - SISC). This department has the most rigorous approach to data collection of current and historical levels of violence in the city. The department analyses data on violence collected from the Colombian National Institute of Medicine: Legal and Forensic Sciences (Legal y Ciencias Forenses Instituto Nacional de Medicina - INML), Colombia’s Department of Criminal Investigation (Seccional de Investigación Criminal - SIJIN) and Technical Investigation Team (Cuerpo Técnico de Investigación - CTI).

The qualitative data used in this thesis are discussed below.

Fieldwork in Medellín

Fieldwork in Medellín was an important part of gathering qualitative data for this thesis. A significant element of qualitative research is ensuring the accuracy of the account in representing the participants’ realities of the social phenomena and is credible to them (Schwandt 1997). Thus, Creswell (2013) argued the importance of using strategies, such as prolonged engagement and persistent observation in the field, writing field notes and memos, accessing multiple sources of data, peer debriefing and conveying findings through ‘thick’ descriptions to enhance credibility.

Fieldwork for this study was carried out on two occasions; May to August 2014 and April to August 2015. These two visits were important and fundamental not only for the eventual orientation of this thesis but also for my appreciation of the complexities of violence and its reduction in an urban environment. This first-hand experience of living in Medellín was particularly important for understanding how the practices and levels of violence are perceived by citizens. During this time, I engaged in both semi-structured interviews and participant observation.

41

Semi-structured interviews

During the two fieldwork periods, semi-structured interviews were conducted with an array of actors including government officials, former government officials, academics, business leaders, and members of community and civil society organisations. These actors all had valuable knowledge about the emergence and decline of violence as well as the different perpetrators of violence in Medellín. Participants for this research were found using two non- probability techniques: purposive sampling and snowball sampling (Valdez & Kaplan 1998).

In the purposive sampling, samples were drawn strategically from actors who were actively involved in or had valuable knowledge about the growth and reduction of violence in Medellín. This sampling was conducted in three ways. First, before leaving for Medellín in 2014, I emailed academics from the Medellín-based University EAFIT and the University of Antioquia and social commentators who were on record as having some knowledge about violence in Medellín. Second, I used my academic contacts in Australia that I knew had academic, political and media contacts in Medellín to put me in touch with potential interviewees. Lastly, when I arrived in Medellín, I had a business elite contact who was able to provide me with access to business and political leaders in the city. This contact also provided me with access to private seminars that were normally closed to the public. Through this purposive sampling, I was able to build networks in the city and reduce the bias that may derive from a limited sample (Yin 2003, 90).

The snowball sample technique was used to complement the purposive sampling by interviewing actors recommended by participants from the purposive sample. This method is the most effective in accessing hidden and/or hard-to-reach populations (Valdez & Kaplan 1998). Snowballing enabled me to obtain access to a wide range of actors including academics, journalists, leading figures in business, former government officials, government officials, and members of community organisations and non-government organisations. This additional group of diverse interviewees further reduced sampling bias (Van Meter 1990; Yin 2003). The variety of networks, I developed in the city introduced me to many different participants and because of the snowballing I gained the trust and cooperation of the different interviewees as I came ‘recommended’ from the previous participant.

Before leaving Australia in 2014, I designed a series of questions that guided my semi- structured interviews (see Appendix 1). These questions focused on themes that had emerged from the literature mainly concerning explanations for the emergence of violence in Medellín 42 and interventions that reduce violence. These questions probed topics such as particular events or situations the participant thought led to the emergence of violence in the city or if the participant could identify any types of success in the reduction or prevention of violence in the city and the reasons for that success. These questions were open-ended and designed to ensure the interviews were more of a conversation with the participants rather than following a set script (Liamputtong & Ezzy 2005, 55). They also allowed a number of topic-related questions about the patterns of violence in Medellín and suggested questions to be asked, allowing for wide-ranging discussion on the research topic (Burnham et al. 2008, 118). These semi- structured interviews were important for enhancing my understanding of documentary material, clarifying ‘grey areas’, facilitating an understanding of the perceptions and beliefs of participants and corroborating accounts from other participants (Burnham et. al. 2008, 119).

A couple of days before the interview, I would email the interview questions and research objectives to the participant. Upon meeting the participant, I would run through the research objectives orally and confirm the participant had received the documents I had emailed. Earlier, I would also ask the participant to sign a consent form, which advised that the interview would be recorded, their identity would remain confidential and any information they provided me would only be used for the purpose of this research. In cases where participants did not wish to be recorded we had more of an in-depth conversation for which I kept detailed notes.

After the two fieldwork periods, I had 15 interviews with four former government officials, three government officials, three community leaders, two NGO members, one academic, one official from a private institutional alliance and one business leader. I then had detailed notes from 13 in-depth conversations with six academics, three community leaders, two government officials, one ex-government official, one social commentator and one journalist. Data from the recorded interviews are presented in the thesis using the following codes: Academic (A1), business leader (BL1), community leader (CL1), community leader two (CL2), community leader three (CL3), former government official one (FOG1), former government official two (FGO2), former government official three (FGO3), former government official four (FGO4), former government official five (FGO5), government official one (GO1), government official two (GO2), government official three (GO3), and an official from a research center focused on evaluating quality of life in Medellín (OF1).

Each interview typically lasted 60 to 90 minutes. They took place either in the participants’ offices or in public venues (such as coffee shops, restaurants or library parks). The interviews

43 with community leaders and NGOs typically required me to travel to the poorer neighbourhoods of the city. I found conducting interviews in public venues was frustrating as the noise levels in the city meant that at times it was difficult to follow the conversation. Therefore, the recording of the interview was crucial in case I had initially missed something.

Most interviewees were surprised to encounter a female from Canberra, Australia, researching the levels of violence in Medellín. This surprise, however, provided a type of ‘ice-breaker’ for the interview, as I would normally explain to participants before meeting them that I was ‘muy mona’ (local term for a blonde female) so they would recognise me. Once I arrived at the interview location we would have a laugh as they would say ‘you are not wrong, you are muy mona’ (a very blonde female). This ‘ice-breaker’ provided a good opening to start the interview and to establish rapport with participants. Furthermore, because of my previous work and study experience living in Latin America (including Cuba, Bolivia and Costa Rica) I have a very particular Spanish accent, which provided another type of ‘ice-breaker’ as participants would often comment on how ‘you look muy mona but sound Latina.’ Therefore, my personal identity along with the various networks I had in the city were important in establishing trust and rapport with my participants. My Spanish language skills enabled me to obtain rich and textured data, as I did not require the services of a translator. Furthermore, the participants had often been victims of violence in the city, from being robbed to having a family member assassinated or kidnapped. Given this sensitivity, I was always very careful when establishing trust and rapport with all participants.

The interviews conducted with an array of actors, from government officials to community leaders, provided me with deep and rich insights into the complexities of violence in an urban environment. The open-ended questions allowed for a wide-range of discussion on the issues concerning changing levels of violence in the city. These insights provide an important contribution to the study of urban violence as it provided an opportunity to establish how different actors understand and comprehend levels of violence in an urban environment. Such deep and rich insights could not have been achieved through a desktop review of the relevant data, such as government reports, academic literature and statistics, on the changing levels of violence in Medellín.

Participant observation

I also engaged in different forms of participant observation during the two periods of fieldwork. According to Goffman (1989, 125) participant observation involves: 44

Subjecting yourself, your own body and your own personality, to the set of contingencies that play upon a set of individuals, so that you can physically and ecologically penetrate their circles of response to their social situation.

During fieldwork, I kept a detailed field diary of my observations. This was an important tool for not only recording these personal observations but also in trying to make sense of the levels of violence in the city. Throughout the data chapters of this thesis, I make reference to these field diary notes which contribute to the rich thick descriptions evident in the data chapters.

During both fieldwork periods, I lived close to the elite neighbourhoods of Medellín where I shared an apartment with a local person from Medellín. The people from Medellín are known locally as paisas. A contact had put me in touch with this paisa with whom I shared the apartment. This paisa spoke limited English and I quickly brushed up my language skills from my six years of living and working in Latin America and he gave me some lessons in paisa avanzado (paisa slang). Sharing this apartment was important for not only gaining knowledge about day-to-day life in the city but also for gaining knowledge about local security measures. I also engaged in different forms of fitness activities during fieldwork periods. These activities included using the services of a personal trainer from the middle-class neighbourhoods who would take me running around different parts of the city. I also took weekly private dance classes with an individual from the middle-class neighbourhoods.

No one day was similar during my fieldwork. Some days were filled with working out with my trainer in the mornings and following up with contacts and potential interviewees or conducting interviews and finishing the day enjoying a beer and arepa (an arepa is a corn cake and a staple food of paisas) on the balcony of my apartment with my paisa housemate. Other days were spent visiting the poorer neighbourhoods in the mornings, attending security seminars in the afternoon and finishing the day having my nails painted and chatting with a señorita (young woman) in the middle-class neighbourhoods of Envigado. While I was interested in attending events in the poorer neighbourhoods during the evenings, given my reliance on public transport and taxis, I was unsure how reliable and secure transport to and from these events would be. On weekends, I would catch up with my paisa friends from the upper-class and middle-class neighbourhoods or participate in fun-runs organised by the Medellín government. On some occasions, I did have interviews or need to attend meetings on Saturday mornings as the ‘working week’ in Medellín is until midday on Saturdays. In 2014, I was in the city for the World Cup, which was a particularly important event for Colombia as it

45 was the first time in 16 years the national team had qualified. Their advancement to the quarterfinals meant that I was sometimes limited in my access to participants, as no one worked when Colombia was playing a soccer match. However, this advancement also meant I was able to spend some evenings watching soccer matches and celebrating their team’s success in public venues.

I considered it was important to take public transport and taxis and to walk to venues during my fieldwork to gain more understanding about day-to-day life in the city. As I often had meetings in different neighbourhoods throughout the city, I considered these were safe and reliable ways to travel. Taking the Metro (city railway) and taxis and chatting to taxi drivers provided me with perspectives of day-to-day life in the city. Taxi drivers were often from the poorer or lower-class neighbourhoods as compared to my more elite interviewees. I found because of my identity, being a tall blonde Australian female, that these paisas were interested in knowing what I was doing in the city and as we were often stuck in traffic we were able to have long conversations about the changing levels of violence in the city. I also found that walking in the different neighbourhoods allowed me to gain a more informed perspective on the use of public space and of safety in the different parts of the city.

To obtain further understanding of the socio-economic changes in the city, I attended a five- day course at the University EAFIT called Social Urbanism in Medellín: Urbanism, Environment and Society (Urbanismo Social en Medellín: Urbanism, Medio Ambiente y Sociedad). This course provided me not only with a theoretical background to the socio- economic developments in the city but also, I was able to visit the intervention neighbourhoods on numerous occasions. This course also afforded me the opportunity to network with social, business, political and academic leaders in the city and to learn more about the urban and social transformations in the city from environmental, social, public administration and community participation perspectives.

To gain further understanding about security policies in the city, I attended a two-day conference organised by the Colombian National Police (Policía Nacional de Colombia) called ‘The 5th International and 8th National Observatory Network and Research Centre of Crime’ (V Encuentro Internacional y VII Nacional de la Red de Observatorios y Centros de Investigacion del Delito). This conference focused on exchanging experiences and approaches on addressing levels of crime throughout different Latin American countries. I also visited the Medellín Ministry of Security on numerous occasions where I met with employees (including the

46

Secretary of Security) to learn about their methods for data collection and analysis of levels of violence in the city.

From my academic contacts at the University of Antioquia and the University EAFIT, I was able to gain more in-depth understanding of the different theoretical approaches for analysing the reduction of levels of violence in the city. Research from the University of Antioquia generally focuses on human security or bottom-up approaches (led by the Human Security Observatory of Medellín (El Observatario de Seguridad Humana de Medellín - OSHM) and research from the University EAFIT generally focuses on public policies or top-down approaches (led by the Centre of Political Analysis (Centro de Analisis Politico). I had the opportunity to meet with researchers from both the OSHM and Centre of Political Analysis on numerous occasions. An academic contact from Centre of Political Analysis organised a focus group with the research team where I was able to ask about their main findings from the report they had recently completed on the design of public security policies in Medellín. These different approaches to data collection and analysis proved to be vital in providing a holistic view of the interpretations of the changing levels of violence in the city.

During the two fieldwork periods, I attended numerous public events such as the presentation of annual results from a program which evaluates the quality of life in the city called ‘How is Medellín Going’ (Medellín Como Vamos), Centre of Urban and Environmental Studies forums (Centro de Estudios Urbanos y Ambientales – Urbam), seminars at the Museum of Memory (Museo de la Memoria) and public security policy seminars. I also participated in public social events such as ‘The Women’s Race’ (Carrera de la mujer) and ‘I am a woman and I decide to live in a new Medellín’ (Soy mujer, yo decido, vivir una nueva Medellín).

These academic, public policy, community forums and social events helped me acquire a holistic view of levels of violence in the city from both a public policy perspective and from the viewpoints of citizens. Through these various forms of participant observation, I gained an understanding of how different citizens perceived levels of violence and the effect this violence had on their daily lives. I would not have been able to achieve such a deep and rich understanding of the complexities of violence in an urban environment from a desktop review of materials on the changing levels of violence in Medellín.

47

Document collection and analysis

Throughout research for this thesis, I collected and analysed primary, secondary and tertiary documents in order to build and deepen my understanding of the emergence and decline of violence in Medellín. Primary sources are sources that are part of or produced by the event, such as policy documents and newspaper articles. Secondary sources are those that relate to the event and are produced soon after in the form of speeches and reports while tertiary sources are written afterwards to reconstruct the event, and include academic papers (see Burnham et al. 2008, 165).

The collection of primary documentary data was conducted during fieldwork and using online sources. I obtained access to the El Colombiano newspaper archives at the University of Antioquia and spent countless hours taking copies from these archives. Due to the time limitations, I was not able to obtain every single record. However, I built up an ample collection of relevant newspaper items on the period studied. I was also able to obtain access to the newspaper archives of El Tiempo via their website. I followed different online media sources, which reported on levels of violence in the city, such as Minuto30, Analisis Urbano, Alcaldía de Medellín, Medellín Como Vamos, Qhubo Medellín, El Colombiano, Denuncias Ciudadanas, Verdad Abierta and Periodico ADN Medellín. My local contacts advised me on the most important documentary materials on the levels of violence in the city, such as No nacimos pa’ semilla, La Sierra, Rodrigo D no futuro, La venedora de rosas and El virgin de los sicarios. I obtained access to influential reports such as: ‘Medellín: Re-encounter with the future’ (Medellín Re-encuentro con el futuro) and ‘PRIMED: Report on the Subnormal District Improvement in Medellín Program’ (Programa Integral de Mejoramiento de Barrios Subnormales en Medellín). The staff at the University of Antioquia library and my local contacts directed me to the most influential tertiary documents such as academic papers and books written by local researchers on levels of violence in the city.

Researching Violence and Conflict

Given the topic of this thesis, it was important to address any safety issues that could have been encountered during the fieldwork periods. Throughout the methodological literature much is written about the issue of researcher and subject safety in dangerous environments (Lee & Stanko 2003; Rodgers 2001; Lee-Treweek & Linkogle 2000). Researchers have even been murdered in Latin America and they must be particularly cautious in conflict zones or areas of

48 intense urban violence (Lee-Treweek & Linkogle 2000, 10). This issue of researcher safety was an important consideration during my time in Medellín, as in 2015 three human rights leaders were assassinated in less than a month and a local journalist was threatened shortly before my arrival (El Espectador 2015a & 2015b).

Given the nature of the research and my personal identity, I needed to find ways to mitigate and compensate for these challenges to ensure the data I collected and have presented in this thesis provide accurate representations of the realities of violence in an urban environment. For example, some methodological literature has highlighted the difficulties of ‘outsiders’ gaining access to participants in environments with high levels of violence and conflict or a history of violence and conflict, as in these environments there are often high levels of distrust (World Bank 2011a, 34). Locals may be sceptical of a researcher’s motives, groups may be united and thus relatively closed to outsiders or locals may be conditioned to dissemble when outsiders are asking questions, as such outsiders are often suspected of being tax collectors, police authorities or housing inspectors (Belousov et al. 2007; Robben & Nordstrom 1995; Goldstein 2014). Methodological literature has also argued how in some fieldwork contexts a researcher’s identity, such as their race, class, gender or age, can place them at risk and ultimately have an effect on the objectivity of their data (Roguski & Tauri 2013; La Pastina 2006; Wesely 2006; Coffey 1999).

Homicide rates were low for the city when I was conducting research in Medellín in 2014 and 2015 with ‘the lowest homicide rates in 35 years’ (Alcaldía de Medellín 2015). Despite this, the following excerpt from my field diary in 2015 gives an indication of how significant levels of violence were still evident:

A young male hung himself today down the road from my house in public view. I saw the pictures - it happened 5 minutes after I walked down that street ... Last night a 13- year-old boy was shot – they [the media] said it was from a bala perdida (stray bullet) from locals celebrating Colombia winning the soccer - but others are saying it could have been from the gangs fighting over territory …. There was a massacre of four people just outside the city on Saturday… With machetes.... The other night a pregnant woman was shot in her sleep in one of the poorer neighbourhoods…

Given these levels of violence, I was required to take certain precautions to ensure I was safe, which required being ‘streetwise’ and using my common sense, local knowledge and language

49 skills (Baird 2009). For example, I only conducted interviews during daylight. I researched the particular neighbourhood I was visiting beforehand and always advised a local contact of where in the city I would be for the day. Often my paisa housemate would remind me ‘ten cuidado’ (be careful) when I departed for interviews. He would also always remind me to order taxis via the secure mobile phone app EasyTaxi. However, I soon realised that the majority of citizens from the elite neighbourhoods had not visited the poorer neighbourhoods of Medellín in recent years, if ever, and I often had more knowledge of these neighbourhoods than these paisas did.

During interviews conducted in public venues, I noticed some participants ‘cautiously’ looking over their shoulders or slightly lowering their voice when discussing the particulars of the perpetrators of violence in the city. In one interview in the San Javier Library Park (Parque Bibloteca San Javier), the participant and myself heard the sound of a camera clicking as if someone was taking a picture of us and we soon noticed someone had been listening to our conversation. Situations like this made me realise that while my particular identity could be an advantage for establishing rapport with participants it also presented disadvantages, as I was not able to easily ‘blend in’ throughout the city.

Despite being ‘street-smart’, not all dangerous situations can be totally managed. While some violence can be partially predictable, other violence is spontaneous and emotive and even the most prudent researchers could be unlucky (Jamieson 2002, 64). For example, after spending the day in the Universidad de Antioquia I would sometimes arrive home to see on the news that the campus had been attacked by encapuchados (hooded vigilantes) and that the riot police had arrived shortly after I left (El Colombiano 2015a). Other days I would read that I had left the centrally located Parque Botero shortly before there had been the public lynching of a rata (rat, local name for a criminal) (Minuto30 2015a). One day I had been running late and narrowly missed a shooting close to the Universidad EAFIT (El Colombiano 2015b). Furthermore, while conducting fieldwork in 2015, one of the potential participants for this research was shot 25 times in a poor neighbourhood. I wanted to interview this community leader about the knowledge he had about bandas co-opting the participatory budgeting process. Some media reported that he had been threatened and assassinated by the bandas because of this knowledge (El Espectador 2015b). Following this assassination, I was advised by local researchers to avoid any ‘unnecessary risks’, with one counselling that ‘you are at a lower risk given that you are foreigner. However, you should be more careful now.’ I interpreted this as advising me to limit my visits to the poorer neighbourhoods of Medellín and if I had to return to them, I would only do so accompanied by someone else. While the individual who was

50 assassinated was not the only person to be assassinated during my fieldwork, this particular event gave me a very strong reminder of the levels of violence in the city.

Furthermore, while qualitative data are important to show how people experience violence, the sensitivity of the topic can create difficulties in generating accurate data, particularly as perceptions of security can alter dramatically with changing circumstances (World Bank 2011a, 10). One example of this during my fieldwork was when following three grenade attacks in the centre of the city in 2014, one individual from a private-institutional alliance commented on how these attacks had increased their fear of visiting the centre of the city. However, another government official did not express any such fear, thus showing that people took different behavioural lessons from such events.

Triangulation of Data

I found that my acquired understanding of the local context combined with my various networks enabled me to cross-check data to ensure I was obtaining reliable data about the incidence, nature and perpetrators of violence in the city. However, given the complexity of the explanations of violence and different interpretations, it was not always possible to obtain a complete understanding of all acts of violence. I needed to be creative, flexible and adopt a wide range of data collection methods to achieve the objectives of this research and ensure that the data presented are valid and reliable.

By analysing the transcripts and notes from my interviews, participant observation and document analysis, three themes emerged. First, it became apparent that despite the significant reduction of homicides in the city, the threat of violence had not disappeared. Second, I began to see the complexity of the makeup of and interactions among the perpetrators of violence in the city. Participants interchangeably used labels such as criminal gangs (bandas), delinquent gangs (combos), the office (la oficina), self-defence groups (autodefesanas), paramilitaries (paramilitares), the mafia (la mafia), demobilised paramilitaries (neo paramilitares), armed illegal groups (grupos ilegales armados) or the armed actors (los actores armados) to identify the previous and current perpetrators of violence in the city. In some interviews, participants even used all of these labels when making reference to one group. I began to understand that from the perspective of citizens it was not important to classify these actors into specific categories, but rather that throughout the city there was a continued presence of organised

51 groups which engaged in acts of violence. The third theme which emerged was I began to understand more about the limitations of the government policies in addressing the levels of violence in the city. The perception of violence from community members in terms of levels of insecurity was very different to the perception of violence from government officials as measured in homicide rates.

The above themes were only made possible through the two periods of fieldwork in Medellín. I would not have been able to obtain such deep and rich insights into the complexities of violence in an urban environment through a desktop review of the relevant materials on the changing levels of violence in Medellín. The methodology of this thesis therefore makes an important contribution to the academic literature and policy-focused analysis on urban violence. The case study approach provides a deep and rich understanding of the complexities of urban violence in a specific urban environment. The reliance on qualitative data also provides important insights into how an array of local citizens perceived levels of violence and the effect this violence had on their daily lives. Using a single method would not have brought such rich returns and the thick description that characterise this thesis. However, as discussed throughout this chapter gathering data in an environment with high levels of violence and conflict is not a straightforward process, and requires a researcher to have flexibility, problem- solving skills and patience.

.

52

Chapter Four: The Emergence of Urban Violence in Medellín

Introduction

This chapter provides a historical narrative on the emergence of increasing levels of violence in Medellín in the 1980s. There is a discussion of the themes and concepts that might help explain the emergence of these high levels of violence, focusing on the effects of the wave of migration to the city that began in the 1950s. The chapter also delineates the involvement of different perpetrators of violence, focusing on their characteristics, methods and motivations. This chapter shows the complexity of the situation and that no single event, actor or circumstance throughout the 1980s can be considered in isolation when analysing the emergence of the high levels of violence in the city. The data presented in this chapter show the importance of taking a multifaceted approach to understand the emergence of violence in Medellín.

The Most Progressive City in Colombia

In 1947, Life magazine dubbed Medellín a ‘capitalist paradise’ (Hylton 2007, 75). Medellín was acknowledged as the most progressive city in Colombia, due to multiple factors, including the production of gold and coffee, the construction of the Antioquia Railway, the discovery of coal in the adjacent region and the construction of hydroelectric plants. This ‘Manchester of Colombia’ had skyscrapers, cinemas, wide avenues, parks, monuments and commercial boulevards for pedestrians (Hylton 2007, 75). The growth in industry generated the development of factories that provided employment and an improved standard of living to many citizens. During data collection, all participants commented on the industrial growth in Medellín with a former government official (FGO3) interviewed in 2015 explaining how this city growth led Medellín to become the leading industrial centre of Colombia. This official said: ‘Medellín was known internationally as the Manchester of Latin America. It was the industrial and business centre of the country. This was because of the previous industrial boom in gold mining and coffee. From this international capital in trade, the profits came back to the city, a city, which had a Mediterranean feel, which was really

53

closed to outside influence. From this international capital came the best companies in Colombia. Many of the smaller banks had consolidated with Bancolombia, there were also textile companies, cement companies, soft drink companies. Basically, all of the important industry companies of today [2015] had started in Medellín. Because of these business industries, there was a city with a superior prosperity and citizen culture from the rest of Colombia. Medellín had a large group of business elites, a private sector and more or less efficient government officials. Medellín had a good electricity system, good public lighting and important public institutions which helped the city’.

As indicated by this government official, the development of Medellín as the leading industrial centre of Colombia was built on the success of the entrepreneurial activities of elites during the mining and coffee booms. During this time of economic growth, Medellín was self- characterised by four ‘Antioquia’ values. Antioquia is the state in which Medellín is located. These values were a ‘strong work ethic, hierarchy, authority’ and the elite paisa cultural value of ‘investing in one’s people’ (a paisa is someone from the region of Antioquia where Medellín is located) (Melguizo & Cronshaw 2001, 113; Lamb 2010, 36; Drummond et al. 2012, 147). This investment meant that businesses prided themselves on having ‘good conditions, housing and schools for their employees and their employees’ families’ (Maclean 2014, 17). Thousands of migrants were attracted to the city because of this economic prosperity and the population of Medellín grew from 60,000 in 1901 to nearly 360,000 in 1951 (Drummond et al. 2012, 147). This was an unexpectedly large and unplanned population growth, with many migrants settling ‘illegally’ on the hillsides of Medellín.

To add to the population of an already expanding city, the violent conflicts in the rural areas between Liberals and Conservatives (known as La Violencia) meant that hundreds of thousands of peasants and farmers fled the countryside for the relative safety of urban settlements throughout Colombia, such as Medellín. La Violencia ended in 1958 with the National Front (Frente Nacional), a power-sharing agreement through which Liberals and Conservatives alternated power. However, the power-sharing agreement had not acknowledged rural peasant groups, mainly left-leaning and communist, which had made demands on the government to recognise their political representation. During La Violencia these groups had occupied the lands of the elites and had been the target of many government-sponsored military offensives (Drummond et al. 2012, 147). By the mid-1960s, these peasant groups had banded together to form leftist revolutionary guerrilla groups such as the Fuerzas Armadas Revolucionarios de

54

Colombia (Revolutionary Armed Forces of Colombia - FARC), the Ejército de Liberación Nacional (National Liberation Army - ELN), the Movimiento Abril de 19 (19th of April Movement - M-19) and others in their conflict against the Colombian state to acknowledge their political demands (Drummond et al. 2012, 148). These increasing levels of violent conflict involving the left-wing groups in the rural areas meant that thousands of families migrated to Medellín, which was largely unaffected by these violent conflicts. During data collection, all participants spoke of how these migrants were escaping both high levels of violence and poverty, with one former government official (FGO3) interviewed in 2014 explaining it as:

‘Families arrived for two reasons. Either they were displaced because of the war in the countryside or they arrived looking for better economic opportunities, such as the families from the Choco region. They settled in the informal neighbourhoods just outside of the city and they created the Medellín version of favelas [a Brazilian shack or shanty town]. These families did not have a father and the mother was the head of the family. So, they did not really have an authority figure in the family and they did not have an economic provider. These families were very unsettled…’.

Another former government official (FGO5) interviewed in 2015 also spoke of the disorganised manner in which these families were arriving:

‘The displaced people from La Violencia and the poorer people started to populate the city in a disorderly manner and started living on the hillsides of the city where they had an extremely low standard of living…’.

As indicated by these officials, the new migrants were settling in an unregulated manner on the outskirts of the city. The local government did attempt to acknowledge the needs of the growing number of migrants with the Medellín Master Plan (Plan Piloto - MMP). This urban plan initiative had been created in 1950 and was sponsored by local government officials, industrialists, entrepreneurs and cultural leaders (Hylton 2007, 73). The MMP had numerous objectives including regulating the growing settlements of the recently arrived rural migrants, the creation of an industrial zone, the canalisation of the Medellín river, and the construction of a soccer stadium and administrative centre (Drummond et al. 2012, 147). But, the rural migrants were arriving so rapidly that the MMP leaders struggled to deal with the growing informal neighbourhoods on the hillsides of the city and ‘were unable to provide them with

55 services such as sewage, water, schools, roads and health care’ (Drummond et al. 2012, 148). All participants interviewed commented on how the state was unable to address the social and economic needs of these migrants with a community leader (CL1) interviewed in 2015 stating:

‘People arrived with nothing and the state did not give them any attention…These families had no way of finding food, of finding housing and they started doing anything to survive… Imagine these people who came fleeing from the rural areas from guns and sadness, from police and state violence, and they arrive to this city and the mayor says ‘I am not going to receive you because you are hungry or you don’t have money to stay here, you need to leave.’ But these families did not leave, they stayed and started living closer to the rivers and the hillsides of the city…’.

As indicated by this community leader not only were these rural migrants economically and socially excluded from the wider society, but also, they were politically excluded. This political exclusion was because traditionally, throughout Colombia, in both rural and urban environments there were structures of reciprocal relations of exchange between patrons and clients (Archer 1990, 10). Patrons always wielded the most power as they could grant access to the resources clients wanted. The patron-client structures were carried over into democratic politics, where incumbents and aspirants for office used them to grant favours using public resources in exchange for electoral support. In this system of political clientelism, politicians were seen as ‘bestowers’ of rights rather than representatives or leaders who supported civic and political rights (Archer 1990, 20). Therefore, local politicians did not need to be responsive to the needs of rural migrants as a politician’s loyalty was only to the person who appointed them. This pattern of social exclusion and inequality in access to basic services was a constant source of tension between migrants and government throughout this period.

Another source of tension was the marginalisation and exclusion of these informal communities from the larger society. This was because the rural migrants that had arrived in Medellín settled on the outskirts of the city in areas that had been considered ‘unbuildable’ by the MMP and were intentionally ‘not incorporated into the city’s urban perimeter’ because of their unstable nature on hillsides (PRIMED 1996, 28). Furthermore, the new migrants often arrived in the city lacking resources to access housing in the high-cost formal market and they were forced to ‘illegally’ occupy the hillsides of Medellín by purchasing fake titles to the land

56 from pirate developers (Roldan 1999; Samper 2012). While there were some state resources provided to these neighbourhoods, the police and army units were ‘sent out routinely to demolish these hillside settlements’ (Hylton 2007, 77). Such demolition exemplified the crisis of local authority in the city. While the state acknowledged the existence of the informal communities and their illegal land appropriation, the state did not have the power or will to deal with the repercussions of evicting residents from their homes (Samper 2012, 7).

Around this time, the elite paisas referred to these growing informal neighbourhoods as las comunas. The term, comuna, which literally means ‘comune’ or ‘community’ in English, is used in Colombian cities to delineate territorial divisions. However, comuna became synonymous with the poorer neighbourhoods of Medellín. Over the years, the local government did officially recognise these informal neighbourhoods and Medellín became organised into 16 comunas, with 271 neighbourhoods and 11 to 26 neighbourhoods in each comuna. The city eventually became divided into the following zones: North-eastern (comuna 1 to 4), north- western (comuna 5-7), east-central (comuna 8-10), west-central (comuna 11-13), south-eastern (comuna 14) and south-western (comuna 15-16) (Alcaldía de Medellín 2016). The distribution of these comunas is noted in Figure 4, showing both the numbers and the names of each comuna.

57

Figure 4: Map of Medellín, Colombia

Image Removed due to Copyright

Source: Alcaldía de Medellín 2016

The corregimientos (territorial subdivisions) presented in Figure 4 include the rural zones of Medellín. From the 1930s, the comunas which experienced unregulated rapid growth were comunas one to eight, thirteen and fifteen. Comunas nine to twelve were part of the MMP, with comuna ten the economic centre of the city and comuna fourteen, El Poblado, was originally a rural area which became the most elite part of the city by the mid-1970s.

Growing Inequalities in Medellín

Medellín emerged relatively unscathed from the La Violencia, as this conflict had been mainly concentrated in the rural areas, and the city’s economy had grown strongly from the 1950s. While Medellín was the most ‘progressive city’ in Colombia it also led the nation in social problems as manifested in high levels of prostitution, alcoholism and imprisonment. These were all ‘undesirable characteristics’ that the conservative elite and Catholic Church sought to overcome by promoting their views of appropriate morality and the Antiqueño work ethic (Drummond et al. 2012, 147; Melguizo & Cronshaw, 2001, 113; Lamb 2010, 39). From the 1950s, the Catholic conservative elites consciously promoted an image of the city as an ‘oasis’ of peaceful capitalist productivity beneficial to the nation due to the social responsibility of its major industrialists (Hylton 2007, 76). These elites and the Catholic church played a major role in promoting adherence to these values. However, as observed by a former government official (FGO3) interviewed in 2015, this type of social contract was not ideal for all paisas:

‘The private industry had a strong influence on people and their behaviour. The society also had a strong Catholic belief and the Catholic church played a strong role in regulating public morals. Society was regulated almost by the church and business. But, because of the increase of displaced people in the city, not everyone was employed and there was also a global change in the role of the Catholic church and in Colombia the church stopped playing such a big role in regulating public morals’.

58

As indicated by this official, the rapid wave of urbanisation and the ensuing social and economic exclusion of the new migrants meant they were not socialised into the mainstream society and its prevailing values. By the late 1950s, social problems in Medellín continued to increase, diversify and intensify. For example, by 1960, 17.6 percent (equal to 64,000 people) of the population were not studying, working or looking for work (PRIMED 1996, 36). The migrants continued to arrive with the city tripling in size from 358,000 in 1951 to 1,071,000 in 1973 largely due to growth in the unregulated and un-serviced settlements (Drummond et al. 2012, 148). The MMP had planned for a population of 600,000 in 1971, but this number had been reached by around 1958 and the MMP leaders struggled to provide work and education for these new migrants (PRIMED 1996). By 1972, 50 percent of the population (equating to 600,000 people) were living in informal settlements (PRIMED 1996, 29). In the ensuing years, these informal neighbourhoods which exceeded the formal city in both physical area and population size and were host to ‘extreme poverty’ (Insuasty et al. 2010, 33; Melguizo & Cronshaw 2001, 113).

To add to these growing levels of poverty, there was an economic crisis in the textile factories of Antioquia and their supporting industries. Inexpensive exports from Asian manufacturers reduced demand for local exports (Hylton 2007, 76). Given that Medellín was the industrial centre of Colombia, its prominent industries went into decline and already high unemployment rates skyrocketed. By the end of the 1970s, Medellín had the highest unemployment rates in the country (PRIMED 1996, 34). Furthermore, as social rights, such as housing and education, were dependent on employment, the results of this mass unemployment were catastrophic and the quality of life for many citizens in Medellín deteriorated significantly (Maclean 2014, 18). Coffee prices had also fallen in the 1960s and 1970s, leading to a further influx of rural migrants seeking a better economic life in Medellín, with the majority of these new migrants inevitably settling in the crowded informal poor neighbourhoods.

Medellín was changing from the ‘capitalist paradise’ to a city with an increasing informally settled population. These informal neighbourhoods were the crowded products of unregulated growth. There were high levels of economic and social inequalities. The newer migrants were both spatially and socially distanced from those who had traditionally controlled and lived in the city in terms of class, rural background and often political ideology (Maclean 2014, 16). Furthermore, these residents were seen as possessing ‘an aura that each inhabitant was, by de 59 facto, a criminal’ and a threat to urban society because of their illegal appropriation of land (Samper 2012, 8). Planning decisions about how to cope with the ‘new migrants’ were predicated on security concerns rather than inclusion or integration. For example, highways were built to bypass these poorer comunas, allowing residents in the more affluent parts of the city to avoid these areas rather than connect with them (Maclean 2014, 16). The poor communities had a lack of state governance and were isolated economically, socially and geographically from the formal population. The inhabitants of the informal communities were excluded from the wider society resulting in vertical inequalities because of the stark contrast in social class, status and power, and, horizontal ones because of the political differences between groups. These horizontal and vertical inequalities limited the life experiences of the new migrants, as predicted by the structural violence approach of Galtung (1969).

The Growth of Narco-Trafficking

During the mass wave of urbanisation in the 1960s, there were small gangs throughout Medellín. These could be broadly separated into the criminal ones, called the galladas and the vigilante groups, called the encapuchados (hooded vigilantes). The members of the galladas were underprivileged young men with few prospects reacting ‘to the pressures of social class segregation and blocking of upward mobility’ (Melguizo & Cronshaw, 2001, 117). These members had low visibility and crime and any violent deaths were generally associated with individual acts rather than from organised group action (Melguizo & Cronshaw, 2001, 115). By contrast, the encapuchados engaged in social cleansing of drug addicts, prostitutes, homosexuals and criminals in the poorer neighbourhoods. They also conducted ‘summary trials’ for some petty criminals in their host communities to maintain social order. These vigilantes were often supported by the elite society because they provided a form of social control for those who did not fit the image of Medellín as an oasis of peaceful capitalist productivity (Lamb 2010, 39).

Around the early 1970s, smuggling was a ‘common side job’ of respected businessmen and officials. This smuggling involved ‘everything from stolen cars and appliances to emeralds and cigarettes’ (Lamb 2010, 39). These smugglers took advantage of the breakdown in law and order, a breakdown to which they had greatly contributed, as some of them were corrupt officials and seemingly legitimate businessmen. By the late 1970s, this ‘common side job’

60 started to be transformed by the global expansion of the marijuana trade which led to aggressive protection of transportation routes. These transportation routes became critical elements in the subsequent growth of one of the most profitable businesses in the world, the delivery and transport of illicit narcotics from cultivation to retail sale, or narco-trafficking (American Security Project 2013, 2). Academics, business elites and officials from NGOs interviewed in 2014 and 2015 stressed how the growth of narco-trafficking was one of the reasons why the violence in Medellín spiralled out of control. The following four interviewees’ comments (A1, FGO4, GO1 and GO3) are typical of the responses:

‘In Medellín narco-trafficking brought a change to the reality’. [A1]

‘I don’t think you can associate the violence in Medellín with one social factor or say that the violence started in a particular year. In Medellín, there are various causes and one of them is narcotrafficking’. [FGO4]

‘You cannot say that it was just one factor [which explains the emergence of violence in Medellín]. I think that there are various causes and one of them is narco-trafficking’. [GO1]

‘Narco-trafficking had an effect on all spheres of life, such as politics, institutions, the private sector and the media’. [GO3]

From the 1970s, the expansion of the illegal drugs trade laid the foundation for the Colombian cocaine industry to become one of the ‘most profitable businesses in the world’ (Forbes 2012). Critical to this growth was the role of one individual from a poor rural family, Pablo Emilo Gaviria Escobar (1949-1993) (Otera 1991, 957). Escobar had a long criminal career that started at a young age and by 1974 he was ‘deeply involved’ in illicit trade. Escobar knew the importance of having established routes to traffic cocaine and by the mid-1970s, he had established a monopoly on the cocaine-trafficking routes out of Colombia. Escobar had first became involved in the cocaine-trafficking business in the mid-1970s when he met the Ochoa Restrepo family. The father of the family, Fabio Ochoa Restrepo, was a horse breeder and a successful and well-respected businessman who ran a side business smuggling stolen home appliances and contraband whisky (Lamb 2010, 50). Escobar persuaded Fabio Ochoa Restrepo’s sons (Jorge, Juan David and Fabio Ochoa or the ‘Ochoa brothers’) to use their family’s smuggling routes and contacts to start transporting cocaine into the US.

61

By 1981, the Ochoa system was at its limit and Carlos Enrique Ledher Rivas entered the business. Previously, the primary method for smuggling cocaine was overland. This method changed when Ledher Rivas begun to use a small private aircraft to smuggle cocaine into the US (Filippone 1994, 324). Ledher had raised enough funds with a marijuana trafficker to buy a small airplane and a small island, Normany’s Cay in the Bahamas, which were used to transport large shipments of Colombian cocaine to Miami on behalf of the Ochoa brothers and Escobar (Lamb 2010, 50). Escobar was paid for each kilo of cocaine that was moved and earned a significant level of financial autonomy. Using small aircraft to traffic cocaine led to domination by Colombian suppliers with 80 percent of the refined cocaine available in the US market coming from Colombia (Bagley 1988). By 1987, Escobar along with the Ochoa brothers were included on the Forbes World Billionaires List with Escobar named as the ‘Seventh Richest Man on Earth’ (Forbes 2012). Escobar and the Ochoa brothers had taken advantage of the breakdown in law and order in Medellín and elsewhere in Colombia, and had contributed to this breakdown to achieve their economic success, thus reflecting the greed perspective as an explanation of the emergence of violence (Collier & Hoeffler 1998; Collier 2000; Collier & Hoeffler 2004).

However, not all was perfect in their multi-million-dollar business. This was because in 1981, a small urban left-wing guerrilla group, called the M-19, kidnapped the Ochoa brothers’ sister (Nieves Ochoa), demanding a very high ransom for her release. The M-19 knew of the wealth of the Restrepo family, so they thought they were an easy target (Lamb 2010, 54). Escobar and the Ochoa brothers however refused to pay the M-19 and instead formed a group called Death to Kidnappers (Muerte a Secuestadores - MAS) with the objective of putting an end to kidnapping. Given the objective of the group, it was supported by not only other narco- traffickers but also police officers, landowners, politicians and businessmen (Lamb 2010, 54). The Ochoa sister was eventually released without any payment to the M-19. However, to make their message clear MAS continued its killing spree and targeted insurgents, suspected insurgents and leftist sympathisers. They killed brutally, ‘disembowelling and hanging bodies from trees to discourage the population from cooperating with the M-19’ (Filippone 1994, 340). This use of brutal violence by MAS presented a challenge to the state’s monopoly of violence as the ‘narco elite’ were imposing their own brand of social control in the city.

The growing cocaine market had made Escobar and the Ochoa brothers millionaires. However, given their core activities involving the transportation of illicit narcotics from cultivation in

62

Colombia to their sale in the US, a cheap form of cocaine paste called bazuco entered the poorer neighbourhoods of Medellín. Given that bazuco was readily available, cheap and extremely addictive it had a similar effect on Medellín’s poorest neighbourhoods as crack cocaine did in the US in the 1980s (Lamb 2010, 46). Bazuco addicts, mainly uneducated and unemployed individuals from the poorer neighbourhoods, started robbing homes, stores, vehicles and individuals to maintain their habit (Jaramillo et al. 1998, 20). These individuals began to be known as chichipatos, ‘low lives’ or ‘disposable people’ and were frequently subjected to the social cleansing campaigns of the encapuchados (Lamb 2010, 46).

Pablo Escobar: Antioquia’s Robin Hood

As the chichipatos were terrorising the poorer neighbourhoods and being victims of social cleansing campaigns, Escobar was becoming more well-known. This was because Escobar provided the poorer neighbourhoods with social services, which the MMP and the Medellín local government were unable or unwilling to provide (Filippone 1991, 337). Examples of these services included when Escobar organised and funded nearly 100 neighbourhood committees to undertake community projects, such as ‘Medellín without Slums’ (Medellín sin Tugurios) seeking to ‘eradicate the slums of Medellín’ with the construction of 5,000 homes and the building of soccer fields, stadiums, churches and schools (Lamb 2010, 52). Escobar used his newspaper, Civic Medellín (Medellín Civico) and television program, The Day of Antioquia (Antioquia al Día) to publicise the good works that he and his fellow ‘businessmen’ were undertaking (Lamb 2010, 52). In 1983, this ‘social work’ led to the Colombian newspaper, Semana devoting a lengthy cover story to Escobar, calling him ‘Antioquia’s Robin Hood’, that he was one of the ‘most popular personalities in the Antioquia department’ and a man of ‘simple presence and personality.’ The wealth of Escobar was a ‘mystery’ to the elite paisas, many of whom did not suspect he was in charge of the world’s largest narcotics business. Escobar commented that his wealth had come from ‘owning a bicycle business, playing a type of lottery (chance), buying and selling cars and land negotiation’ (Semana 1983). Individuals from the poorer neighbourhoods had been abandoned by the state and Escobar had provided them with crucial services. During data collection, all participants commented on how Escobar was perceived favourably in the poorer neighbourhoods. For example, a former government official (FGO5) interviewed in 2014 commented on how Escobar was seen as an idol:

63

‘There is a neighbourhood called Pablo Escobar. Escobar built houses for the poorer people, he invested in many things for these poorer people. He had also come from a poor background so he wanted to do many things to resolve the needs of the poor people. He invested in football fields, social investments. There was even a football club which was financed by him and narco-trafficking. Escobar was seen as an idol!’.

Escobar also provided a sense of a governing body in these communities. For example, he did not encourage or support members taking bazuco and claimed to be working for the good of the community (Filippone 1994, 328). The sense of a governing body Escobar provided was in stark contrast to local political leaders. This was because the clientelistic political structures in the city meant that local politicians did not need to be responsive to the needs of these poorer communities, as a politician’s loyalty would only be to the person who had appointed them rather than being a representative who supported civic and political rights for all the city’s inhabitants. Escobar, however, was perceived as a leader who did support civic and political rights. Another participant, a government official (GO3) interviewed in 2014, commented on how Escobar was perceived favourably by residents in these communities:

‘Escobar…ended up looking like a type of Robin Hood, he becomes a local popular hero…’.

These acts of providing welfare and social services to the poorer neighbourhoods led Escobar into politics. He had gained the support and loyalty of many individuals from the poorer areas and, in 1982, campaigned for a seat in the House of Representatives. Escobar was firstly publicly rejected by the New Liberalism Party (Partido Nuevo Liberalismo), an anti-patronage party led by Rodrigo Lara Bonilla and Luis Galan. They rejected Escobar as they suspected that his wealth did not come from legitimate sources but from narco-trafficking. Because of this rejection, Escobar joined the Liberal Party with Jairo Ortega. Ortega won the congressional seat, in which he alternated with Escobar, as permitted by Colombian law. By gaining a seat in the House of Representatives, Escobar acquired both formal political power as well as additional support from the poorer neighbourhoods of Medellín.

One of Escobar’s key political concerns was the Extradition Treaty of 1979. This treaty removed the drug traffickers from their ‘untouchable niche’ in Colombia by enabling their extradition to the US. Escobar was known nationally for his campaign against extradition, calling it a ‘violation of national sovereignty’ (Lamb 2010, 58). He had organised the first anti-

64 extradition forum in Medellín in May 1983 where over 500 people attended (Semana 1983). Escobar and the Ochoa brothers even referred to themselves as ‘The Extraditables’ (Los Extraditables) with the slogan ‘we prefer a tomb in Colombia to prison in the US.’ By 1983, the US had made ‘dozens of requests’ for extradition under the treaty (including for Carlos Ledher Rivas) but the then Colombian President Betancur, had not signed a single extradition order. This was because Betancur preferred to try these Colombian drug traffickers in the Colombia judicial system, appealing to nationalist sentiments (Otera 1991, 969). Initially, Escobar’s campaign against the Extradition Treaty was peaceful, but in the following years, it became a violent campaign of killing, bombing and kidnapping that placed Medellín on the frontline of a war against it.

Escobar only had a short political career. This was because he was forced to withdraw from his positon in late 1983 after the Minister of Justice, Lara Bonilla, publicly raised questions about Escobar’s involvement in narco-trafficking. With this public denouncement came increased public attention and Escobar went into hiding in January 1984. Lara Bonilla continued to ‘wage a war against drug traffickers’ having a major breakthrough in 1984, when Tranquilandia, a large cocaine processing lab located in the jungles of Caquetá, was raided and destroyed (Rubio 2013, 83). This raid was the largest one of the time, with 27,500 pounds of cocaine seized, with an estimated street value of $1.2 billion (Otera 1991, 970). After the raid, it was revealed the processing lab belonged to a group which the Drug Enforcement Agency (the DEA) called the ‘Medellín trafficking syndicate’ which later became known as the ‘Medellín cartel’, and the left-wing guerrilla group, the FARC provided protection for the aircraft runways and laboratories (Salazar 1990, 122). The relationship between the narco-traffickers and the left- wing group was significant as it was one of the first times that an armed left-wing group was known to be cooperating with a narco-trafficking business (Jaramillo et al. 1998, 397). The participation of this left-wing organisation in narco-trafficking also showed how this group was not only using violence to redress their exclusion from the state, but they were also taking advantage of the breakdown in law and order to profit from doing business with narco- traffickers to fund their struggle.

The political exclusion of Escobar and the raid on his Tranquilandia business led Escobar to retaliate by putting out a contract for the assassination of the Minister of Justice, Lara Bonilla, who was gunned down in Bogota in April 1984 by two young paisas on motorbikes. This type of hired assassin was known as a sicario (Lamb 2010, 57). The assassination of Lara Bonilla 65 was significant for two reasons. The first reason was because this was the first time such organised and selective violence had been used against a leading public official. The second reason was because it marked the beginning of the employment of this type of violence (the hiring of sicarios) which became a significant feature of drug-trafficker’s techniques from Medellín over the ensuring years. This public assassination brought with it an escalation in levels of violence in Medellín as public authorities searched for the person responsible for hiring these sicarios. A former government official (FGO4) interviewed in 2014 recalled the significance of the assassination of this presidential candidate and the response of the Colombian state against the narco-traffickers:

‘In 1984, the Colombian government decides that they needed to fight back following the assassination of the Minister of Justice, Rodrigo Lara Bonilla, by Pablo Escobar. The assassination of Lara Bonilla was like the breaking point and the Colombian state used the extradition treaty as a powerful tool against the narcotraffickers. This is because in the United States, the jails are more severe than Colombian ones and the criminal sentences are longer’.

As indicated by this official, the Colombian government used the extradition treaty as a powerful tool against the narco-traffickers. President Betancur had reacted to this public assassination by declaring war on the narco-traffickers. Betancur invoked state-of-siege powers that produced a ‘wave of arrests, confiscating property and drug-processing labs’ especially in Medellín (Otera 1991, 970). Over the next two and half years, Betancur signed and the Colombian Supreme Court approved extradition orders for thirteen narco-traffickers. However, in response to these extradition orders, 13 judges were assassinated throughout Colombia by members of what was becoming known as the Medellín cartel (Otera 1991, 970). The narco- traffickers continued to use organised violence on selected targets to control their highly profitable business. The rising levels of violence in Medellín can thus be explained by the individuals involved in the trafficking of coca and taking advantage of the breakdown of law and order, a breakdown which the cartel members had also contributed to (Collier & Hoeffler 1998; Collier 2000; Collier & Hoeffler 2004).

Escobar: Plata o Plomo?

66

By 1985, it was public knowledge that Escobar was the leader or capo of the Medellín cartel. This cartel was hierarchically structured with Escobar supervising the trafficking of cocaine from Colombia to the US. To maintain his control over trafficking routes from Colombia to the US, Escobar operated his business through two methods: silver or cash (plata) and lead or bullet (plomo). A generous bribe (plata) was offered to judges, police officers, journalists, prosecutors or politicians for their assistance. The plata method was particularly successful because during this time, Colombian judges were ‘overworked, underpaid, badly protected, and heavily maligned by the media and other public officials’ (Otera 1991, 963). Escobar took advantage of these easily corruptible public officials and if they were not cooperative they were met with plomo, a rain of bullets (Lamb 2010, 50). Anyone who violated the strict rules of the cartel or any attempt to double-cross them brought an order ‘not just to obliterate the violator, but their entire family - wife, children, brothers, sisters and parents’ (Filippone 1994, 329). Soon, all of Medellín was affected by these rules. The drive-by shootings often resulted in the deaths of innocent bystanders, which would then trigger a cycle of revenge killings by the victims’ friends and families (Lamb 2010, 53).

The perpetrators of plomo were young males from the poor neighbourhoods of Medellín. These individuals worked for the Medellín cartel and were organised into criminal bandas. They admired their boss or their capo, because of the social services he had provided them, such as housing. These young males were also attracted to banda membership because of the economic benefits it provided. This was because in the poorer neighbourhoods during this time, the male son was often the economic provider of the household, as often the family had been abandoned by their father, or he had been killed during La Violencia (Bedoya Marin & Jaramillo Martinez 1991, 76; Salazar 1990, 27). These young males were frequently unable to obtain employment in the formal economy because of the economic crisis, low levels of educational attainment and because young male individuals from the poorer neighbourhoods were largely perceived by the wider society as criminals or potential criminals (PRIMED 1996, 36; Salazar 1990, 33). Thus, young poor males looked to the bandas to provide them with economic resources for their family’s day-to-day survival which they were unable to obtain in the formal economy. All participants interviewed in 2014 and 2015 commented on how young males from these neighbourhoods saw their employment in a banda as one of the only ways to survive in Medellín. For example, a government official (GO3) interviewed in 2014 commented on the economic benefits narco-trafficking offered:

67

‘The narco-traffickers started to give a salary to people which was higher than the formal economy. People from the poorer neighbourhoods started to see narco- trafficking as a solution to their economic problems. There was almost like a competition between the formal economy and narco-trafficking and narco-trafficking won – it offered an alternative economy for these people’.

A community leader (CL1) interviewed in 2015 also commented on how employment in a banda enabled these males to provide the family with household needs:

‘In the poorer neighbourhoods, the mother never asked why her son arrived with a new fridge or a new motorbike or a new television. She never asked, she only said ‘You are such a good son, go and find money and be rich…’’.

This community leader also explained how because of the economic benefits these individuals obtained from their employment they were perceived as someone important in their neighbourhood:

‘All the money from narco-trafficking came to Medellín… in the form of houses, cars, pretty women and parties, of people coming to your party and taking photos… This started to create a certain symbol of status and this symbol of status was transferred to the poorer neighbourhoods. For example, a person from the poorer neighbourhoods would work as a driver or as a bodyguard for a rich person. They would attend their parties and then would bring the leftovers to the poorer neighbourhoods and have their own type of party, but on a smaller scale than the rich person. Just like the narco- trafficking boss looked like a dandy in the media, the poorer person looked like a dandy, but on a smaller scale’.

These young poor males were also attracted to banda membership because of the social benefits membership provided them. For example, a banda provided these young males with membership of a social group and a figure of authority in the banda leader (Bedoya Marin & Jaramillo Martinez 1991, 103). A former government official (FGO4) interviewed in 2014 spoke at length at how a banda provided these individuals with a form of structure which they could not obtain from the wider society:

‘The young people [in the poorer neighbourhoods] of Medellín had a lack of landscape. They had a lack of a future because they were unemployed and uneducated. They did

68

not see a reason for being employed or educated by the formal sector. They did not have a purpose for being in society. As a result of this unemployment, they did not have economic benefits or opportunities in society. This unemployment brought not only economic consequences but also psychological consequences…. There was a growing population of mafia bosses from the narco-traffickers, bosses that needed protection. So, they created bandas that provided bodyguards who were the young people who had a lack of purpose and were a lost population. So, these bandas provided the young people with some form of structure’.

Not only did membership provide these young poor males with a form of structure but also, they were perceived as someone important and prestigious in their neighbourhood and not as a poor criminal as they were seen by the wider society (Bedoya Marin & Jaramillo Martinez 1991, 72). Working as a sicario, was even perceived as the most ‘important achievement’ as the assassin felt as though they were powerful because they were the ‘owners of another person’s life’ (Bedoya Marin & Jaramillo Martinez 1991, 125). Membership therefore provided these poor young males with a form of social mobility which they could not obtain from the wider society. One community leader (CL3) interviewed in 2015 who witnessed the emergence of these groups commented on how these social benefits were attractive to these marginalised individuals:

‘During this time [in this community] whenever you asked a young male what they wanted to be when they grew up, they would say they wanted to be a boss of an armed group. This was because for these young people the boss of an armed group was someone important in the community. They had money, they had a car, a motorbike… Whatever they wanted they got! They also had this sensation of power, that they could do or say whatever they wanted and these were attractive qualities for the other young males in the community...’.

The insights provided by these local actors complement the grievance perspective when explaining the rise of violence, as the young males involved in the criminal bandas were descendants of the families who had arrived in Medellín in the wave of uncontrolled migration. These young males had been neglected by the state and they had taken up arms to redress this neglect (Sambanis 2002; Gurr 1970). Homicidal violence was increasing in Medellín, not only because of the Medellín cartel’s methods, plomo o plata, but also because social cleansing continued against ‘undesirables’, such as drug addicts, prostitutes and homosexuals (Lamb

69

2010, 45; Salazar 1990, 152). By the 1980s, these acts of violence were carried out by the encapuchados in the poorer neighbourhoods and a group called ‘Love for Medellín’ (Amor por Medellín). The former emerged with the support of local authorities and elites, who used the ‘services’ of this group to remove any ‘undesirables’ and maintain order in the city. This practice of social cleansing by vigilante groups showed how state institutions did not have a monopoly over the use of violence. According to one official from a non-government organisation (NGO) interviewed in 2014 (NGO1), it was easy to use the services of vigilante groups to resolve conflicts in the city: ‘In the 80s there was a group called Love for Medellín, how crazy is that name! This group was killing homeless people, prostitutes, and homosexuals. The city was happy for this badly named social cleansing. I mean you could just pay someone for the ‘favour’ of killing the homeless person in front of your house!’.

The Rise of the Militia

The cartel and the social cleansing groups were not the only perpetrators of violence in the city. Another group that used violence to achieve their objectives were the militia groups. The militia groups had been present in Medellín since around 1984. Around this time, President Betancur signed a ceasefire with the left-wing guerrilla group, the M-19. The ensuing peace talks brought ‘Peace Camps’ (Campamentos de Paz) to Medellín. The peace camps were set up in Medellín (and other major cities) as a way to build popular support for M-19 for when they would become a political organisation once the peace agreements concluded (Gutiérrez & Jaramillo 2004, 21). The M-19 organisers were ‘hedging their bets in case the peace talks broke down’ and also ‘showed that the guerrillas could aspire to a more passive presence in the city’ (Lamb 2010, 86). These camps were set up in five neighbourhoods across four comunas in Medellín and they promoted their left-wing political cause, provided social services and helped the communities maintain order (Gutiérrez & Jaramillo 2004, 21). However, just nine months after the peace talks, the peace camps were outlawed and their participants arrested. The army and most police had never supported these peace talks and they had broken down at a national level (Lamb 2010, 86). Not all of the participants were arrested and over the following years, different outlawed militia groups emerged in these neighbourhoods. These groups were broadly distinguished between the Popular Militias of the People and for the People (Las Milicas

70

Populares del Pueblo y para el Pueblo - MPPP) linked to the M-19 and the People’s Militia of the Aburrá Valley (Las Milicas del Valle de Aburrá - MPVA) which were supported by, but not dependent on, the left-wing ELN.

In the neighbourhoods where the militia operated, community members were required to make a payment to the militia, called a vacuna, and in return they were given security services by the militia group. This payment was based on donations that residents gave to the militia when they legally operated in the peace camps. The security services that the militia provided included confronting criminal bandas and other ‘common criminals threatening the security of their communities, such as rapists, drug dealers and bullies’ (Gutiérrez & Jaramillo 2004, 21- 22; Medina Franco 2006, 24). One example of the security services provided by the militia and how they were perceived by the community was given by a former government official interviewed in 2014 (FGO2):

‘Initially the militia were confronted by small disorganised groups of young males, who were normally stoners, rapists or robbers. So, when the militia arrived they started to kill these individuals and the community applauded their actions because they were killing the rapists, the stoners and the robbers. So, the militia are seen as saviours. Because they used this type of local justice, as well as clandestine justice’.

Some of the militia used violence, such as acts of social cleansing, while others used dialogue and persuasion to maintain order in their communities. The militia acted as a type ‘of local congress for those who had the most trivial social or political issues’ and were considered an ‘autonomous form of security’ that had support from the community (Medina Franco 2006, 37- 39; Angarita 2003, 6). They also had a revolutionary ideological platform related to their left- wing affiliations with one militia leader commenting on how their struggle consisted of ‘military undertakings in the night and social work in the daytime’ (Gutiérrez & Jaramillo 2004, 21). State forces had attempted to confront the militia expecting them to be poorly armed and undisciplined. However, the security forces expectations were confounded as not only were they well-armed, well trained, and disciplined but also, they had strong backing from the community because of their provision of services (Gutiérrez & Jaramillo 2004, 22). Certain neighbourhoods preferred the presence and work of these groups to what the state had to offer.

71

The militia groups generally confronted the criminal bandas which threatened the security of their community members. However, Escobar also encouraged the activities of the militia. This encouragement was not because Escobar supported their left-wing political ideology but because they were another force against the Colombian state. Escobar therefore began to ‘funnel arms and other resources’ to them to support his anti-state efforts (Gutiérrez & Jaramillo, 2004, 22). Escobar’s conflict against the Colombian state was motivated by his campaign against the Extradition Treaty and to protect his economic interests. Escobar’s encouragement of the left-wing militia activities was in contrast to his previous dealings with the M-19 when in 1981 they had kidnapped the Ochoa brothers’ sister and MAS (led by Escobar) had retaliated with brutal levels of violence against any leftist sympathisers and insurgents. The relationship between M-19 and Escobar was made public following the attack on the Palace of Justice in Bogota in November 1985. This attack resulted in the destruction of the office that handled extradition, the death of more than 100 people including 25 Supreme Court justices and 35 M-19 members. While the attack had been executed by members of the M-19 it had been partially funded by the Medellín cartel (Lamb 2010, 54). This not only highlighted the power of the Medellín cartel, but also showed how actors apparently pursuing different goals could find common ground to cooperate in violent acts (Jaramillo et al. 1998, 397). Escobar’s support of the M-19 showed the complexity of the situation as the Medellín cartel also continued their use of sicarios against leftist organisations and leftist sympathisers. A former government official (FGO2) interviewed in 2014 recalled how the conflicts between these groups contributed to the rising homicide rates in Medellín:

‘The criminal bandas were defeating the militia groups and vice versa. However, the narco-traffickers were providing resources to both the militia and the criminal bandas. The criminal bandas and the militia were involved in violent conflicts. As a result, there were an appalling amount of deaths’.

The left-wing groups and the narco-traffickers appeared to be winning their conflict against the Colombian state. This was because shortly after the attack on the Palace of Justice, the Colombian Supreme Court declared the legislation of the extradition treaty unconstitutional. The court had declared this because ‘an interim President and not the president of Colombia had signed the legislation into effect’, but they could have also been subject to the intimidation and threats of the cartel (Otera 1991, 970). Almost six months later, in December 1986, the

72 newly elected President Virgilio Barco Vargas brought the treaty back into effect. However, a short time later the Colombian Supreme Court ruled the treaty’s enabling legislation unconstitutional and the President was forced to postpone extradition proceedings indefinitely (Otera 1991, 971). Nevertheless, the Medellín cartel continued its selective and organised assassinations, against anyone critical of their organisation. This use of violence against key opponents was clearly demonstrated with the assassination of the head of the anti-narcotics unit of the National Police of Colombia and the editor of the Colombian newspaper El Espectador in 1986.

Despite the postponement of the extradition treaty by the national government, by 1989, it was back on the political agenda. This was because of the public assassination of the presidential candidate Luis Carlos Galan. This politician had previously prevented Escobar from joining his political party and was well known for open criticism and denouncement of violence and his support of the extradition treaty (Otera 1991, 972). Given this open criticism, Galan was a key target of the Medellín cartel and Escobar was the number one suspect behind ordering the assassination (Lamb 2010, 58). His assassination rebounded on the cartel as the then President Vargas resurrected the extradition treaty through state-of-siege powers in an Emergency Decree. This power allowed the president to extradite cartel members to the US without first obtaining a judge’s signature on the extradition order. President Vargas vowed to ‘drive the drug traffickers from his country’ and within a month the cartel’s accountant was extradited to the US (Otera 1991, 972). However, the ‘cat and mouse’ game continued with the cartel retaliating and assassinating an investigator from the Colombian attorney general’s office, the wife of a police major and the wife of a Colombian intelligence officer (Otera 1991, 973). The Medellín cartel was showing that they would brutally retaliate against any extraditions attempted by the Colombian state.

In Medellín, not all was well with the cartel’s claims to be championing the poor. Despite the redistribution of resources by the cartel, by 1989, the income distribution in Medellín was ‘more unjust in than it was in 1969’ (PRIMED 1996, 35). Throughout the 1980s, the unemployment rate had remained high, ranging from 14 percent to 17 percent (PRIMED 1996, 35). The informal economy was estimated to provide more than 50 percent of employment (Betancur 2007, 2). By 1989, the city was divided more than ever between the informal and formal economy, with the bandas providing a major source of young male employment in the poorer comunas. Two local political events in 1989 brought some change to the situation. The 73 first change was that the informal settlements were recognised by the government with the urban land reform mandating that urban policies needed to address the integration of the ‘subnormal [informal] settlements into the formal city’ (Calderon 2008, 56). The second change was that mayors were to be elected by popular vote for the first time for four-year terms. This political change was a significant departure from the previous 40 years when mayors had been appointed from within their own social and political circles rather than by popular vote. This mode of appointment meant that Medellín had 48 mayors from 1950 to 1990 lasting from one to 18 months and with less or no dependence on the voters (Lamb 2010, 127). These changes were the important first steps in moving away from the clientislitic political structures and the political exclusion of the poorer neighbourhoods in Medellín.

The situation in Medellín had changed dramatically in a short span of time. Not only were there high rates of unemployment, inequality and poverty but there had also been an increase in violence. Homicide rates had dramatically increased from 43 per 100,000 in 1980 to 248 per 100,000 in 1989. This dramatic increase was linked to conflicts between the different non-state armed groups in the city and between them and state authorities. For example, the militia used homicidal violence, in the form of social cleansing against the criminal bandas and other delinquents, to maintain order in their communities. The criminal bandas were employed by the Medellín cartel for selective assassinations against anyone critical of the cartel’s business. For example, in 1989 a presidential candidate, the Commander of the Antioquia police, a Superior Judge, a journalist and the Governor of Antioquia were all assassinated by orders of the cartel. The cartel’s use of violence attracted the attention of the authorities and these authorities used whatever means possible to capture and eliminate members of the cartel. Another conflict had started around the end of the 1980s when a drug cartel based in the Colombian city of Cali, called the , obtained control of the cocaine market in the US cities of New York and Washington. Obtaining this control led to conflicts between the Cali and the Medellín cartels.

Conclusion

This chapter has discussed the emergence of violence in Medellín. The discussion of violence was contextualised in the rapid urbanisation the city experienced and the high levels of inequality and deprivation that ensued. These circumstances created an environment conducive

74 to violence. The chapter also traced the rise of narco-trafficking in the city, focusing on the methods, characteristics and motivations of the Medellín cartel. The methods of this drug cartel led the levels of violence in Medellín to spiral out of control. The chapter also discussed the rise of the left-wing non-state armed group, the militia, and their use of violence in maintaining control in their territories. This chapter demonstrated how complex the situation was in the city and that no single event, actor or situation can be considered in isolation when analysing the emergence of the high levels of violence in Medellín. The data presented in this chapter show the need to take a multifaceted approach in explaining the emergence of violence in Medellín. Examples of the emergence of violence in Medellín were shown to derive from rapid urbanisation which itself gave rise to the emergence of narco-trafficking and the criminal bandas. No one single approach can explain why the levels of violence in the city spiralled out of control and it only by drawing from these multi-layered and inter-related approaches can a holistic explanation of the emergence of violence in Medellín be provided.

The following chapter examines the levels of violence in Medellín in the 1990s. In this period, the levels of homicides reached their peak and Medellín was perceived as the most dangerous city in the world. Throughout this period homicidal violence was the preferred mechanism to maintain and/or obtain territorial control by the different non-state armed actors in the city, including the militia, criminal bandas and right-wing paramilitary groups. In this period, there was increased attention from international, national and local stakeholders to reduce the violence and prevent future outbreaks.

75

Chapter Five: The 1990s: Medellín, ‘the Most Dangerous City in the World’

Introduction

The previous data chapter discussed the emergence of the high levels of violence in Medellín. The discussion was contextualised in terms of rapid urbanisation and the high levels of inequality and deprivation. Evidence was also presented on the growth of narco-trafficking in Medellín and how this growth led to the levels of violence in the city spiralling out of control.

This chapter looks more closely at the rising levels of violence in Medellín in the 1990s. There is a discussion of the multi-pronged approaches used by government to reduce the violence and prevent future outbreaks. The data presented show how the policy makers in Medellín took a multifaceted approach in explaining the emergence of violence, thus justifying the need for their chosen policy approaches. There is also examination of the different non-state armed actors who were active in the city, including the militia, bandas and right-wing paramilitaries, focusing on how they used violence, their motivations and their characteristics.

Medellín: The Battleground of the Medellín Cartel

By 1990, Medellín had changed significantly from the ‘capitalist paradise’ of the 1950s. The city was certainly no longer known for being the ‘most progressive city in Colombia’ but rather it was known for its high levels of violence and poverty. Homicides had become so common that some paisas used the page in the newspaper that reported the death notices as ‘a social page of knowing the whereabouts of people that may have been reported missing’ (Field Diary, May 2014).

This increase in homicidal violence was linked to three conflicts involving non-state armed groups active in the city. The first of these conflicts was between the Medellín cartel and the Colombian state. This cartel was in conflict with the Colombian state over the Extradition Treaty. To force abolition of the treaty the Medellín cartel employed criminal bandas for acts

76 of violence, such as selective assassinations or car bombs, with the objective of assassinating anyone critical of their organisation. Escobar had also employed a new tactic in his war against the Colombian state. Not only would he pay for the assassination of judicial and social leaders critical of his organisation but he also began to pay for the assassination of police officers. This task was taken up with vigour by individuals working in the bandas with 180 police officers assassinated in the city in June 1990 alone (Lamb 2010, 60). During my participant observation and informal conversations paisas commented on how the payment for the assassination of police officers was a turning point in the levels of violence in the city (Field Diary May 2015). This was not only because citizens now realised it was not safe to be close to police officers but also how diabolical the situation in the city had become. During interviews, all participants mentioned this as a significant change in Escobar’s conflict strategy against the Colombian state and one which had a dramatic effect on the larger society. The following comments from former government officials (FGO3, FGO5), an individual from an NGO (NGO1) and a community leader (CL1) reflect Escobar’s war on the police:

‘When I was about 15 years old the boss of our city, Pablo Escobar, would pay a reward of $2000-$3000 for assassinating a police officer, so to be close to a police officer was dangerous’. [FGO3]

‘You would earn about $1000 to $2000 for killing a police officer’. [FGO5]

‘There were around 1 to 2 police officers shot every day and Escobar had paid for this’. [NGO1]

‘In this period Escobar was responsible for the highest criminal rates… Bombs against the police and the general population. He consciously created terrorism; he paid $2000 for killing a member of the police!’. [CL1]

The second reason for the increase in homicides was because of the conflict between the Cali cartel and the Medellín cartel. The Cali cartel had recently obtained control of the New York and Washington cocaine market and taken away Escobar’s monopoly control of the US market. Thus, conflicts increased between these drug cartels. The conflict between these two cartels was battled out in the cities of Medellín and Cali, with car bombs and kidnappings daily occurrences (Riaño Alcalá 2006, 40). By 1990, there were estimates of 150 different bandas operating in Medellín, with 45 linked directly to the Medellín cartel and the remainder indirectly linked, responsible for controlling the drug market in the city (Bedoya Marin &

77

Jaramillo Martinez 1991, 93-97). The local drug market in Medellín was largely comprised of bazuco and these addicts continued to threaten the security of their neighbourhoods. They were also frequently victims of social cleansing ‘campaigns’ by vigilante groups who ‘removed any undesirables’, such as drug addicts and prostitutes, often with the support of the elites and authorities (Salazar & Jaramillo 1992; Bedoya Marin & Jaramillo Martinez 1991, 93-97). While the Cali and Medellín cartels were hierarchically structured, the Cali cartel did not use violence in the way the Medellín cartel did. For example, the Cali cartel did not have a conflict against the Colombian state and concentrated its efforts on maximising profits from their criminal activities (Gutiérrez & Jaramillo 2004, 22). However, the Cali cartel did engage in acts of violence such as kidnapping, physical threats and bribes to intimidate associates, competitors and Colombian government officials (DEA 1994).

The third reason for the increase in homicides was because the state had begun to use repressive force to weaken the structures of the bandas and assassinate anyone suspected of being a member of these groups (Human Rights Watch 1994, 31; Riaño Alcalá 2006, 40). Given that banda members were typically young poor males, the majority of homicide victims were young males, to the extent that the city had been described as ‘in danger of becoming a city of old people’ (Riaño Alcalá 2006, xxxii). The state’s use of violence was noted in June 1990 when a ‘para-police death squad’ assassinated 150 young males in 20 different massacres throughout the city (Riaño Alcalá 2006, 40). The most brazen attack by these state agents was the Villatina massacre in 1992 in the eastern comuna of Medellín. In this attack, eight children and one adult from a Christian group were brutally murdered by the police, with the attack linked to either a case of ‘mistaken identity’ or a revenge attack because the bandas were assassinating police officers (Human Rights Watch 1994, 31). Officials from NGOs (NGO1 & NGO2) interviewed in 2014, commented on how this use of violence by state actors contributed to higher levels of violence rather than reducing them. They also commented on how these killings adversely affected the already shaky trust residents had in state authorities:

‘In the Escobar period, narco-trafficking organisations acted against the state, which was a type of open war with car bombs in public places, assassinations, paying for assassinating police officers etc. However, the state also used groups of sicarios to attack groups of young people aligned with Pablo Escobar and this was a type of dirty war from the public forces. This generated high levels of violence in the city’. [NGO1]

78

‘Around the beginning of the 1990s anytime a police officer was assassinated there were rumours that the police would kill 10 young males. The police created their own type of sicarios. People from the poorer neighbourhoods spoke of the unidentified cars driving around at night. For example, the Villatina massacre, the army assassinated some young kids and captured others and then it turned out that they were not even criminals!’. [NGO2]

The insights gathered from an array of actors show how the homicides in Medellín had spiralled out of control because of conflicts between the Medellín cartel and the Colombian state over the extradition treaty, conflicts between the Medellín cartel and the Cali cartel and the repressive violence used by the Colombian state to weaken the structures of the criminal bandas. These different conflicts show the importance of taking a multifaceted approach in explaining the emergence of violence in Medellín. Some individuals engaged in violence because of their neglect by the state, others because they found they could profit through their control of trafficking coca and taking advantage of the breakdown of law and order. There were also implicit forms of violence, such as the vertical and horizontal inequalities throughout the society. These multi-layered and inter-related approaches provide a holistic explanation of the emergence of violence in Medellín and why the violence was spiralling out of control.

‘The Most Dangerous City in the World’

In 1991, Medellín had a very serious problem with violence. The levels of violence reached their peak and Medellín was portrayed as the ‘most dangerous city in the world’ with over 6,500 homicides that year (Maclean 2014). Internationally the city had become known for its high levels of violence with one example of this noted by the New York Times (1991) running a headline of ‘Medellín Journal, Colombia’s tortured city.’ Medellín’s homicide rate (381 per 100,000) was over four times the national homicide rate of Colombia (90 per 100,000) and almost 40 times the United Nations definition of epidemic violence of 10 per 100,000 (Maclean 2014, 5). This violence was exceeded in the world only by Beirut, Lebanon, which was the scene of a civil war. Not only did the city have these high homicide rates but also homicide rates for children (any individual under the age of 18) had leapt by 566 percent between 1986 to 1991 (from 171 to 1,021 individuals), and up to 17 young poor males were assassinated each day (Human Rights Watch 1994, 30).

79

During this time, homicide data in the city was largely collected by three government institutions: The Colombian National Institute of Medicine: Legal and Forensic Sciences (Legal y Ciencias Forenses Instituto Nacional de Medicina - INML), the Department of Criminological Studies and Identification, (Departamento de Estudios Criminológicos e Identificación - DECYPOL) and Metropolitan Planning (Planeacion Metropolitana). Despite these records, there was not a reliable body of data and there were discrepancies between the statistics established by these institutions (Restrepo et al. 2012, 88). Just one example of these discrepancies was noted in 1991 when Metropolitan Planning recorded 6,595 homicides, whereas DECYPOL recorded 6,349 homicides and INML recorded 6,810 homicides. One official from an NGO (NGO1) interviewed in 2014 even provided a much higher number of homicides from 1991:

‘Look I have heard from researchers from the national government that there is a report, which of course will never be shown to the public, that there were more than 10,000 homicides in that year... They say it was an astounding statistic!’.

Despite the disagreement on the official number of homicides, there was agreement that the high levels of violence had raised levels of insecurity and fear among citizens. For example, during participant observation, many paisas spoke of what these levels of violence meant for their daily live. Some commented of having to walk past up to 30 dead bodies a day, witnessing assassinations in broad daylight on street corners, not being able to socialise in public venues or having their sleep interrupted because of the sounds of car bombs blowing up local police stations (Field Diary, May 2014). For the paisas, the effects of these levels of violence had a dramatic effect on their wellbeing in terms of their livelihoods, feelings of safety and erosion of social capital. For example, an official from a research centre interviewed in 2014 (OF1) spoke of these effects:

‘I remember this time living in the city, the city did not have any life, it was such a cold city… There was no future…’.

A former government official (FGO4) interviewed in 2014 also recalled this effect:

‘Escobar was putting bombs in restaurants, in places where the authorities would go. It didn’t matter if innocent people were killed, as the most important thing was to 80

assassinate a particular person. The bandas also made mistakes, which also meant that innocent people were killed. The city was terrified and you could not go out at night’.

This former government official (FGO4) had lived in the city during this time and recalled how the levels of violence became almost normalised and that citizens accepted that nothing would be able to change the situation:

‘I think that a culture of violence was created, which sounds a little dramatic, but well, the society started to accept that there are deaths, disappearances, tortures, rapes. It started to be daily news and something normal, an everyday thing… People would take precautions; you would know you could not go to certain sectors of the city… Unfortunately, there was not really a fight against the violence from the authorities... We just accepted that this was our destiny, it was part of our life.’

The Presidential Council: A Strategy to Address the Violence

On a national level, in 1991 President Gaviria declared that the violence in Medellín was ‘the country’s gravest problem.’ A Presidential Council for Medellín and the Metropolitan Area (Consejería para el Área Metropolitana de Medellín) was formed to reduce the high levels of violence and prevent future outbreaks. Given the international scale of narco-trafficking, funding for this Council was obtained from both national and international sources (Maclean 2014, 37). An important role of this Council was gaining the perspectives of non-political actors about why the violence had emerged, who were the perpetrators of violence and what they thought were the best solutions to reduce or prevent future levels of violence. Thus, numerous seminars were held over the following years and were attended by a wide range of stakeholders including community leaders, business elites, national and local government representatives and academics to provide solutions to the rising levels of violence. During data collection, all interviewees commented on the importance of this Council as the first step in addressing the high levels of violence in Medellín. For example, a former government official (FGO4) interviewed in 2014 commented on how it was the first time that the high levels of violence had been acknowledged by the national and local government:

81

‘I think when this Council started it was when almost as if a conscience was created about confronting these very complicated problems. We were very lucky that we had some very sensible and aware municipal administrations in this time period’.

A community leader (CL1) interviewed in 2015 who was initially involved in the Presidential Council also recalled how it was an important step in acknowledging the high levels of violence in Medellín. When speaking about the Presidential Council, this leader highlighted the importance of involving different stakeholders in understanding the urgency of the problem:

‘It was an important time when rich and poor people, whites and blacks, academics, business leaders and community groups all came together. We all sat down at the same table and realised we were all scared to death and that this was the only place where we could meet and be responsible and make some agreements… However, as these seminars were going on, the violence continued in the rural areas of the country and more displaced people were arriving in the city. So, there was no time for planning, it was more about how to improve things’.

Not only did participants comment on the importance of the Council in finding solutions to the violence but also community leaders and NGO personnel interviewed for this study in 2014 and 2015 remarked on the significance of these seminars as it was the first-time individuals from the marginalised communities were provided with a voice. These groups had been excluded from the policy debate given the history of clientelistic political patterns in Medellín, and Colombia. For example, a former government official (FGO3) interviewed in 2015 spoke about the involvement of the different political, social and academic leaders who were invited to contribute to the development of solutions to the violence in the city:

‘The people involved in these seminars had spent many years working to help young people escape the violence. Individuals like the future mayor Alonso Salazar who wrote the famous book No Nacimos Pa Semilla [Born to Die in Medellín] and Victor Gaviria who made the famous movie Rodrigo D no Futuro [Rodrigo of no Future], public health experts like Luis Fernando Duque from the University of Antioquia and academics like Hector Gomez, Hernan Henao, who were also human rights leaders and were eventually assassinated. All of these people had written and thought about the violence in the city and they had an opportunity to express their views in these seminars. It presented a space for these different social, cultural and academic groups to discuss their ideas’.

82

A community leader (CL1) interviewed in 2015 recalled how during the seminars he had the opportunity to speak to the business elites about the problems in the poorer neighbourhoods of Medellín. The community leader recalled what he said to the business elites:

‘We [the community groups] knew firstly there had to be more employment.. We said to the business elites, ‘you generate employment; you are the owners of wealth in the city. Wealth has to be redistributed, salaries have to be improved, life conditions have to improve, and we need more education and health’’.

The most significant findings from these seminars were outlined in the report, Medellín: Re- encounter with the future (Medellín Re-encuentro con el futuro). Local academics, local NGOs and the UNDP all contributed to the findings of this report. The report emphasised how the high levels of violence in the city were related to what they called ‘structural reasons’ that is, inequalities, high levels of unemployment, low education, poor health, the increased incidence of informal housing and a significant lack of public space (Presidencia de la Republica 1991). These reasons were all related to the rapid urbanisation the city experienced from the 1950s and largely based on how individuals engaged in violence as a mechanism of redressing their neglect by the state. The report also emphasised how the growth of narco-trafficking had caused the levels of violence to spiral out of control. Individuals involved in narco-trafficking had taken advantage of the breakdown in law and order, a breakdown to which they had greatly contributed. These individuals used organised and selective violence to protect their economic interests. There was no one reason or causal factor identified in the report which could explain the emergence of high levels of violence in the city. There were multiple causes which necessitated the development of a multi-pronged approach.

The stakeholders looked to develop policy solutions to address the ‘structural reasons’ for the high levels of violence. All participants interviewed for this study acknowledged that there is no one reason or cause which explained the emergence of violence in the city and thus they needed to develop multi-pronged solutions to address the violence. For example, one former government official (FGO4) commented on the complexity of the situation and the need to develop many solutions:

‘They looked to find a strategy to address the climate of violence in Medellín. This strategy had many facets. They started strengthening the police force, improving the

83

justice system, increasing programs for young people. But they also looked at the large inequalities in the city. There were investments in education, health, housing, neighbourhoods. The violence in Medellín is a very complex problem which required simultaneous approaches’.

One of these policy solutions was the program called ‘Subnormal District Improvement in Medellín Program’ (Programa Integral de Mejoramiento de Barrios Subnormales en Medellín - PRIMED). The stakeholders had obtained international funding from the Kreditanstalt für Wiederaufbau (Reconstruction Credit Institute), the German Development Bank, to launch this program. The objective was to improve public services and public spaces, upgrade housing, secure land tenure for residents in informal settlements and other issues identified by the stakeholders involved in the Presidential Council regarding the state neglect of these neighbourhoods. Given the international funding of the program, it was run independently and, therefore, avoided the largely clientilistic political structures in the city. PRIMED also focused on ensuring the community was engaged with any project and encouraged state-civil society partnerships (PRIMED 1996; Bahl 2012). There was an assumption by the policy makers that by modernising the essential infrastructure and improving the housing in the poorer neighbourhoods they could reduce the violence and prevent future outbreaks. This assumption reflects the social disorganisation approach of Shaw & McKay (1942). For the paisas from these poor neighbourhoods, it was the first time any state investment was made in their communities. Any previous state presence in these neighbourhoods was repressive, in the form of extra-judicial killings or forcible removal by the state.

Not only were there social development projects to address the high levels of violence, but also the Colombian government had begun negotiations with the international drug baron, Pablo Escobar. Negotiations had started between the national government and Escobar, with President Gaviria, arguing to the Colombian population that Escobar’s surrender would bring an end to the organised and selective assassinations (Lamb 2010, 62). Escobar was interested in negotiating, as reforms to the Constitution in June 1991 had placed a ban on extradition and he also saw prison as a place to be safe from his Cali cartel rivals. A condition of Escobar’s surrender was that he could choose his own prison, and he selected a former drug rehabilitation centre just outside of Medellín called The Cathedral (La Catedral). Escobar moved in with some associates and was now safe from his rivals, the Cali cartel and Colombian security forces. The Cathedral, however, was not like any other prison in Colombia and there were rumours it ‘resembled Colombia’s best five-star hotels’ (Lamb 2010, 62). Furthermore, just

84 because Escobar was removed from society did not mean he would stop working as a narco- trafficker and he continued to collect ‘taxes’ from other narco-traffickers amounting to an estimated $US100 million a month (Lamb 2010, 63).

For a short time, there was a decline in car bombs and kidnappings in Medellín. However, just shy of Escobar’s one-year anniversary in the Cathedral, local authorities found some tortured corpses within the surrounds of the prison. It was soon revealed these corpses belonged to former members of the cartel, who had been suspected of stealing from the cartel and had been violently assassinated (Lamb 2010, 62). The national government was made aware of this discovery and could no longer justify keeping Escobar in this ‘prison.’ The national government decided to move Escobar to a higher security prison in Bogota. However, Escobar became aware of the plan and escaped (Lamb 2010, 63). The re-capture of Escobar, dead or alive, was now the Colombian government’s number one priority. International assistance from the DEA and CIA was provided to the Colombian authorities and the Search Block (Bloque de Búsqueda) was formed. To assist in capturing Escobar, the increase in the numbers of state officials throughout Medellín provided a more visible state presence than in previous years.

However, the Search Block’s objective would not be achieved so easily. This was because Escobar recruited more young men into his ‘army’ and started a new campaign of violence. Escobar increased his payments for killing state officials; with rumours that he was offering up to $27,000 per member of the Search Block assassinated (Lamb 2010, 69). This significant payment resulted in 620 police officers assassinated in Medellín in 1992 (Lamb 2010, 60; Human Rights Watch 1994, 31). The Search Block was also confronted with another problem. The use of repressive force by the state in previous years meant that many paisas from the poorer comunas did not trust their actions. Many young males continued to prefer the more ‘localised power’ from associations with the Medellín cartel, which provided them with social and economic resources which the larger society did not offer them.

Despite the support Escobar had in the poorer comunas, his use of brutal violence, such as torturing and assassinating members suspected for stealing from the cartel, had meant he had a growing list of enemies. Some cartel individuals also did not support Escobar’s open war against the Colombian state as it put pressure on the narco-traffickers from the authorities which ultimately reduced the profits that could be obtained from narco-trafficking. Thus, Escobar’s enemies included members from the Cali cartel and former members of his own cartel, such as an individual called Diego Fernando Murillo, who used the alias of Don Berna.

85

These enemies obtained funding from some industrialists, politicians and ranchers to form the vigilante group, People Persecuted by Pablo Escobar (Perseguidos por Pablo Escobar - PEPES). This group was led by the leader of the largest paramilitary group, the Autodefensas Campesinas de Cordoba y Urabá (Peasant Self-Defence Forces of Cordoba and Urabá - ACCU), Fidel Castaño, who had previous dealings with the Medellín cartel (Semana 1994; Maclean 2014, 25). While the PEPES was comprised of vigilante groups, they also ‘moonlighted’ for state authorities, as the latter wanted Escobar dead and by using the services of the PEPES this could be achieved more quickly (Lamb 2010, 68). Using the state’s significant financial and military support, PEPES worked to ‘annihilate Escobar’ (Lamb 2010, 66). They violently and publicly assassinated anyone associated with Escobar to ‘maximise the psychological effect’ and over the following year, more than 125 people associated with Escobar had been violently killed (Salazar 1990, 150; Lamb 2010, 67).

Escobar was killed in Medellín on 3 December 1993 by members of the Search Block. His death was perceived as ‘a crucial step for the end of narco-trafficking in Colombia’ (President Gaviria in El Pais 1993). There was hope that with the removal of this drug baron, who had paid for the assassination of public authorities and deliberately created a climate of terrorism that the levels of violence would decrease. However, the problems of violence in the city would not be resolved so quickly. This was because the young males who had previously worked for Escobar in the criminal bandas were now left without an employer. The banda lifestyle had become part of their socialisation and existence, a model of upward mobility and a way in which they could earn easy money (Melguizo & Cronshaw 2001, 115). Many of these poor young males had access to weapons and extensive knowledge of using them and wanted to avenge the death of their fathers or friends by either state or non-state groups (Lamb 2010, 71).

Changes began to occur in the patterns of violence in Medellín as various would-be leaders jockeyed for power over non-state armed groups. Of these aspiring leaders was one former Medellín cartel and PEPES member, Don Berna, who had been watching the situation very carefully (Gutiérrez & Jaramillo 2004, 27; Insight Crime 2015a; Hylton 2007, 83). All participants recalled how following Escobar’s removal, Don Berna played a significant role in the re-structuring of organised crime in the city. For example, a former government official (FGO4) interviewed in 2014, who commented on how Escobar’s death was not the ‘crucial step for the end of narco-trafficking’ authorities had hoped for, recalled how the vacuum created by Escobar’s death was filled by Don Berna:

86

‘I am not sure if you can necessarily say that when Escobar died Don Berna took his place. It is similar to ‘The king is dead, long live the King’. Just like how the guerrillas say that from every guerrilla assassinated, a rose grows. One person dies and the space is just occupied by the same people’.

While Don Berna did occupy the same space as Escobar, he operated his business in a different way. Don Berna had inherited a network of specialist killers (sicarios) from Escobar called the Oficina de Envigado (Hylton 2007, 84). In contrast to the hierarchically structured Medellín cartel, the Oficina de Envigado was a network or a ‘clearinghouse’ for the Colombian criminal underworld which connected customers to contractors and charged a ‘finder’s fee’ (Lamb 2010, 168). Don Berna made the decision to keep this network structure as it would not draw the attention of the authorities in the way the hierarchically structured Medellín cartel did (Gutiérrez & Jaramillo 2004, 27; Lamb 2010, 168). Don Berna had also learnt from the mistakes of Escobar of having an open war against the Colombian state. An NGO worker interviewed in 2014 (NGO2) recalled how the criminal actors changed how they used violence to obtain territorial control. These criminal actors acknowledged the need to act in less visible networks to ensure their economic interests were protected:

‘There is an undisputed fact that from 1994 until today [2014], no narco-trafficker confronts the state.. The narco-trafficking groups have understood, by taking almost a rational and intelligent decision, that confronting the state is a lost war and it is better to collaborate. As a result, there are almost kinship-like relationships between criminal organisations and sectors of the state, almost like invisible government pacts, with the logic of ‘you do not come after us and we will not use obvious violence.’ This was a fundamental qualitative change in 1994’.

The Militia: Semi-criminal and Semi-legal Organisations

Another reason why the problems of violence in the city were not going to be resolved quickly was because the militia continued to pose a threat to the state’s monopoly on the use of violence as the militia used violence in maintaining order in their communities. By 1991, local sources had estimated the militia groups had up to 8,000 members (El Tiempo 1991). The militia continued to use violence, such as social cleansing against delinquents and drug dealers, to maintain control in their territories (Melguizo & Cronshaw 2001, 122; Gutiérrez & Jaramillo

87

2004, 22; El Tiempo 1991). For the communities where the militia were present, they were perceived as the legitimate authority, as they regulated social action and provided a sense of a governing body (Estrada & Gómez 1992, 67; Lamb 2010, 96; Arco Rivas 2005, 20-23). These groups provided communities with a means to resolve local tensions and grievances that the state was unable or unwilling to provide. Furthermore, residents of these communities were often victims of the extra-judicial by state forces, which created a type of ‘boomerang effect’, as the ineffective official governance in the communities contributed to the rise of these militia groups and a perceived need to rely on them to regulate social action.

The militia were loosely divided between the Popular Militias of the People and for the People (Las Militias Populares del Pueblo y para el Pueblo - MPPP), linked to the M-19 guerrillas predominantly in the north-east comunas and the People’s Militia of the Aburrá Valley (Milicia Populares del Valle de Aburrá - MPVA) supported by, but not dependent on, the ELN, predominantly in the north-east and east-central comunas. From the perspective of residents in these neighbourhoods, the main distinction between the different militia groups was how they dealt with criminals. For example, the MPPP would generally use violence against criminals whereas the MPVA claimed they acted more as ‘peace negotiators’ with criminals (Lamb 2010, 94). However, outsiders, such as government officials and the media, did not see the same distinction and the militia were perceived as one group.

Of concern to local authorities was expansion of the militia following the death of Escobar. One interpretation for this expansion was that, following the death of Escobar and the dismantling of the Medellín cartel, many young males were left without an employer or a social structure. They joined the militia as a mechanism for redressing their social and economic grievances and not because of political reasons. For example, many joined the militia as a mechanism for ‘avenging the deaths of their friends or relatives’ by state or non-state groups (such as other bandas) (Lamb 2010, 146). Given the previous work experience of these ‘new militia’ members in the criminal bandas, they were more violent than the ‘old’ militia members and the community perceived them as ‘more like gangsters than leftist militants’ (Lamb 2010, 146).

Around this time there were also changes in militia practices. This was because some militia members perceived the vacuna payments as being inadequate compensation for the services they provided to their communities. Thus, some militia members began to seek donations more 88

‘coercively’ but others also ‘rationalised’ their involvement in narco-trafficking in terms of using the profits to ‘give back’ to the community in the form of social investment projects (Lamb 2010, 114). Some of the militia members had previously been involved with narco- trafficking when Escobar had funnelled arms and other resources to them in his war against the Colombian state (Gutiérrez & Jaramillo 2004, 22). However, their involvement in narco- trafficking now varied from hiring out their security services to providing strategic planning for narco-trafficking (Barrero Tapias 2001, 217). This increased involvement in narco- trafficking led to two changes within the militia groups. The first change was that many members were now motivated for economic interests, rather than political ones. The second change was that conflicts increased between the different militia groups. This was not only because not all members supported the involvement in narco-trafficking but also because obtaining and maintaining territorial control for economic reasons was now of great importance to them (Barrero Tapias 2001, 271; Lamb 2010, 114; Gutiérrez & Jaramillo 2004, 22). The situation was becoming more complex as it was not as easy to distinguish the difference between the militia and criminal bandas because of these ‘blurred practices’.

From 1991, some of the militia had engaged in informal negotiations with local authorities. These led to formal negotiations starting between the national government and the MPPP, the MPVA and a dissident faction of the latter, the Independent Militia of the Valley of Aburrá (Milicia Independientes del Valle de Aburrá) in 1994. There were other militia groups in the city at this time, but they rejected any possibility of negotiation. The local government acted as a ‘neutral third party’ in these negotiations (Gutiérrez & Jaramillo 2004, 23). The national government was particularly interested in this opportunity to demonstrate its commitment to peace, especially given the recent national failed negotiations with the FARC (Gutiérrez & Jaramillo 2004, 23). The Presidential Council was also interested in driving the negotiations as they continued working to address the still high levels of violence in the city. There had been negative repercussions of the trust residents had of the police following the high degree of force used to capture Escobar and this was an opportunity for all to demonstrate their commitment to peace. It was seen as a unique event in Latin America, as it was one of the first times a government in the region had negotiated with an urban militia group (Rozema 2008, 434; Melguizo & Cronshaw 2001, 123). The militia came to the negotiation table hoping to obtain some political recognition and ‘some advantages to consolidate their areas of influence’ (Gutiérrez & Jaramillo 2004, 23). Despite the best intentions of both parties the negotiations were plagued with problems before they even began. These problems were linked to the 89 significant internal divisions, internal corruption and leadership problems in the different militia groups. The militia leaders also lacked the technical capacity to negotiate with the state, especially given their poor socio-economic backgrounds and low levels of education (Melguizo & Cronshaw 2001, 123; Gutiérrez & Jaramillo 2004, 23).

Following months of complicated negotiations, the Media Luna Agreement was signed on 26 May 1994. The agreement proposed the demobilisation of 600 individuals from the militia, state social investment in some neighbourhoods, and some political privileges for the militia and judical benefits in the form of an end to any criminal procedures if they confessed to their crimes (Gutiérrez & Jaramillo 2004, 24). A security and community service cooperative (Cooperative de Seguridad y Servicios a la Comunidad - Coosercom) was also created which required the demobilised members to work with state security services in crime prevention. These were all difficult proposals, which despite the best intentions of both parties, would be difficult to implement and evaluate.

The first problems started within three months of the end of the negotiations. One of the militia leaders, Pablo Garcia, was violently assassinated. Investigations by local authorities revealed that a leader from another militia group had assassinated him. This assassination led to vengeance killings between the different militia groups (Gutiérrez & Jaramillo 2004, 23). Also, the demobilised militia were not viewed favourably by militia groups outside of the Media Luna Agreement and were seen as ‘traitors’ because of their negotiations with the national government. The group Death to the Demobilised Militia (Muerte a Milicianos Desmovilizados) emerged with the specific aim of assassinating demobilised militia (Human Rights Watch 1994, 48). The levels of vengeance killings were the highest number that have been seen in any reinsertion process in Colombia (up until 2013), with 22.2 percent of demobilised militia violently killed (Giraldo & Mesa 2013, 225). For example, in the first month following the negotiations, homicide rates had dropped by 23 percent; but they increased by 47 percent in the second month. One interpretation for this increase was because when the militia groups physically vacated their communities for the negotiations, there was no one left to maintain control in their territories or provide any type of social control (Gutiérrez & Jaramillo 2004, 23). Just as the defeat of Escobar did not mean the neutralisation of banda activity, the Media Luna Agreement did not stop or neutralise the militia or its armed opponents from engaging in violent acts (Melguizo & Cronshaw 2001, 124). Within two years, many of the militia leaders were assassinated because of vengeance killings and only scattered militia

90 groups remained, with most losing their traditional ‘revolutionary’ left-wing political motivations.

The increase in homicides was not the only problem following the negotiations. The security and community cooperative created after the Media Luna Agreement, Coosercom, also presented significant challenges. This was because the cooperative was highly criticised as participants were never given clear instructions on how to change from being a member of a ‘semi-criminal and semi-social organisation’ to becoming an appendix of the police (Giraldo & Mesa 2013, 236). Given their lack of training and supervision, there were numerous reports of Coosercom members committing abuses against civilians and accusations of homicides and threats, and many members became involved in criminal activities (Gutiérrez & Jaramillo 2004, 23). The cooperative was dismantled around the middle of 1996. However, by creating this co-operative the state had facilitated the use of violence to resolve tensions and grievances and demonstrated that the state did not have the control of legitimate violence in the city.

Peaceful Co-existence Pacts with the Bandas

By 1996, local government sources estimated there were over 180 different bandas predominately working with the Oficina de Envigado (an increase from 150 in 1990) and 19 separate militia groups (Melguizo & Cronshaw 2001, 130). However, these figures were not reliable, given the Oficina de Envigado was purposefully organised as a ‘network of contractors’ rather than a hierarchically structured organisation like the Medellín cartel and it acted in ways which avoided drawing the attention of the authorities. Another reason why the figures were not reliable was because of the ‘blurred practices’ between the militia and the bandas. For example, some militia groups had become involved in narco-trafficking and some bandas were providing their community with services such as ‘taking drunks home and resolving family conflicts.’ These bandas had learnt from the militia that to maintain order in their communities, they needed to have some loyalty from community members, which they could not obtain by just using violence (Gutiérrez & Jaramillo 2004, 25; Melguizo & Cronshaw 2001, 125). Therefore, the line between the militia groups and the bandas was not as clearly defined as before. From a government perspective, some groups were still labelled militia, despite them not having any left-wing political motivations. However, from a community perspective, it did not appear to matter if their territory was controlled by the bandas or the militia, as long as a group provided a sense of security and a constructive communitarian set 91 of activities (Gutiérrez & Jaramillo 2004, 25). Therefore, even though these non-state armed groups were not seeking to overthrow the government, they presented a significant challenge to the government, as they were perceived by the community as part of their environment or legitimate actors.

Violent conflicts were common between the bandas and militia. These conflicts concerned a range of issues such as, territorial disputes relating to controlling the local crime rackets and drug market, schoolyard fights escalating to neighbourhood conflicts, revenge attacks or even different political ideologies (Baird 2011, 97; Insuasty et al. 2010, 54). Despite the different reasons for the conflicts, homicidal violence was the preferred mechanism for resolving them. To protect the territory of each group, invisible borders (fronteras invisibles) were created throughout the poorer comunas (Lamb 2010, 144). These borders marked out the territory controlled by the different groups and were patrolled by armed individuals. They enforced the death penalty for any individual, even if they were not a member of an armed group, crossing into another group’s territory (Lamb 2010, 144). These borders resulted in communities becoming extremely divided with high levels of distrust between them, with anyone from a different community assumed to be ‘an enemy of the other’ (Lamb 2010, 145). During my participant observation, community members commented on how these borders affected their wellbeing and social mobility, as they were often not allowed to travel across ‘borders’ to attend schools, their workplaces or hospitals. One participant even told the story of how her infant daughter died of an asthma attack because the border-control police would not let her leave or let medical attention enter the area (Field Diary, July 2015). These invisible borders not only affected residents’ wellbeing but also had an impact on personal networks and relationships, as often an individual’s network did not extend outside their immediate community. This lack of networks across the city created further reliance on the group protecting the territory as a coping mechanism for community members to deal with the effects of the high rates of violence and distrust.

There were some short periods of ‘peace’ when free movement was allowed between the neighbourhoods and shooting ceased between the different groups. These ‘peace periods’ were because of ‘peaceful coexistence pacts’ (pactos de conviviencia). A Catholic priest living in one of the poorer comunas had initiated these pacts. This priest stressed to the bandas the importance of ‘ceasing hostilities and ending violence’ for the good of the entire community (Rozema 2008, 435). Following the initial pact, more local priests in the poorer

92 neighbourhoods began to negotiate with the bandas in their neighbourhoods. These pacts were never written down, as they respected the ‘internal laws’ of the banda and thus, they relied on ‘word of mouth’ (Rozema 2008, 436). Some pacts banned public drug consumption, others provided rules against personal conflicts turning into wars and some prohibited firing guns on public holidays. Local sources estimated there were 57 pacts in 71 different neighbourhoods between 1995 and 1999 (Rozema 2008, 436).

By the end of the 1990s, the local government had taken notice of the success of these pacts. The local government had recently created the Office for Peace and Co-Existence (Oficina de Paz y Convivencia) with the objective of taking part in these local negotiations and providing some financial support for the processes (Rozema 2008, 436). Some small-scale investments were made in terms of human rights workshops and training to run small businesses, and there were even promises to people they would be able to enter the workforce if they improved their behaviour, despite not having the qualifications and experience to obtain employment (libreta militara) (Rozema 2008, 436). The Office for Peace and Co-Existence had the objective of working specifically with the bandas, but because of the increasing ‘blurred practices’ between the militia and the bandas, some militia groups could have also benefited from this support.

There were some successes with the pacts and many lives were saved. For example, homicides reduced from 209 per 100,000 in 1996 to 184 per 100,000 in 1997. However, these pacts did not offer long-term solutions to the targeted communities. The lack of funding from the local government also meant that many promises were not followed through (Melguizo & Cronshaw 2001, 112; Rozema 2008, 436). Furthermore, the Office for Peace and Co-Existence never had a political or legal mandate to negotiate with these bandas. However, by recognising and negotiating with these groups, the state acknowledged that its organisations were not the only ‘legitimate locus of power’ in the city and that the norms and rules in the poorer communities were set by non-state actors and not by the government (Melguizo & Cronshaw 2001, 127; Velez 2001, 70). The local government’s efforts to make and maintain pacts were inadequate to address the larger scale of the problem and violent conflicts started again between the different groups. Many groups were never disarmed as part of the ‘peace process’ (Rozema 2008, 436). Just as the Media Luna Agreement did not stop or neutralise the militia, these pacts did not stop or neutralise the bandas (Melguizo & Cronshaw 2001, 124). It showed yet again how the state was sharing control over citizens and communities with these non-state groups.

93

At least the Office of Peace and Co-Existence had provided some form of support to the poorer neighbourhoods, while the social development policies from the Presidential Council were also doing the same. The internationally funded PRIMED program focused on improving the built environment in the poorer neighbourhoods and addressing the levels of inequality. The program had many successes which were published in the report ‘PRIMED: A Successful Experience in Urban Intervention’ (1996). PRIMED was listed as urban development ‘best practice’ by UN-Habitat with estimates it had benefitted around 51,000 people living in the informal settlements (PRIMED 1996).

Despite initial support for the nationally led Presidential Council, difficulties emerged in the relationship with the local government concerning the implementation of the council’s policy programs. This was because the local government wanted control over activities in its territory and used recent reforms of the National Constitution (1991) and local government to justify their claims. The local government did not wish to be ‘dictated to’ by national policies as to how they should address levels of violence in the city (Moncada 2013, 15). Furthermore, even though the original intentions of the national local government were about promoting the involvement of the community and other non-political stakeholders, the local government, along with business leaders, used their preferential access to local media to discredit groups working with the Presidential Council, even labelling some of them criminal groups. They did this because the views of these community organisations did not align with those of the local government (Moncada 2013, 15). A community leader (CL1) interviewed in 2015 who had been initially supportive of the Presidential Council recalled how the relationship between community groups and the government had become strained. This leader said:

‘… The groups that made the first agreements [in the Presidential Council] were the groups which should not have really been there. There were other groups that wanted to be in the agreements, but well, you should ask why they were not invited! It turned out to be a very difficult negotiation [between the stakeholders involved in the Council]’.

The different positions of stakeholders also reflected the continuing political clientelistic relations in city. Despite the central and local governments’ efforts to reduce urban violence and improve the welfare of poor communities, these neighbourhoods continued to be politically, economically and socially excluded from the larger society. Whilst some state

94 services were now present in these neighbourhoods, they were generally perceived as weak and ineffective. Thus, given these difficulties Restrepo et al. (2012) argued that the 1990s was a period largely characterised as one of ‘fragmented implementation’ of policy.

The Paramilitaries Enter Medellín

By 1997, a solution for reducing the violence and preventing future outbreaks in Medellín had not yet been found. The non-state armed groups were now largely comprised of the left-wing militia and the criminal bandas. The majority of the criminal bandas worked for Don Berna as ‘contractors’ for the Oficina de Envigado and they were in conflict with the militia groups over obtaining control of territories. Both the militia and the bandas used invisible borders to mark out the territory between themselves and different groups and enforced the death penalty for any individual crossing into another group’s territory. While both the militia and the bandas used homicidal violence to maintain control in their territories they also provided their communities with services, such as conflict resolution or social cleansing to maintain this control. The levels of violence in the city were however about to change dramatically with the arrival of the right-wing paramilitary groups.

The rise of paramilitary groups in Colombia has a complex history and the following will provide an overview to understand how their presence contributed to patterns of violence in the city. On a national level, paramilitaries formed in the 1960s. At this time, the Colombian military did not have the resources to counter left-wing guerrillas, so the military forces recruited civilians in the guerrilla-controlled areas to support the military (Lamb 2010, 172). The use of these civilian recruits was legalised in 1968 and lasted for 20 years. However, in response to fears of ‘narco-terrorism’ and the assassinations of presidential candidates by sicarios, who had links to paramilitaries, the setting up and use of paramilitaries had been outlawed. However, in 1994, the Special Services Surveillance and Private Security (Servicios Especials de Vigilancia y Seguirdad Privada - CONVIVIR) was created in 1994 in response to growing guerrilla activity (Lamb 2010, 172). CONVIVIR was created to provide legal paramilitary structures to maintain control over areas from where the guerrilla forces had been expelled or which could face future attacks (Amnesty International 2005, 9; Lamb 2010, 173). However, the organisation was poorly regulated and operated more as a ‘paramilitary death squad’, with a Colombian Army Officer even noting in 1997 that they had become ‘very

95 difficult to control’ (Lamb 2010, 175). Government sources were unable to advise how many CONVIVIR groups existed throughout the country (Lamb 2010, 174). By this time, CONVIVIR had also become corrupted with links to narco-trafficking and criminal organisations (Semana 1997). The national government reacted to these issues by clarifying the standards of behaviour of CONVIVIR and suspending licences for any individual who violated the new standards. Following this regulation, many groups lost their licences and the organisation was officially dismantled (Lamb 2010, 175).

Many of the previous members of CONVIVIR were however simply integrated into the recently formed United Self-Defence Forces of Colombia (Autodefensas Unidad de Colombia - AUC). This organisation united most of the until then disunited paramilitary groups in the country and was commanded by the Castaño brothers (Vicente, Fidel and Carlos). These brothers were farmers whose father had been assassinated by the FARC in the 1980s. Carlos Castaño had previously formed the Peasant Self-Defence Forces of Cordoba and Urabá (Autodefensas Campesinas de Córdoba y Urabá) (ACCU) in the late 1980s, which had been a group well-known for their violence, such as leaving behind beheaded or disembowelled bodies as a warning sign to any leftist sympathisers (Lamb 2010, 171). Fidel had worked with Pablo Escobar in the Medellín cartel, but when Escobar threatened to assassinate him, Fidel played a role in the PEPES, the vigilante group which had the number one objective of annihilating Escobar.

The AUC was comprised of different but isolated groups or blocks throughout the country which worked together for two general objectives. Their first objective was their provision of a ‘civil-defence service’ against the left-wing guerrillas, a service the national government welcomed. Their second objective was providing security services for coca leaf farmers and cocaine manufacturers (Lamb 2010, 179). The latter was particularly important as while rural farmers had initially funded this right-wing group, by the late 1990s, Carlos Castaño claimed that ‘70 percent of the AUC’s revenue came from narco-trafficking’ (Semana 2008; Amnesty International 2005, 8). The large scale of their involvement in narco-trafficking was noted in May 1999, when the Colombian National Police seized a large AUC-run drug lab in the Middle Magdalena area of North Colombia (El Tiempo 1999a). While the paramilitaries were initially formed because of political ideologies, they were now increasingly concerned with generating income opportunities from the production and distribution of cocaine from the rural production of the coca leaf. The AUC’s involvement in narco-trafficking meant that the city of Medellín 96 was crucial, as its geographical location meant that it provided corridors to all parts of the country, crucial for the trafficking of illegal narcotics from Colombia to the US.

In the late 1990s, the local media and government reported how the AUC entered Medellín under the rural paramilitary unit called Metro Block (Bloque Metro) (BM), led by a man who used the alias Doble Cero (Verdad Abierta 2016). Doble Cero adhered to the right-wing ideology of the paramilitaries and did not allow narco-trafficking to corrupt his work. He had previously worked in the Colombian Army and spent many years in the right-wing ACCU. The BM initially targeted Medellín communities where residents ‘complained’ about abuses from the existing urban militia (Lamb 2010, 183). By this time the urban militia in Medellín had links to the FARC and the ELN. There was also a new group called the Peoples Armed Commandos (Comandos Armados del Pueblo) (CAP). The latter had been founded around 1996 from dissident factions of the ELN and MPP militia following the failure of Coosercom and the Media Luna Agreement (Lamb 2010, 188). In contrast to the ‘old’ members who were seen as saviours the ‘new’ militia members were more violent (Lamb 2010, 146). The BM thus began to offer security services to these communities, such as engaging in ‘bloody social cleansing campaigns’ to maintain order in these communities. The BM justified their presence and their use of violence on the grounds that ‘the citizenry was fed up with delinquency and asked us to come’ (El Colombiano 2002a). For example, a former government official (FGO2) who had good knowledge of the paramilitaries in the city interviewed in 2014 confirmed how the paramilitaries initially provided these services when they arrived in a particular urban community:

‘They [the BM] defeat the urban militia and take over the neighbourhoods. They socially cleanse the urban militia groups and become an established group…’.

As indicated by this official, by providing these services the paramilitaries were able to maintain order in a community. This presented a turning point in the history of Medellín, as the right-wing paramilitary group had provided some security and vigilante justice to the communities when they entered Medellín in the late 1990s. These types of services had won the paramilitaries some support or acceptance from the community. The Medellín communities had little concern with the political motivations of the paramilitaries or their excessive use of violence in the rural areas. Although affected urban communities may have also have feared the paramilitaries rather than support them, the communities were nevertheless interested in

97 the services which the paramilitaries provided and that state authorities were not willing or able to provide.

To obtain control of territories in Medellín, Doble Cero identified any criminal bandas who were fighting against the militia and hired or coerced them into joining or backing the BM in their conflict against the militia (Lamb 2010, 184). One of these bandas which worked alongside Doble Cero was a banda called La Terraza. This banda was already allied with Don Berna and the Castaño brothers (Gutiérrez & Jaramillo 2004, 26). La Terraza was led by a man who used the alias of El Negro Elkin, and he had worked for and then against Escobar. This banda was made up of 200 to 300 members who were ex-militiamen, delinquents and ex-police officials (Lamb 2010, 170). The banda was known for being responsible for crimes such as the kidnapping of a Colombian senator, and the assassination of almost 3000 individuals, including a peace activist, two human rights lawyers and 12 employees from a division of the Colombian Office of the Attorney General (Cuerpo Técnico de Investigación) (El Tiempo 1999b; Semana 2001a). The alliance between the right-wing BM and the criminal group, Banda La Terraza, showed how armed non-state actors pursuing different goals could find common ground and cooperate.

This alliance between Don Berna, the Castaño brothers and La Terraza however only worked well for a short time. This was because around 1999 a conflict started between Banda La Terraza and Carlos Castaño. There are different accounts as to why this conflict started with some relating it to El Negro Elkin demanding more payment from Don Berna for services and others saw it in terms of the Banda La Terraza committing crimes against a paramilitary unit or in the name of a paramilitary unit (Lamb 2010, 185). The conflict further escalated when a member of the banda assassinated Don Berna’s brother. Following this assassination, Don Berna paid Carlos Castaño a substantial amount of money and formally joined the AUC (Lamb 2010, 185). The formal involvement of Don Berna in the AUC created a powerful and extensive network of paramilitary and criminal elements (Gutiérrez & Jaramillo 2004, 26). Don Berna created a new block of the AUC called the Block Cacique Nutibara (Bloque Cacique Nutibara - BCN). He informed the bandas who were already working for him as contractors in the Oficina de Envigado that they now worked for the BCN of the AUC. However, their day-to- day employment duties, such as working as sicarios remained the same. This substantial powerful network (of the BCN and the Oficina de Envigado) was able to eliminate any opposing bandas or militia groups. Furthermore, the BCN claimed they were eliminating any 98

‘undesirables’ i.e. left-wing militia, which meant they could count on the ‘passivity’ of the authorities (Gutiérrez & Jaramillo 2004, 26). However, the involvement of Don Berna in the AUC was largely motivated by economic reasons, rather than the right-wing political objectives of the paramilitaries.

At the end of the 1990s, the situation in the city had slightly improved from the crisis of 1991. There had been a reduction in homicides from the high rates of 1991, although there was still a significant presence of non-state armed groups in the city. These groups, the paramilitaries, the militia and the criminal bandas all used homicidal violence to obtain and/or maintain territorial control. However, there had been a change in how these groups used violence against the state. Since the removal of Escobar, individuals involved in narco-trafficking had refrained from using high levels of open violence, such as bombs, to avoid drawing the attention of the authorities to their activities. These non-state criminal actors had acknowledged that if they were involved in violent conflicts, this would draw the attention of the state which would adversely affect the profits they could obtain from their criminal activities, namely narco- trafficking.

Another change from the non-state armed groups of the 1980s was it had become increasingly difficult to categorise them. Rough categories, mainly used by government officials and the media, labelled bandas as generally being associated with narco-trafficking, militia groups tended to be associated with left-wing politics and paramilitary groups were generally affiliated with the AUC. There were however blurred boundaries between these categories and they did ‘not do justice to the fluid nature of illicit activity’ in the city (Lamb 2010, 340). For example, all of the non-stated armed groups provided their host communities with services, such as in social and security matters. These groups presented a significant challenge to state authority, as in their communities they were perceived almost as legitimate actors because of this provision of services. Another similarity between the non-state armed groups was the characteristics of members. The members of these groups were largely young males from the poorer neighbourhoods of Medellín. These young males appeared to be attracted to a group, left-wing, right-wing or criminal, for the same reasons, which were the social identity gained from their membership and the economic returns they obtained. For example, research from Baird (2011) found from 40 life-history interviews with youths who sometimes called themselves gang members and sometimes paramilitaries, not one interviewee claimed to have joined gangs for political reasons. During data collection in 2014, a government official (GO3)

99 with in-depth knowledge of the non-state armed groups provided further insight on the allegiances of these young males:

‘In the period of the militia, a lot of the individuals changed to be involved in narco- trafficking. There were people from ELN who changed to be part of the BCN. Narco- trafficking is such a strong business. There were organisations who originally had left- wing revolutionary character and the narco-trafficking organisations got them to work for them…Here is one example, there was a leader who worked for Don Berna. He was known as Antonio, he was originally from the ELN. However he ended up working as a paramilitary leader. He was assassinated because of internal problems. A lot of people think the problem of violence in Medellín is easy to understand, but just look at how complicated it all is!’.

Conclusion

This chapter has provided an overview of the pattern of violence in Medellín in the 1990s. In the early part of the 1990s, state and non-state groups acted very violently and targeted homicides were the preferred mode of resolving territorial disputes. Homicides reached appallingly high levels with the city portrayed as the ‘most dangerous city in the world’ in 1991. The data presented in this chapter showed how the increase in homicidal violence was linked to three conflicts involving non-state armed groups active in the city. These different conflicts show the importance of taking a multifaceted approach in explaining the emergence of violence in Medellín and why the violence was spiralling out of control.

Another theme discussed in this chapter was the state investments in the poorer neighbourhoods to reduce the violence and prevent future outbreaks. However, these investments were largely perceived as inadequate and these communities continued to be marginalised from the larger society. The third theme discussed in this chapter was the difficulties in categorising the different non-state armed groups in the city because of their ‘blurred practices’ and the common characteristics of members. Furthermore, the right-wing paramilitaries, the left-wing militia and the criminal bandas all used homicidal violence to obtain and/or maintain territorial control.

The following chapter provides a discussion of the reduction of violence in Medellín in the 2000s. In this period, Medellín left the list of the ‘50 most dangerous cities in the world’ and

100 by 2015 the city had the ‘lowest homicide rates in 35 years.’ The chapter discusses the policy approaches used by the local government to reduce the violence and prevent future outbreaks. It also analyses the conflicts between the non-state armed groups in the city, their use of violence and their motivations.

101

Chapter Six: 2000-2015: A Significant Reduction in Homicides

Introduction The previous chapter provided an overview of the pattern of violence in Medellín in the 1990s. In this period, homicides reached appallingly high levels with the city portrayed as the ‘most dangerous city in the world’ in 1991. The problems of violence were acknowledged by national and international stakeholders and there were some government policies to reduce the violence and prevent future outbreaks. The policy makers took a multifaceted approach in explaining the emergence of violence, thus the need for justifying the need for their multi-pronged approaches to reduce the violence and prevent future outbreaks. However, this period was largely characterised by fragmented implementation of these policies. Throughout the 1990s, different non-state armed groups such as urban militia, criminal bandas and paramilitaries used homicidal violence to obtain and/or maintain territorial control. This chapter provides an overview of the decline of levels of violence in the city in the 2000s up to November 2015. In this period, there was a reduction from the extremely high levels of homicides in the 1990s to moderate ones, and by 2015 Medellín departed from the list of the ‘50 Most Dangerous Cities in the World.’ The purpose of this chapter is to provide a discussion on the policy approaches by the local government to reduce the violence and prevent any future outbreaks. These policy approaches were widely cited as the reason for significant reduction in homicides as they appeared to justify the assumptions made in the most influential urban violence reduction and prevention approaches. There is also a discussion on the non-state armed groups, militia, criminal bandas and paramilitaries, focusing on their use violence to obtain and/or maintain territorial control, their characteristics and their motivations. The final part of the chapter provides insight into the relationship these non-state armed groups have with the state and their communities. These insights provide important lessons for policy makers in the region seeking to address urban violence.

A New Wave of Violence

102

By 2000, the right-wing paramilitary group, the AUC, had two blocks operating in the city. The first block was the BCN led by Don Berna, who also led the criminal organisation, the Oficina de Envigado. The second block was the BM, led by Doble Cero. Doble Cero was known as a ‘traditional paramilitary’ and had the number one objective of defeating the left- wing militia groups (Lamb 2010, 182). The BM and the BCN generally had to use force, in the form of homicidal violence, to co-opt any criminal bandas or militia groups that were against them.

There were however some cases when the BCN and the BM did not have to use force to co- opt rival groups. For example, one left-wing militia group, the November 6 and 7 Militia (Milicias 6 y 7 de Noviembre - M-6&7) affiliated with the ELN, changed their alliances to the right-wing block of the AUC, the BM. The M-6&7 had left-wing sympathisers, as their name celebrated the M-19’s attack on the Palace of Justice in 1985, and it is unclear as to why their allegiances changed (Lamb 2010, 103). Some interpretations are that they changed because of death threats from the BM or the prediction the AUC would eventually win the conflict. Another interpretation was as one former member put it ‘the urge was always to have a motorcycle, to get girls, to be in a gang’, rather than any political agenda (Lamb 2010, 195). This interpretation shows how members might not be motivated by the political ideology of a group, but rather by the social and economic benefits which they could obtain from membership. There were some individuals in the M-6&7, who were motivated by the left-wing political objectives of this group and they found sanctuary in the remaining left-wing FARC and ELN militia groups in the city.

These urban militia groups active in Medellín were essential for larger FARC and ELN forces with military and other resources throughout the country. Around this time the ELN was operating in East Antioquia and they had some groups in the north-eastern comunas of Medellín, while the FARC was operating in the north-west of Colombia and had some militia groups in the western comunas. There was also an urban militia group which did not have a political affiliation operating in the central-western comunas. However, for the AUC these areas were all important access routes in and out of the city and thus there were violent conflicts between these right-wing and left-wing groups to obtain control of these territories. The access routes were particularly important for the AUC in terms of their involvement in narco- trafficking, from which they obtained 70 percent of their revenue (Amnesty International 2005, 8). The conflicts between these groups over territory were noted when the BCN faction of the AUC attempted to take control of the central-western comunas and the ELN retaliated with the

103

Anti-Paramilitary Revolutionary Front (Frente Revolucionario Anti-Paramilitar) (Lamb 2010, 188). The use of such politically charged names indicated that the Medellín violence was not simply about narco-trafficking but that the city had also become a battleground in the broader Colombian civil conflict. While the battles appeared to reflect political motivations, the fight for the control of urban territory still had a strong economic character, most often related to narco-trafficking.

Not only were there violent conflicts between the left-wing and right-wing groups but also the conflict continued between Banda La Terraza, the Castaño brothers and Don Berna. This conflict had started around 1999 for inconclusive reasons and was played out in both the poorer and elite comunas. These non-state armed opponents were not only using selective assassinations but also other methods of violence. For example, in 2001 there were two car bombs in the elite comuna, of El Poblado. The first bomb was in a Medellín mall, which was ‘Colombia’s most modern shopping mall’, and resulted in one death and 53 injuries (El Tiempo 2001a). The second was detonated in a park frequented by university students and resulted in the death of eight people and 138 injuries (El Tiempo 2001b). While authorities never found out who was responsible for the attacks or their motivations, Banda La Terraza was the number one suspect (Semana 2014). The first bomb was seen as more of a warning message to the Castaños, as the banda knew their affiliates had business interests in this mall. The second bomb in the park was clearly intended to kill innocent civilians (Semana 2001b). Both the car bombs had a significant impact on individuals from this elite comuna, as the explosions demonstrated that the conflicts between warring armed non-state groups were no longer confined to the poorer comunas. Following the attack in the park, a public vigil was held and businesses re-opened within 24 hours, with locals refusing to allow these perpetrators to take away their public space (El Tiempo 2001c).

By 2002, the area in the city with the majority of militia combatants, a comuna in the west- central part of the city, Comuna 13, had constant violent conflicts between the AUC and the militia, and state authorities and the militia. This comuna had been known for its violent conflict between the different non-state armed actors, such as the ELN, the M-19 and the FARC militia groups in the 1990s. This was because the geographical location of the comuna meant it provided important access routes to the rest of the country, and was thus crucial for the trafficking of drugs, guns and money (Lamb 2010, 188; Insight Crime 2011). The militia group with direct relations to the FARC and ELN, called The Peoples Armed Commandos (Comandos Armados del Pueblo - CAP) had obtained control of this area since around 1996.

104

By 2002, the conflict between the CAP, the AUC and state authorities reached its peak. In the first six months of 2002, there were 442 homicides in this comuna, which included targeted assassinations of community and religious leaders. These leaders were targeted because the militia prohibited any public meetings and the leaders had disobeyed their rules (Martin 2014, 445). These high levels of violence in this comuna had an effect on all spheres of life for its residents. For example, during participant observation in 2014 and 2015, residents commented on how the constant conflicts meant they could often not attend work or school or had to wait for the fighting to stop for them to return home from school, with one individual commenting on how most people in this area only received half of their schooling (Field Diary, July 2014). This semblance of governance that was provided by the militia, while it was often violent, was however in stark contrast to the inadequate formal state governance and the comunas’ history of the repressive use of force by state actors.

The levels of violence were so bad in Comuna 13 that the situation was described as resembling a ‘type of civil war’ (Franco 2004, 298). As a community police member commented, the militia prevented public authorities from entering the area and ‘dead bodies would spend days in the streets’ (Martin 2014, 445). The violent conflicts continued despite a national government operation to stop it (Operation Mariscal) in May 2002. This operation consisted of a large-scale joint raid by police and intelligence units designed to take the area by force (Lamb 2010, 189). During the operation, 31 individuals were captured and nine civilians were killed (Martin 2014, 448). However, the operation was deemed a failure, as it did not make a significant impact on the actors in this area (Lamb 2010, 189). Not only did the militia have support from the national left-wing groups, such as the FARC, but they also had some passive and direct support from the community as the CAP provided a semblance of governance by engaging in practices of social cleansing of delinquents and other undesirables (Angarita et al. 2008, 78).

Pacification of the City

With the election of Luis Pèrez Gutièrrez of the Liberal Party to the position of mayor in 2001 came increased public visibility of state authorities throughout the city. This was because Pèrez was of the opinion that the high levels of violence were because of a ‘lack of security management’ and so he focused on strengthening local institutions, such as the police and

105 judiciary (Martin 2014, 432). Pèrez also closed the Office of Peace and Co-Existence that had previously negotiated with the bandas, arguing that the office had not ‘fulfilled its purpose’ (Riaño Alcalá 2006, 177). There was, however, a significant distrust of Pèrez and his policy approaches, as he was perceived as yet another corrupt politician in the traditional mode. One NGO (NGO2) worker summed up what many citizens said:

‘Pèrez was a very traditional, corrupt politician. They called him Luis of 15 percent, because this was the percentage of money he stole from contracts. A typical corrupt crook of a politician who only used his position to steal public resources’.

On the national level the newly elected President, Alvaro Uribe, commenced his administration in 2002 with the promise of a new security policy, which he called ‘’ (Fajardo & Andrews 2014, 6). This policy involved adopting a mixed approach of force against the left-wing FARC and negotiations with the right-wing paramilitaries (Kruijt & Koonings 2007). The force element of the strategy was especially noted in October 2002 when the national government sent a ‘strong message’ of its intent to the remaining militia groups in Comuna 13 when launching Operation Orion. During the operation, over a thousand troops from different agencies invaded the Comuna 13 territory, resulting in the deaths of civilians, state forces and militia members (Martin 2014, 458). The authorities viewed the operation as a success, as the militia left the city shortly after the operation. However, the use of state repressive force was not viewed favourably by residents from this comuna. Comments from residents emphasised how they needed employment and social development policies rather than military actions by government troops (El Colombiano 2002b). During personal participation observation, residents of the area frequently commented on how the state never had any real presence in this area and even now their main experience of the state was in terms of repressive force (Field Diary, July 2014).

Operation Orion was only a short-term success as although the militia groups left the city, the power vacuum was soon filled by the BCN. It had been widely acknowledged that the BCN had ‘quietly provided intelligence and tactical assistance’ to state forces during the military operation (Richter & Miller 2007, 2). For citizens of Comuna 13, it was simply a matter of the changing of the guard from the militia to the paramilitaries, with the repressive methods of the paramilitaries similar to those of the militia. The key difference for residents was that paramilitaries employed the tactic of ‘making a person disappear without any evidence’

106

(desaparecidos), whereas the militia had simply ‘assassinated people’ (Rozema 2008, 441). On obtaining control of the area, the paramilitaries introduced their own sets of rules to control the daily life of citizens, such as prohibiting formal gatherings in the streets and collecting compulsory donations (vacunas) from businesses and transport companies. The paramilitaries also announced over 400 employment opportunities and, similar to the criminal bandas, young males were attracted to the money and status offered by this group, even if they did not align with their political ideologies (Rozema 2008, 441). The BCN had obtained control of the territory through the state’s military operation and they maintained control of the territory through their use of violence and provision of services to the community. The military operation in Comuna 13 may have cleared out the militia but it had not helped with the marginalisation of the youth from this area. They continued to lack access to legitimate employment opportunities in the larger society.

With the militia groups now all but removed from the city, it was hoped the levels of violence would decrease. However, a conflict soon started between the two AUC blocs in the city, the BCN and the BM. This conflict was mainly related to the ‘strong dislike’ the leader of the BM, Doble Cero, had for groups protecting cocaine production facilities and narco-trafficking or ‘narco-paras.’ Doble Cero preferred to obtain funding from renting out private security services and collecting vacuna payments from residents. Doble Cero was one of a few para leaders who refused to allow money from narco-trafficking to corrupt his work (Lamb 2010, 182). His motivations were in contrast to leaders of the AUC, who obtained 70 percent of their revenue from narco-trafficking (Amnesty International 2005, 8). Given that the Castaños did not have the support of the BM, they ordered and provided support for the BCN to defeat the BM, and BCN members threatened BM members with death if they did not join the BCN (Amnesty International, 2005, 30). Security forces also ‘unofficially’ aligned with the BCN, because of their use of force against any remaining left-wing militia in the city (Lamb 2010, 203).

The BM was defeated around May 2003 and Doble Cero was assassinated. This defeat was significant as now Don Berna, the leader of the criminal organisation, the Oficina de Envigado, and the paramilitary bloc, the BCN, had monopoly control over the criminal underworld and all of the territories in the city (Lamb 2010, 205). There was no need for territorial conflicts between the non-state armed groups. As an NGO worker (NGO2) commented, this control had an effect on the levels of violence (homicides) in the city. This NGO worker said:

107

‘From 2003 there was a reduction of violence, even though there were different structures of criminal groups, there was a control of the city and this was because of the pacification of the city’.

Not only were homicide rates declining in Medellín, but also there was a significant political change. Sergio Fajardo from an independent party had been elected as mayor in December 2003. This was the first time an independent local party had obtained political power in Medellín. Fajardo had campaigned in the 1999 local elections and came third. Since this defeat, he had been campaigning in the city, building an image of ‘having sympathy with the average citizen’ (Devlin & Chaskel 2010, 2). This campaigning enabled Fajardo to gain support from all sectors of society. For example, he gained the backing of the influential business elites because of his professional background, as an academic and business leader. Fajardo also had a following from NGOs, social organisations and academics because of his experience in the civic movement, Citizens Commitment (Compromiso Ciudadano), which had formed in the late 1990s from stakeholders involved in the Presidential Council (Maclean 2014, 38; Gutiérrez et al. 2013). In the election, Fajardo received the most votes ever obtained by any mayoral candidate (210,000 votes) and entered office with a 72 percent approval rating (Lamb 2010, 252).

The defeat of the BM and the arrival of Fajardo occurred at the time when the national government under President Uribe was negotiating with the right-wing paramilitary groups. This negotiation was significant as it was the first time a national government had ever negotiated with them (Rozema 2008, 428). The paramilitaries had entered into these negotiations as a ‘political organisation with paramilitary backing’ with the leader of the AUC, Carlos Castaño, publicly denouncing any members’ involvement in narco-trafficking. This denouncement had angered many AUC members and created tension within the organisation, as Carlos had gained significant wealth from his involvement in narco-trafficking (Lamb 2010, 224). Carlos was assassinated in 2004 with disgruntled members of the AUC believed to be the main suspects. While the paramilitaries previously had support from the public, given their objective of destroying the left-wing groups, they had recently lost a lot of this support. This was because of their use of excessive violence, such as leaving behind beheaded or disembowelled bodies and they had been branded a ‘terrorist organisation’ by the US, following the 9/11 attacks on New York City and Washington in 2001(Lamb 2010, 171 & 224).

108

Despite the lack of public support and the internal problems of the AUC, on July 2003, the AUC and government negotiators signed the framework for negotiations. A pilot demobilisation program in Medellín was to begin in November 2003 and if the project was successful, then it would be implemented throughout Colombia. In a televised ceremony, on 25 November 2003, 873 paramilitary combatants handed in their weapons with Don Berna the first to demobilise. He and all combatants pledged to re-enter society as ‘peaceful civilians’ (Lamb 2010, 190). They had their criminal records checked, ensuring there were no pending investigations, before they received a de facto amnesty and returned to their communities (Amnesty International 2005, 38). The national government recognised the NGO, Corporation Democracy (Corporacion Democracia) as the ‘corporate representative of the combatants’ (Amnesty International 2005, 38).

The Fajardo administration had the tough task of implementing this demobilisation program that had arisen from the national negotiations. The Medellín local government under Fajardo developed a demobilisation program which they called the Peace and Reconciliation Program ‘Return to Legality’ (Programa Paz y Reconciliacion ‘Regreso a la Legalidad’). They had taken advice from national and international experts. The program was led by a team of psychologists and social workers engaging with the combatants to address any mental health problems and also to assess the possible educational needs and employment of each combatant (Rozema 2008, 443). Subsidies were also provided in the form of financial incentives for employment and for businesses set up by the combatants (Rozema 2008, 443; Amnesty International 2005, 43). A government official (GO1) interviewed in 2014 explained how the government had learnt from previous failures (such as Coosercom in the 1990s) and also had taken on advice from national and international experts:

‘…This program did not happen overnight, we had to design the program, understand it, revise previous programs. We also had advice from the UN and we worked closely with the national government…’.

This government official (GO1) then commented on how the local and national government acknowledged the characteristics of individuals who would be involved in the program:

‘The majority of the combatants had basic school levels, most of them had not finished secondary school, some of them did not even finish primary school…Through the program they could have improved their school levels. There was also psychological

109

work that needed to be done, not only with them, but also with their families and the neighbourhood where they would be living…Every combatant had working opportunities and received a little bit more than the minimum salary. This was a significant experience about showing an example of how a person who used to be an armed actor could be a peaceful citizen in their community…’.

One former government (FGO2) with knowledge of the demobilisation program also interviewed in 2014, commented on the importance of ensuring the program focused on acknowledging these characteristics in implementing the program:

‘We worked one on one with each combatant. This was about generating trust… Each psychologist had 100 demobilised combatants, but within six months they were required to know what each individual combatant was up to, where they were going, if they were attending workshops, visiting their families…We knew about the family structure, personal transformations of each combatant, such as if they consumed illegal substances, if they had a criminal record… It was about creating a personal perspective for each individual’.

The hard work of the local administration soon paid off and the demobilisation program gained national and international recognition as ‘an innovative model of intervention to deal with situations of complex violence’ (Abello-Colak & Pearce 2015, 205). By May 2005, 87 percent (762) of the demobilised combatants had paid employment (Rozema 2008, 434). Furthermore, to avoid undermining this ‘innovative’ peace process, the national government began to use the label ‘criminal bands’ (bandas criminals - BACRIM) to differentiate between any paramilitary groups and organised crime groups in the country (McDermott 2014, 5).

Despite the best intentions of the program and the international recognition of its success, there were a number of issues prior to its implementation. One of these issues was concerning the number of BCN members who had demobilised. For example, before the first televised ceremony, the strength of the BCN was estimated at over 2000 combatants in Medellín. However, only 863 combatants demobilised in this ceremony, and there was evidence that a large number had ‘redeployed’ to East Antioquia shortly before the program began (Amnesty International 2005, 40). Another issue was concerning proof that an individual was actually a member of a paramilitary. For example, a last-minute change to the requirements of the demobilisation plan meant that the paramilitary leaders only needed to submit a list of names and not actual proof that the ‘combatant’ was a member of the group. This resulted in a number

110 of banda members who had never been part of a paramilitary group ‘joining’ the BCN and participating in the demobilisation program because of the economic benefits offered (Lamb 2010, 233; Rozema 2008, 443). A former government official (FGO4) interviewed in 2014, provided some perspectives on these issues when he said:

‘The demobilisation provided a way back to civilian life. But a lot of the members were not paramilitaries. They had an inventory, but it was not precise. They thought there were 15,000 or 18,000 but the program attracted 36,000 people or something like that. This was because the program created very favourable conditions They were also some narco-traffickers attracted to the program, but not all were narco-traffickers…For example, many young people from the poor comunas were told ‘You don’t have anything to eat? Well if you come into this program, there will be advantages, in two years you will have subsidy of $500 or $300, you will be able to go to university, you will have access to health, the only thing you have to do accept that you were in one of these [paramilitary] groups.’ As a result, artificial paramilitary structures were created…Many people got a lot of advantages from the program, such as Don Berna, who was a criminal’.

Not only were there these issues prior to the implementation of the program, but also there were numerous other issues during its implementation. For example, one issue was the inadequacies of the employment preparation training combatants were offered. An official engaged in this program, best described this situation as ‘they [the combatants] were sent to work as mechanics and didn’t even know how to use a spanner’ (in Martin 2014, 486). Training was noted as difficult particularly due to the low education levels of the combatants. As one government official involved in the program indicated, ‘90 percent were illiterate, 43 percent had not finished primary school education and 85 percent did not have secondary education’ (in Martin 2014, 486). There was also criticism of the types of job offered to these individuals. For example, over 200 security positions were offered to these combatants in the private security initiative called ‘Secure Zones’ (Zonas Seguras). This initiative was coordinated by the Medellín local government and sought to privatise security arrangements in some neighbourhoods in the city. This initiative however had echoes of Coosercom, the semi- criminal and semi-legal organisation of the 1990s following the militia demobilisation program. Participants who were involved in this program had committed many abuses against civilians and many had become involved in criminal activities (Gutiérrez & Jaramillo 2004,

111

23). Following criticism from Amnesty International and other human rights organisations, this program was abandoned (Amnesty International 2008, 41).

Another issue concerned the types of employment the former combatants were offered. For example, while the young men were offered training and vocational courses many were only ever offered temporary work in low-status employment such as in construction or cleaning (Martin 2014, 485). One of the reasons why they were not offered higher status employment was because there were significant levels of distrust of these actors because of their violent histories. For example, during participant observation, one citizen commented ‘You shouldn’t talk to the supposed demobilised paramilitaries, they are criminals and liars, I am surprised they did not rob you’ (Field Diary, July 2015), a sentiment which was expressed by many other paisas. A former government official (FGO2) who had knowledge and experience in the demobilisation program spoke at large of the failures of the demobilisation program and cited the widespread distrust of these combatants as one of the reasons for this failure:

‘One day, 150 combatants graduated from the program. It was a great day, the auditorium was full, with all the mums and kids… But you know that same day an ex- combatant attacked a bus with a knife… Guess what was the first story in the 6pm news? That the demobilised combatants – see the plural? That they, not one, were attacking a bus… The news that 150-young people graduated from the program didn’t even make the news! Another example, an ex-combatant graduated from University with honours and he came to my office and said ‘Please don’t attend the graduation because then people will know that I am a demobilised person and I don’t want anyone to know that’… Another example, a demobilised combatant is walking with their friend and the friend has a gun and people see the friend and comment ‘look, the demobilised have not been disarmed!’.

Integral Urban Projects

The Fajardo administration continued to work hard to address the problems in Medellín. They set the agenda of addressing the inequalities in the city and improving its international image. Fajardo had a young and dynamic team, made up of architects, anthropologists, urbanists, social workers, academics, community leaders and members of NGOs, with the majority

112 having some experience in the Presidential Council in the 1990s. They were determined to make a change in the city and brought a wealth of experience and knowledge with them.

In addressing the agenda of lessening the inequalities in the city, the first year of Fajardo’s administration saw slow progress. The administration was criticised by the wider public, with Fajardo responding that they were ‘planning and not improvising’ (Fajardo & Andrews 2014, 7). They finally released their first projects, which they called Integral Urban Projects (Proyectos Urbanos Integrales - PUIs). The objectives of these projects were to increase public investment and extend public services into the poorer comunas under the broad areas of transport, education and health, ensuring ‘real impact’ in a short time (Fajardo & Andrews 2014, 6 & 9). The projects had three components: physical (public spaces), social (community involvement) and institutional (collaboration between agencies) (Alcaldía de Medellín 2011). The local government argued that the improvement of the social, economic and spatial problems of these comunas would contribute to improvements across the city (Alcaldía de Medellín 2011). Thus, they would be able to obtain the support from all sectors of society. These urban development policies soon became known as ‘Social Urbanism’ (Urbanismo Social). They included initiatives that invested in infrastructure for the poorer areas of the city and iconic architectural projects to address the ‘historical social debt’ owed to these marginalised comunas (Maclean 2014, 5).

The first PUIs were implemented in the north-eastern comunas, which were generally seen as the poorest, and most dangerous, parts of the city (Maclean 2014, 14). Residents of neighbourhoods in these comunas were initially ‘sceptical’ of the projects, having built up over many years of neglect an overall negative perception of the local government (Martin 2014, 503). Many of the residents were from families that had arrived in the rapid wave of migration from the 1950s and had been marginalised from the established Medellín society for many years. The residents of the poor comunas were not used to receiving social development projects and benefits from state authorities and had generally only experienced the state as an agent of repression. The local government was aware of this distrust, especially because some members of Fajardo’s administration came from these marginalised communities and some had work experience for NGOs in the communities. Therefore, the Fajardo administration ensured they worked closely with community members. For example, to promote greater involvement and ownership of these projects by the local population, the administration specified that ‘all public work contractors must hire a significant component of local manual labour’ (Dávila & Daste 2011, 9).

113

The new work in these comunas actually continued that begun by the previous mayor, Luis Pèrez. He had started planning to create a public transport system to connect these neighbourhoods of approximately 100,000 inhabitants with the rest of the city using ski-lift technology, called the MetroCable (Martin 2014, 43). The MetroCable addressed the limitations of the public transport system (Metro), as this part of the city is characterised by its steep gradients and hilly topography (Dávila & Daste 2011, 1). Residents from these neighbourhoods could now, comfortably and in an innovative form of public transport, travel to the economic centre of the city in ten minutes, instead of two hours while travel between places within these neighbourhoods had been reduced in some cases, from forty-five minutes to fifteen (Dávila & Daste 2011, 2; Maclean 2014, 30). However, not only did the MetroCable reduce commuter times but also it was the first-time ski-slope technology had been used a means of public transportation in an urban setting (Dávila & Daste 2011, 2; Maclean 2014, 30). Of particular significance was that this reduced commuting time facilitated increased access to employment opportunities for these residents.

Another significant project of the PUI in the north-eastern comunas was the construction of a library park (parque bibloteca) in the Santo Domingo neighbourhood at the end of the MetroCable line. Prior to the construction of the library park, government workers had engaged in community consultation, in the form of ‘Workshops for Imagining Projects (Talleres de Imaginarios Por Proyectos) to gain the perspective of community as to what the library park should look like and what it should provide for them (Fajardo & Andrews 2014, 7; Calderon 2012, 5). The final product was a building which provided the community with a library, a community centre and a childcare facility. The Spanish monarchy had given a substantial amount of funding for the library park and because of this it was named the Spanish Library Park (Parque Bibloteca España), with the King and Queen of Spain even attending its official opening in March 2007.

Among the results of the PUI in the north-eastern comunas was the creation of 125,000 square metres of public spaces. Neighbourhood parks increased from three to 17 in nine neighbourhoods, 3,439 new jobs were created and 290 programs were provided in education, health, social welfare, recreation and sports (Alcaldía de Medellín 2011, 77). The average journey time for those using the MetroCable and Metro combination dropped from 120 minutes to 65 minutes, thus improving their access to employment opportunities and services (Área Metropolitana del Valle de Aburrá 2006). Over the period 1997-2007, there was a 13-point

114 increase in the quality of life index in the intervention neighbourhoods, compared to an average of eight points in other districts of the city (Medellín Como Vamos 2007).

During interviews and informal conversations in 2014 and 2015, policy makers involved in the PUIs spoke of the importance of improving the built environment, modernising essential infrastructure, promoting public spaces and facilitating access to education in the poorer neighbourhoods. These policy makers argued that these interventions were done with the objective of reducing the levels of inequality and poverty in the city. For example, a former government official (FGO3) who worked in the Fajardo administration spoke of how policy makers saw that improved access to education in the poorer neighbourhoods would provide these residents with the tools to progress in the formal society:

‘We knew that sicarios would not give up their guns to sit down and read. But, a young person who was nine years old when we started the project would have the opportunity to see the world [via the internet], have contact with social networks, read books and/or learn to play a musical instrument. Our hypothesis was that, to see the world, to learn about other people. Well they would learn to be a better human being…’.

An academic (A1) who was also involved in the Fajardo administration spoke of how the social urbanism policies assisted in providing alternatives for individuals from the poorer neighbourhoods. This academic also commented on how these improvements could provide one of the solutions to the violence in the city:

‘You can’t conclude that social urbanism or the education projects would provide a solution to the violence, but it did enable important civic processes and social development, which help in making violence more difficult. It was about ensuring that a young person would have other alternatives…’.

A government official (GO2) who was also involved in the Fajardo administration spoke of how the improvements to the built environment were an important part of transforming the city:

‘These projects were not about addressing the violence and insecurity in the city, but about addressing the inequalities. So, we placed an importance on state investment in the poorer sectors for their necessities, such as schools and libraries, access routes and childcare centres… The social investments such as the library parks, mega-colleges, boulevards, child care centres, cultural spaces, renovation of the botanical gardens, the

115

football parks, the creation of the interactive science and technology park, recreation spaces were made not to improve levels of security in the city, but instead to transform the city’.

Along with these interventions in the north-eastern comunas, the local government also further developed the Participatory Planning and Budgeting Program (first implemented in 1998). The aim of this program was to bring government and planning closer to the citizenry and civil society, as well as legitimating local government and promoting transparency (Maclean 2014, 33; Fajardo & Andrews 2014, 9). This was an opportunity for the local administration to promote positive interaction between local residents and state authorities. In the local planning process, projects such as community kitchens and football parks were completed and funds were made available to finance 1,200 scholarships for young people to pursue university studies (Maclean 2014, 33; Dávila & Daste 2011, 9). Importantly, the local planning process offered to local communities living under violence ‘opportunities to develop a preference for legal modes of participation’ (Uran 2010, 152). This political change in the city was especially significant for individuals from the poorer marginalised comunas. For example, a community leader (CL3) from one of the intervention neighbourhoods who was interviewed in 2015 openly supported Fajardo. This leader praised the politician for his work and expressed how for the first time the community organisation he was involved with felt like they had a voice and an opportunity to show the wider society that they were more than being from a poor and violent comuna:

‘In this period, we could have influenced at least three or four projects which came from our neighbourhood, or from our community and ended up being projects in the whole city. We had never seen this before! It was about showing that the community organisations also had proposals for improving life conditions in the city. It was about recognising that these community organisations had knowledge. It was about recognising that these community organisations have contributed to development in the city’.

The local administration also took concerted steps to improve the international, and national, image of the city, which had been severely damaged from over 30 years of high levels of violence and conflict. This agenda was important to the business elites in the city and thus Fajardo had their continued support in pursuing this objective. During interviews with former government officials, NGO workers and community leaders, they commented on this objective

116 of improving the international image of the city. For example, a former government official (FGO1) interviewed in 2015 said:

‘It was about doing a lot of work to find an exit from the story people knew about the city… The marketing of the city was about being a model of urbanism…’.

A government official (GO1) interviewed in 2014 who was involved in the Fajardo administration provided examples of how Fajardo invited international, and national, political and non-political stakeholders to the city:

‘Fajardo always said ‘whenever I go to other countries to speak with their governments or the King of Spain I would always finish the meetings saying, I came here, but next time I want to see you in Medellín to show you what we are doing.’ The excuse was to bring them to the city, so that’s why the library parks all have auditoriums to host local, national and international events… There was a strategy about people dedicated to selling the good points of the city’.

A worker from an NGO (NGO2) interviewed in 2014 spoke of the different techniques, such as event management, used by the administration in improving the international image of the city:

‘For this model to be successful there needs to be a lot of marketing… Medellín has a group of international businesses dedicated to bring international events, such as congresses for lawyers and engineers, the campaign for the youth Olympics, CLASCO in 2015, a network of more than 500 Latin America research centres who have their assembly every three years. The marketing of the city was done with publications and this image is so important for the city. For example, in the World Urban Forum, Joseph Stiglitz did a conference and he was taken around the city and wrote a column about the city in the New York Times praising the city. This type of international media coverage is so important for the city and when anyone criticises the city in the media, this criticism is considered as bad for business’.

A community leader (CL2) also interviewed in 2015, commented on how the investment in the infrastructure was about this international image:

‘They wanted to make Medellín an international city, so they have dressed it up with this mega-urban transformation, to make it one of the most important cities of Latin

117

America… They invested heavily in the physical infrastructure so that the city is modern and competitive, all to bring foreign investment’.

The Medellín Miracle

Sergio Fajardo was a well-respected individual in the city, and had always been known for his ‘charisma and wit’ (Maclean 2014, 53). By the end of his administration, he had done something that no previous mayor had been able to do: he improved transparency in finances, demonstrated clearly by a 47 percent increase in individuals and companies submitting tax returns (Devlin & Chaskel 2010, 8; Lamb 2010, 245). Under Fajardo, the city had at last built what appeared to be a capable, accountable, effective and responsive public sector. During participant observation and interviews, many citizens spoke of their respect for Fajardo. They commented on how his administration provided a ‘breath of fresh air’ with many perceiving Fajardo as an honest person who did not follow the clientelistic political patterns of the past (Field Diary, May 2015). The following comment from a prominent business leader (BL1) reflects this sentiment:

‘Fajardo changed the way of being a politician in the city. People were tired of the same corrupt politicians. Fajardo showed people that without corruption there is money for everything’.

However, rather than only remarking on positive individual characteristics of Fajardo, many interviewees stressed the need to look at the nature of Fajardo’s administration. This administration was the result of effective collaboration between business leaders, academia, NGOs, community leaders and government authorities, a relationship that had begun with the Presidential Council in the early 1990s. The ‘transformation’ of the city did not start with Fajardo, rather it had been the result of over twenty years of hard work from these individuals (Field Diary, May 2014 & 2015). The following comments from an academic (A1) and an NGO (NGO2) were typical of all participants interviewed in 2014 and 2015 when asked about Fajardo’s administration:

‘I think that in 2004 a more civic government arrived, which allowed a space of trust between different actors, such as people and institutions who were previously working on different projects. For example, a group of researchers were previously working,

118

without any government support, on how to transform the poorest and more violent areas of the city and Fajardo was interested in collaborating with this project’. [A1]

‘Fajardo did some important work. However, in his administration there were echoes from previous administrations. Yes, there were some important developments under the Fajardo’s administration; however, you cannot say that in that specific moment the city started to improve…’. [NGO2]

The Fajardo administration brought about a ‘significant shift in local and political planning culture’ (Dávila & Daste 2011, 10). Fajardo finished his term with an unprecedented 90 percent approval rating from all sectors of society (Fajardo & Andrews 2014, 10). Most significantly, there was a ‘deep change from the negative attitude the communities of the lower socio- economic neighbourhoods had towards the administration’ (Calderon 2012, 10). There was some evidence that the increased access to employment opportunities in the intervention comunas had improved their economic situation, as indicated by the rise in the quality of life indexes. The built environment of these comunas had been improved. The local planning process had empowered previously marginalised individuals. The international image of the city had also improved. For example, during the Fajardo administration the city underwent a ‘construction boom’ of luxury hotels, apartments and shopping malls and saw the arrival of 70 foreign enterprises (Hylton 2007).

Not only were there these social and economic improvements but also homicide rates had fallen to 37 (2004) from 58 per 100,000 in 2007. This was a 32 percent reduction from the beginning of Fajardo’s administration and almost a 90 percent reduction from the high rates of 1991. Some international organisations started to make claims about how the social urbanism policies caused the reduction of homicides. For example, the UN-Habitat’s Report on Enhancing Urban Safety and Security: Global Report on Human Settlements 2007 compared Medellín very favourably to other Colombian cities (2007, 38):

It is interesting to observe differences between cities within countries, such as reductions in crime and violence achieved in Medellín, Colombia, where a strong mayor and civil society were determined to overcome drug violence and dominance of drug lords compared to continued problems in Cali.

Shortly after the Fajardo administration finished its term, the Inter-American Development Bank (2009) praised the policies used in Medellín for reducing the homicides:

119

…After two decades of improved law enforcement applied in conjunction with massive infrastructure and social investment, Medellín has regained its reputation as a city with an attractive quality of life and a good place to do business…Results were striking as in addition to becoming more appealing as a city with modern architecture and green spaces, the homicide rate was cut by more than two-thirds from its peak rates in 1991.

Despite these claims of the success of the social urbanism policies, community leaders, non- government workers and some government officials interviewed in 2014 and 2015 commented on the inadequacy of the interventions in addressing the day-to-day problems in the poorer neighbourhoods. For example, a government official (GO3) interviewed in 2014 commented on how, despite the involvement in the community in the development of policy in the Fajardo administration, policy makers still did not understand the day-to-day problems of these comunas:

‘These people [the government] have not felt the harshness of the social problems in the poorer comunas. Until they understand these types of problems they think the solutions they are providing will resolve everything. But it is not so, these problems are still in the comunas. Investments are made such as the Metro, MetroCable, and Library Parks and of course they have an effect, but they do not resolve all of the problems this city has’.

An NGO worker (NGO1) interviewed in 2014 was very critical of the interventions and their inadequacy in addressing the day-to-day issues of residents when he commented on the continuing problems in the city:

‘If you want to have an innovative city, why is the city full of casinos? Why do we have so many people addicted to gambling? Is this what we want to show the world? Do you know how many young boys and girls are involved in prostitution? Do you think that in an innovative city they sell virginity? A foreigner or someone of high status can take the virginity from an 11-year-old girl – it costs just seven million pesos [US$2,000]! Is this innovation? Is it innovation when there are houses where they dismember people? Or torture people? … Yes, I agree the MetroCable is very nice, it is cheaper for the poorer people to get to the centre of the city. But what about the library parks where they don’t even have internet access?’.

120

There has also been criticism from national and international researchers that the interventions were inadequate in addressing the particular needs of the poorer comunas. For example, many of the users of the MetroCable are informal workers travelling to and from the economic centre of the city, which often requires them to travel with bulky items, such as market goods. However, because of the design of the MetroCable bulky items are banned from it (Dávila & Daste 2011, 7). The reduction of commuter times (from forty-five minutes to fifteen) did not take into consideration waiting times at stations, which may be from 10 minutes to over an hour at peak times. Another criticism was the cost of using this form of transport. While using the MetroCable is cheaper if users are travelling long distances, for shorter distances, such as to and from the centre of the city, the bus may be cheaper and more convenient (Dávila & Daste 2011, 7; Brand 2011, 8). Other criticism concerned the library park. While the building did provide the community with a library, a community centre and a childcare facility, it has been criticised for being more a ‘formal show of fashion’ (Hernandez-Garcia 2013, 49). For example, as an academic and activist in 2012 commented ‘why on earth is it called after the King and Queen of Spain if it’s all meant to be about community and bottom up development?’ (Maclean 2014, 31). The cost of the building has recently been subjected to criticism, with the building closing in October 2015 for repairs to structural problems, with estimates that the repairs will cost more than the original construction (Americas Programme 2015). An NGO official (NGO1) interviewed in 2014 did comment on this when he said:

‘They only officially opened the Spanish Library Park because the King of Spain was coming. Today it is falling apart and they have to rebuild it. They only reason why they wanted to show off the Library was because the King of Spain was visiting!’.

There has also been some criticism that the agenda of the international image was more important than the agenda of addressing the inequalities dividing the city’s population. For example, an architect at a Bogota-based university criticised the use of the term ‘social urbanism’ as simply a type of branding exercise for the damaged international image of the city to assist in its transformation into an ideal hub for business, tourism and investment (Hernandez-Garcia 2013). International and national urban planners also criticised the use of the term ‘social urbanism’ to describe the social development policies on the grounds that ‘all urbanism has a social element’ (Kalandides & Hernandez-Garcia 2013, 47). During participant observation in 2014 and 2015, the phrases ‘la innovación es maquillaje’ (the innovation is just a cosmetic change) or ‘el alcalde solo quiere maquillar la cuidad’ (the mayor just wants to show off cosmetic changes to the city) could be heard on a daily basis in both the elite and

121 poorer neighbourhoods. There were even comments by respondents that the international campaign to improve Medellín’s image possibly offended the victims of the violence in the city, as the victims of violence were forgotten in this international campaign (Field Diary June 2014). Many academics and NGOs also reported that if they criticised the international image of the city, they were perceived as being ‘an enemy of the city’ (Field Diary June 2015). Furthermore, despite the improvement in the quality of life (13-point increase from 1997 to 2007, compared to 8 points in other parts of the city) in the intervention comunas, these improvements had broadly followed the overall trends in the city. Also, comparisons between composite indices, such as quality of life and human development index, indicated that these improvements were much less than seen in the wealthier comunas and Medellín was listed as the ‘most unequal city in Colombia.’ (Brand 2011, 9)

Despite the international claims that the social urbanism policies from the Fajardo administration caused the reduction of homicides, some international and national researchers wondered whether this reduction was because of the decline in territorial conflicts between the non-state armed groups in Medellín (Fukuyama 2011; Hylton 2007; Semana 2005). For example, in late 2003 Don Berna had obtained control of all territories in Medellín, as the militia groups had all but been removed from the city, the BM block had been defeated, Don Berna’s rival, Banda La Terraza had been defeated and all other criminal bandas worked for Don Berna in the Oficina de Envigado. Thus, Don Berna had a monopoly control over all criminal actors and territory in the city. This control was so complete that some traded the term, ‘governability’, for ‘Donbernability’ (Donbernabilidad) as Don Berna managed illicit networks in the city, so much so that not one assault, robbery or assassination could take place without his permission (Lamb 2010, 238; Fukuyama 2011; Hylton 2007; Semana 2005 & Semana 2009).

A ‘Tenuous’ Reduction in Homicides

Fajardo finished his term in December 2007 and his Chief of Staff, Alonso Salazar, was elected as mayor. Not only was this the first time in the history of Medellín that an independent political party had won for the second time but also in contrast to Fajardo’s elite background Salazar was from a middle-class background. His election was a considerable change from previous mayors who had been put in place by the city’s conservative elite. Salazar had spent years working as a social researcher and journalist and was the author of the bestselling book, ‘Born

122 to Die in Medellín’ (No Nacimos Pa Semilla) which was based on his interviews with sicarios in the late 1980s. During interviews and informal conversations, all participants commented on how the election of Salazar showed how much politics had changed in Medellín with the following comment from a former government official (FGO3) summing up a generally expressed sentiment:

‘No one could have imagined that a former journalist like Alonso Salazar would have been elected as a mayor!’.

Despite Salazar inheriting a very different city to the one Fajardo did, his first year of government presented significant challenges. One of these challenges concerned the paramilitary demobilisation program. The initial challenges of this program had now turned into larger ones. For example, some former paramilitaries had kept in close contact with their former commanders and the extortion of buses, an activity attributed to the paramilitaries, continued (Rozema 2008, 444). Community members from the Comuna 13 also spoke of up to ‘150 paramilitaries wearing plainclothes during the day, but AUC armbands and military-style uniforms during the night and patrolling the area’ (Amnesty International 2005, 42). Residents from this comuna still felt that the former paramilitaries were the controllers of their neighbourhood, with one resident commenting on how these ex-combatants were ‘punishing people who were accused of wrongdoing’ and how residents asked them to ‘solve certain problems, such as domestic violence’ (Rozema 2008, 447). These non-state armed actors continued to provide their own interpretation of law and order to which residents had to submit, as these residents perceived that the state could not adequately provide these services.

Salazar had generally been supportive of the demobilisation program and had argued that any suspicious behaviour by ex-combatants did not have ‘conclusive evidence’ (Semana 2008). However, in July 2008, he publicly referred to a recently assassinated demobilised combatant, Job Lopez, as ‘never disassociating himself from criminal activities’ (Lamb 2010, 276). This sent a message to the combatants and the wider society that the local government had given up on the demobilisation process. Thus, during participant observation, one demobilised combatant commented on how he ‘felt like the government had given up on him, just like any government programs in the past’ (Field Diary, July 2015).

The comments from Salazar were then echoed by the national government with the decision to extradite Don Berna to the US with other ‘narco-paras.’ Don Berna had been arrested in October 2005 for being connected to the murder of a local politician. However, his extradition,

123 with other narco-paras, to the US on drug trafficking charges in 2008 created a power vacuum in the criminal world of Medellín (Rozema 2008, 440). There was not an obvious successor to take over Don Berna’s empire. This led to violent conflicts for the leadership of the Oficina de Envigado. The power vacuum also opened up space for a new group Los Urabeños, which was loosely linked to the AUC. Los Urabeños used the removal of Don Berna as an opportunity to expand their growing drug trafficking network and attempt to take over the Oficina de Envigado (McDermott 2014, 5). Los Urabeños and the Oficina de Envigado used homicidal violence to obtain control over territories in the city. The conflict between these criminal groups was noted in the dramatic increase in homicides from 45.6 per 100,000 in 2008 to 94.4 per 100,000 in 2009. There was even a massacre of eight individuals in a local nightclub in 2010, with the Oficina de Envigado the number one suspect (Semana 2010). During data collection in 2014 and 2015 when participants were questioned about this rise in homicides they spoke of the vacuum that had been created when Don Berna was extradited. For example, a former government official (FGO4) interviewed in 2014 commented on how the banda members were now left without an employer when he said:

‘In this time, shooting started again in Medellín, because these criminals were left without a defined structure, without a leader. There was no one to pay their salaries’.

An official from a research centre (OF1) interviewed in 2014 commented on how this increase in homicides was because of Don Berna. This official commented on how this increase was a criticism of the view that the policy approaches used by the Fajardo administration were the cause of the reduction in homicides:

‘When Don Berna was extradited to the United States, there was a chain of homicides that followed. This is very clear, as it had left these [criminal] structures free to do whatever they wanted. This is when we started to see an increase in homicides. This is really a criticism of Fajardo, because it is saying that the initial reduction was not because of Fajardo’s government’.

The local government reacted to this rapid increase in homicides with repressive and restrictive policy measures. For example, by 2010, there were 2,510 private security guards guarding 609 public institutions with 180 security cameras and 200 alarms around the city (Alcaldía de Medellín 2010). Legal curfews were introduced for young people in Comuna 13 (Abello-Colak 2013, 100). The number of police officers and soldiers patrolling the city increased. A new model of policing by neighbourhood quadrants (Policia de Cuadrantes) was introduced

124

(Abello-Colak 2013, 99). This model, based on community policing, was about having more police presence in a neighbourhood and ensuring that community members personally knew the police in their neighbourhood. The intention was that the police could build a better relationship with the community and in return, the community could provide them with information about the whereabouts of criminals (Bulla et al. 2014, 38).

These security approaches however have had their limitations. For example, given the history of state-violence in the poorer comunas, the citizens perceived the police as ‘corrupt and ineffective’ and ‘resented and did not trust them’ with even the police acknowledging this distrust (Abello-Colak & Pearce 2015, 209; EAFIT-CAP 2014, 31). Furthermore, despite the increase in police numbers, few citizens ‘knew how to access the neighbourhood policing service’ and when they did, they were not satisfied with the service they received (EAFIT- CAP 2014, 32). The curfews in the Comuna 13, while supported by some sectors of the population, especially the more elite comunas, were strongly rejected and criticised by civil society and community organisations (Abello-Colak 2013, 100). The young people from Comuna 13 also interpreted them as an ‘unfair punishment’ characterising young people as a ‘particular dangerous category of citizen’ and resembling curfews imposed by other armed actors in their communities in previous years (Abello-Colak 2013, 100).

Concurrent with these security approaches, there was also an investment in crime prevention programs. For example, the local government implemented a program called Youth Force (Fuerza Joven) designed for young people (14 to 29 years old) at risk of being recruited to a criminal organisation (Abello-Colak 2013, 100). This initiative combined different programs and offered participants an alternative to criminal activities and incentives not to ally with criminal organisations. Participants were offered a subsidy, access to education and cultural and social activities (Abello-Colak 2013, 101). The program was recognised internationally as ‘good practice’ and seen as a success with 1500 participants signing up (Abello-Colak 2013, 101).

However, Youth Force reinforced the perception among residents that former members of armed groups linked to illegal criminal activities were ‘more successful at obtaining special government benefits than ordinary citizens’ (Abello-Colak & Guarneros-Meza 2014, 17). This was seen with a key requirement to access the program being that a participant ‘needed to be a

125 member of a banda1 or a group connected to a type of banda’ resulting in many young people joining a banda just to obtain the benefits of the program (Abello-Colak 2013, 101). Furthermore, many beneficiaries of the program continued to participate in illegal activities and only used these state provided resources to ‘buy more drugs and arms’ (Abello-Colak 2013, 101). Also, the bandas sometimes threatened the program participants, forcing them to pay a percentage of their subsidies or be killed (Abello-Colak 2013, 101). This was noted with the assassination of a program participant in January 2011 by a suspected local banda which had been trying to extort the money this individual was receiving from his participation in the program (Caracol Radio, 2011).

Despite the good intentions of these violence reduction and prevention policies, they could not affect change in the criminal underworld of Medellín. This was because since the extradition of Don Berna to the US, the Oficina de Envigado lacked his ‘shrewd and rational’ leadership (Gutiérrez & Jaramillo 2004, 26). Most of the Oficina de Envigado now belonged to Erick Vargas, who used the alias of ‘Sebastian’, and much of the international trafficking was handled by Maximiliano Bonilla, who used the alias of ‘Valenciano’ (Insight Crime 2015b). Sebastian and Valenciano were however in conflict to obtain complete control over this highly profitable criminal organisation. The authorities had some success when Valenciano was captured in November 2011, leaving Sebastian as the leader of the Oficina de Envigado. However, more violent conflicts followed with many of the factions of the Oficina de Envigado who had originally supported Valenciano now aligned with Los Urabeños. These individuals also wanted to take over the Oficina de Envigado (Franco & Caputo 2013, 21). By 2011, local sources estimated there were between 100 and 300 separate groups operating in the city with links to the Oficina de Envigado or Los Urabeños and the majority of members were younger than 35 years old (Semana 2012). There was another success for the authorities with the capture of Sebastian in August 2012. However, a massacre of nine Oficina de Envigado members in December 2012 illustrated how brutal the conflict still was. Interpretations of this massacre relate to either Sebastian or Valenciano ordering it, despite being in prison, to ensure their interests in the Oficina de Envigado were being protected (Insight Crime 2013; El Tiempo 2013). This event was just one example of how, in contrast to ‘Donbernability’, homicide rates

1 In some literature on gangs in Medellín, these groups are referred to as combos or bandas. But some literature on gangs in Medellín does separate these groups into criminal bandas and delinquent combos. For reasons of clarity, the term banda is only used. However, the particular term that a participant used has not been changed. 126 were not stable, as the criminal actors were fighting over the control of territory and the state could not prevent these actors from using extreme violence in the pursuit of territorial control.

Not only had there been increased conflict between members of the Oficina de Envigado and Los Urabeños but also the poorer comunas were being subject to less bloody modes of violence. For example, a 2013 study from Humanitarian Actions in Situations Other Than War (HASOW) emphasised how these criminal actors were adopting the practice of threatening residents to get them to leave the group’s territory or face severe consequences such as death. This is a type of violence known as forced urban displacement, which was employed by criminal actors to maintain their control in a territory. The HASOW findings were confirmed with the number of urban displacement victims increasing more than 40 percent from 2010 to 2011 with 8434 victims in 2011 (Abello-Colak & Pearce 2015, 208). While the non-state armed groups were not necessarily engaging in direct violence in their territories, such as the use of homicidal violence, through the use of this indirect type of violence, they continued to maintain control over their communities. A respondent interviewed in 2014 (OF1), who had knowledge of forced urban displacement in the city, provided further insight into this increase in this practice and its relationship with criminal actors in the city:

‘In the 2006 and 2007 Medellín Como Vamos report [which evaluates the changes in quality of life in different socio-economic zones] combos were only mentioned as a small threat to insecurity. However, from 2008 to 2011, combos appeared as a threat and there was a relationship between this appearance and the increase in forced urban displacement…The combos had not been using force and then they started to’.

‘Reformers have Combatted Crime and Poverty in Medellín’

During Salazar’s administration in Medellín, there had been further claims from international organisations that the social urbanism policies had caused the 32 percent reduction of homicides during the Fajardo administration. For example, the first global study on urban violence by the World Bank in 2011, Violence in the City: Understanding and Supporting Community Responses to Urban Violence, commented on how the social urbanism policies had caused the reduction in Medellín’s crime rates (World Bank 2011, 79):

127

Central and local government have developed successful prevention policies to reduce violence. In a number of cases, these policies have yielded impressive results. Bogota, Medellín, Sao Paulo, and others have seen their crime rates fall impressively.

The monthly international architectural magazine, the Architectural Review, also linked the reduction of crime with the social urbanism policies. It was argued that the urban improvements caused the reduction in violence (2011, 2):

Reformers have combated crime and poverty in Medellín with programmes that foster decent, dignified architecture and responsible urban planning… As recently as 2003, the hillside slums of Santo Domingo…were a no-go zone for outsiders, and residents were advised to stay off streets that were controlled by urban militias. A systematic programme of urban improvements changed all that…

There had also been an important study examining the effects of the interventions in the north- eastern comunas. The objective of this study by Colombian and US epidemiologists was to analyse these interventions, such as the MetroCable, and whether their implementation can be linked to changes in violent activity, i.e. homicidal violence (Cerdá et al. 2011). Their study found that the decline in homicides was 66 percent greater in the intervention neighbourhoods than in control neighbourhoods (without a MetroCable) and resident reports of violence decreased 75 percent more in intervention neighbourhoods than in control neighbourhoods. Thus, these epidemiologists concluded that the physical and social integration of informal urban neighbourhoods can reduce criminal acts. Their findings have since been widely cited, notably in the 2015 Inter-American Development Bank Report Closing Knowledge Gaps: Toward Evidence-Based Crime Prevention Policies in Latin America and the Caribbean.

Salazar finished his administration in 2011 and the election of the new mayor, Aníbal Gaviria, was significant as it broke the shift in political culture of the eight years of the independent governments of Fajardo and Salazar. In contrast to Salazar and Fajardo, Gaviria was from the Liberal party, and had an elite background and experience in politics, previously serving as the Governor of Antioquia (2004 to 2007). The first change from the independent government of Salazar and Fajardo to the Liberal government of Gaviria was noted in an increase in repressive approaches to reduce violence. For example, there was increased investment in security measures including electric police vehicles, local police units known as Centres of Immediate Action (Comandos de Acción Inmediata - CAIs), intelligence patrols and security cameras (Field Diary, May 2015). There was also ‘a record public investment’ (100,000 billion pesos

128 or just over US$30 million) in justice and security funds from the Medellín-based state-private telecommunications company, UNE-Millicom (Medellín Como Vamos 2015, 66). The numbers of police neighbourhood quadrants increased from 120 in 2012 to 411 in 2014, or an increase of just over 300 percent and by the end of 2015, there were just over 8300 police officers patrolling the city (El Tiempo 2015). The number of security cameras had also increased from 383 in 2012 to 823 in 2014, so that by 2015 Medellín had the ‘most sophisticated camera system’ in the country (El Tiempo 2015). During data collection, participants were asked about the effectiveness of these security and policing approaches. Community leaders, officials from NGOs and some government officials spoke of the limitations of these approaches in preventing future outbreaks of violence. For example, a government official (GO3) commented on how the private security cameras were infiltrated by ‘illegal businesses’ when he said:

‘You can invest all you want… You can put 1500 cameras, but the social and economic structure is infiltrated by illegal businesses. Has anyone ever pointed this out to you?’.

An NGO worker (NGO2) argued that the increased surveillance and police presence would not reduce the violence in the city:

‘They have an obsession with video-vigilance…they think video-vigilance will resolve everything. However, this is really following a broken windows approach, an approach which has been heavily criticised… Having a police station on every corner and a security camera on another corner is not the solution to reduce the indicators of violence’.

A community leader (CL2) interviewed in 2015 also commented on these limitations:

‘When a territory wants to be controlled by another group this is when things get ugly. The police cannot prevent this conflict, nor can a security camera’.

Along with this increase in security technology and police presence to reduce the violence, the Gaviria administration also changed the involvement of community organisations in the development of policy. While community organisations were seen as ‘legitimate interlocutors’ in the Fajardo and Salazar administrations, under Gaviria these organisations were now no longer viewed as important participants in local decision making (Abello-Colak 2013, 96). Many community groups and NGOs highlighted this lack of community consultation during data collection in 2014 and 2015. For example, during participant observation in 2015, one community leader from the Comuna 13 even commented on how the only time the Gaviria

129 administration made contact with his organisation was to publicise the effectiveness of local government policies as shown in the recent reduction of homicide rates. This community leader commented that while the area had the most security cameras and police presence in the city, their presence did not dissuade the criminal actors (Field Diary, June 2015). For the poorer comunas, the change in the local government engagement with the populations of these neighbourhoods reflected the return of clientelistic political relations as expressed in the following comments from community leaders (CL1, CL2 and CL3):

‘For us [the community group] Fajardo and Alonso were possibly the only candidates and mayors who visited our territory. The dialogue changed under these governments. They were more receptive to the suggestions we proposed…Now [the Gaviria government] the state looks at us [the community group] from above instead of from below. We want the state to look at us on the same level…. We are so sure that if there were not groups like ours, the violence in the city would have been worse. There would have been so many more young males killed in this absurd war. Also, there would have been less opportunities for young people and people would have been more desperate…The local government now looks at the community groups as if they are mediators between the local government and the community... As if the role of the community groups aren’t necessary…. I think it is a dangerous model.. He [Gaviria] even told our group that our building was too big and he said ‘give me a part of the building and I will put my government projects in and whatever the community asks for I will give it to them…’. [CL1]

‘These governments from 2004 to 2011, from Fajardo and Alonso, generated some interesting conditions. They delivered programs, which the community asked for. They organised some programs and projects in their government plans, which had not happened in previous governments… In this [the Gaviria administration] government, the participation programs that were created in the Fajardo and Alonso administration have been destroyed’. [CL2]

‘Under the Gaviria administration, we [the community organisation] get very small resources and very little recognition from the local government’. [CL3]

These NGO officials and community leaders provided insights into the limitations of the policy approaches in reducing the violence and preventing future outbreaks. Obtaining these deep and rich insights was only made possible through the methodology of this research, that is using

130 the context-specific approach. The policy approaches which had been widely praised by international organisations and epidemiologists for causing the reduction of violence, relied on homicides and did not take into account other measurements of violence, such as perceptions of insecurity.

‘The Most Innovative City in the World’

In 2013, the city received more international attention for the social urbanism policies and the reduction of homicides. This attention was because Medellín was selected as the ‘Most Innovative City in the World’ by Wall Street Journal, Citigroup and the Urban Land Institute. In awarding Medellín this prize, the Wall Street Journal wrote of the improvements to the built environment, the enhanced access to education, the improving social relations in the city and the reduction of homicides (2013):

Medellín’s homicide rate has plunged, nearly 80 percent from 1991 to 2010. The city built public libraries, parks, and schools in poor hillside neighbourhoods and constructed a series of transportation links from there to its commercial and industrial centres. The links include a metro cable car system and escalators up steep hills, reducing commutation times, spurring private investment, and promoting social equity as well as environmental sustainability. The local government, along with businesses, community organizations, and universities worked together to fight violence and to modernize Medellín. Community organizations, health centres, and youth groups have formed, empowering citizens to declare ownership of their neighbourhoods.

International praise for the policy approaches in causing the reduction of homicides also came from a well-known specialist in security and development, Robert Muggah. In a TEDGlobal Talk in Brazil in 2014 entitled ‘How to protect fast-growing cities from failing’, he spoke of how the improvement to the built environment, the modernisation of essential infrastructure and coalition building had caused the homicides to reduce in Medellín (Muggah 2014b):

Social cohesion matters. Mobility matters in our cities. We’ve got to get away from this model of segregation, exclusion and cities with walls. My favourite example of how to do this comes from Medellín. When I lived in Colombia in the late 1990s, Medellín was the murder capital of the world, but it changed course, and it did this by deliberately investing in its low-income and most violent areas and integrating them with the middle-class ones through a network of cable cars, of public transport, and first-class

131

infrastructure, and in the process, it dropped homicide by 79 percent in just under two decades.

Muggah (2015, 30) then reiterated these claims in a 2015 publication entitled ‘A manifesto for the fragile city’ in the Journal of International Affairs when he said:

There are many examples of how to design-out crime emerging from global cities… Arguably the most stunning case is Medellín… A succession of mayors led by Alonso Salazar and Sergio Fajardo turned things around by devoting more attention to tackling the poorer and most dangerous comunas, or neighbourhoods. The slums were purposefully connected with middle-class areas by a network of cable cars, bus transport systems and first-class infrastructure. Homicidal violence declined by almost 80 percent…

There were also quantitative studies which argued the policy approaches had caused the reduction. For example, researchers in spatial economics at a Medellín-based university and researchers from a geo-environmental cartography and remote sensing group at a Valencia- based university tested ‘which land cover, structure and texture descriptors were significantly related to intra-urban homicide rates in Medellín Colombia, while controlling for socioeconomic confounders’ (Patiño et al. 2014). These quantitative researchers argued how their findings were ‘expected and supported by social disorganisation theory’ and broken windows theory by concluding:

Areas with higher homicide rates tended to have higher local variation and less general homogeneity; that is, the urban layouts were more crowded and cluttered, with small dwellings with different roofing materials located in close proximity to one another, and these regions often lacked other homogeneous surfaces such as open green spaces, wide roads, or large facilities (Patiño et al. 2014, 48).

The homicides continued to reduce and by 2015, the city even departed from the list of the ‘50 Most Dangerous Cities in the World’ and had a homicide rate of 20 per 100,000. Not only was this rate lower than some US cities such as Baltimore which had a rate of 54 per 100,000, but it was also over a 90 percent reduction from the high rates Medellín experienced in 1991 (Business Insider 2016). During fieldwork in 2014 and 2015, there were numerous government sponsored advertisements throughout the city celebrating how the reduction in homicides was the ‘first success of [the government’s] 100 successes’ (Hecho No 1 de 100 hechos). A

132 government official (GO2) interviewed in 2015 provided and argued that local government policies were an integral reason for this decline in violence:

‘In 2014, Medellín had the lowest homicide rates in 35 years… We started our government in 2012 and in 2011 the homicide rate was 63 per 100,000 or 1649 homicides. Our first year of government the rate reduced to 52 per 100,000, a 24 percent reduction, about 400 homicides less than the year before… In 2013, the rate was 38 per 100,000 or 925 homicides… In 2014 it was 659 homicide cases or 26.8 per 100,000.. The lowest homicide rate in the last 25 years and in the first four months of this year almost 98 less homicide cases than the year before… We are going to finish this year [2015], even though you cannot predict anything, below 18 per 100,000, the lowest homicide rate in the last 40 years! This is an absolutely significant reduction!’.

‘I think that Medellín has a history of the last 14 or 15 years that has been making progress on several issues that focus on reducing levels of violence. The first element I think is addressing the issues of inequality… For example, programs which mean there is more equality in the city such as the ‘colegios de calidad’ (quality schools) the library parks… Everything that has to do with education is the fundamental way to address levels of inequality… The current mayor says the E from equality starts with the E from education! All the investments in colleges, education, library parks, health, and neighbourhood football parks have allowed Medellín to understand that investing in the poorer people is a business… I also think another important element is capacity building; I am an advocate for institutions… The other element which is especially important with this administration is life, so our slogan is ‘Medellín: Everyone for life’… Our central focus has been about more just, human, free and happy citizens. The second element is equality, which is achieved through health, education, social inclusion, economic development, sustainable, development of infrastructure, education, transport and institutional capacity’.

The city was now seen as a benchmark for other Latin American cities seeking to make and implement successful violence reduction and prevention policies. For example, the European Forum for Urban Security organised a study visit for mayors of seven Guatemalan municipalities to Medellín to learn about the city’s experience in implementing citizen security and crime prevention policies (European Forum for Urban Security 2013). Politicians, planners and police officials from cities such as Rio De Janeiro, Johannesburg and Washington also

133 visited the city to further understand the ‘Medellín experience’ (The Guardian 2013a). Headlines such as ‘Medellín, Colombia: Reinventing the world’s most dangerous city’, ‘Medellín: Reborn: Colombian city moves out of Escobar shadow’ and ‘Medellín: How Colombia’s second city overcame its drug cartel past’ were common throughout international media (The Guardian 2013b; CNN 2013; International Business Times 2013). The then US Secretary of State, Hillary Clinton, had even recently commented on how emerging Arab democracies should look to Medellín for a ‘frame of reference in tackling their security problems’ (Colombia Reports 2013). These international appraisals of the social urbanism policies all relied on the objective indicator of violence, homicides, to demonstrate the success of the policy approaches in Medellín.

There had also been important advancements in the collection and analysis of homicidal data in the city. For example, the government agency, System of Information for Security and Cohabitation (Sistema de Información para la Seguridad y la Convivencia - SISC) had been created under the Salazar government. This agency analyses homicide data taken from the Colombian National Institute of Medicine: Legal and Forensic Sciences (Legal y Ciencias Forenses Instituto Nacional de Medicina - INML), Colombia’s Department of Criminal Investigation (Seccional de Investigación Criminal - SIJN) and Technical Investigation Team (Cuerpo Técnico de Investigación - CTI). However, their data can only reveal where a homicide victim was found and not where the victim could have been assassinated. Another advancement was the creation of the Medellín-based research centre called the Centre of Political Analysis (Centro de Analisis Politico) in 2006. This research centre had the objective of interpreting, explaining and providing recommendations to the local government about levels of violence in the city (Centro de Analisis Politico 2016). Through the collaboration of SISC and the Centre of Political Analysis they were able to analyse homicides in the city from 2003 to 2013 and mapped these homicides in the city. Their analysis found that just over a third of homicides (32.1 percent) are located in 11 polygons (or nodes) which are dispersed throughout 9 of the 16 comunas in 53 of the 249 neighbourhoods in Medellín and cover an area of about 11.78 kilometres (Giraldo & Preciado 2015, 9) (Figure 5). As a researcher from the Centre of Political Analysis clarified during data collection, the location of these polygons, being between neighbourhoods, indicate that just over a third of homicides in Medellín concern territorial conflicts between bandas. This researcher also identified how this homicide mapping could assist policy makers in allocating government resources in the identified polygons to reduce and/or prevent violence (Field Diary, May 2015).

134

Figure 5: Homicide Polygons in Medellín

Image Removed due to Copyright

Source: Giraldo & Preciado 2015

In explaining homicides which were not located in the polygons the researchers from the Centre of Political Analysis stated: ‘we could intuitively suggest two of every three homicides corresponded to diffuse spontaneous violence and citizen intolerance’ (Giraldo & Preciado 2015, 11). While they do not provide an explanation of what is meant by ‘spontaneous violence’ and ‘citizen intolerance’ it is assumed they are referring to incidents of conflicts between residents which turn into violent acts. Their suggestion is confirmed with one of the most violent days of the year, in all parts of Colombia, being Mother’s Day. On Mother’s Day 2015 in Medellín, there were 793 reported fights and four homicides, with one journalist even expressing that ‘Mother’s Day should be prohibited’ because of these high levels of violence (El Colombiano 2015c; Minuto30 2016). There is no precise understanding as to why Mother’s Day is such a violent day. Some interpretations are linked to the high amounts of alcohol consumed during family gatherings or long-running disputes turning into violent conflicts (El Colombiano 2015c; Devlin & Chaskel 2010, 6; Minuto30 2014). An NGO worker (NGO2) interviewed in 2014 spoke of the high levels of spontaneous violence on Mother’s Day:

‘Not all homicides in Medellín are related to confrontations between the armed groups. For example Mother’s Day is a very dangerous day… People drink a lot and there are a lot of street fights and fighting in families and since there is both a large presence of guns and knives, people end up stabbed and die as a result and they don’t have anything to do with narco-trafficking’.

A community leader (CL3) interviewed in 2015 spoke of the high levels of spontaneous violence to resolve grievances following soccer games:

‘People kill each other after soccer games.. They leave the game and attack each other and end up being stabbed, just because they are wearing the shirt from the other team… The only way to explain this is because life doesn’t have any meaning…’.

However, the Centre of Political Analysis acknowledged that care must be taken with relying on homicide rates in measuring levels of violence and the presence of non-state armed groups

135 in the city (Giraldo & Preciado 2015; Casas-Casas & Giraldo 2015; EAFIT-CAP 2014). This care was also acknowledged by participants during data collection. For example, community leaders (CL2 and CL3) interviewed in 2015 argued that while the social urbanism policies may have caused some of the homicide reduction, the non-state armed groups were still present in the city and their actions greatly determined homicide rates:

‘The public policies can be effective. However, the reduction of homicides cannot just be explained by public policies. Unfortunately, the existence of armed groups still determines homicides in the city’. [CL2]

‘You cannot deny that there has been a reduction of homicides since the 1990s. But this reduction is not because of public policies’. [CL3]

As indicated by these community leaders, the reduction in homicides cannot just be explained by the social urbanism policies. Some interviewees argued that a more plausible explanation for the 90 percent reduction in homicides was a change in how the non-state armed groups use violence (homicides) to obtain and/or maintain territorial control in the city rather than the social urbanism policies. The reasons why these criminal actors changed how they used violence was best explained by a government official (GO3) interviewed in 2014:

‘I will recognise that homicides have reduced, but they have also reduced because the groups are not in conflict, because it is not convenient for them to have confrontations. They need to run their businesses effectively and make sure there aren’t any arrests…’.

A worker from an NGO (NGO2) interviewed in 2014 also expressed this sentiment:

‘They [the homicides] have reduced because the criminal organisations have taken the decision not to use such obvious violence [such as homicides]. They have understood that if they use such obvious violence, then it will call attention to their presence. It is almost like a type of tactic’.

These insights into the patterns of violence (homicides) in Medellín were only made possible through the time spent on-the-ground in Medellín, interviewing an array of actors and engaging in various forms of participant observation. Whilst previously cited international organisations and researchers claimed that the social urbanism policies had caused the reduction of violence (homicides), the insights provided from interviewees and from participant observation indicated how the leaders of the criminal organisations in Medellín made decisions to operate

136 less conspicuously to avoid drawing attention to themselves from the authorities. This is a rational decision as to how they use violence in the pursuit of their prime aim of maximising profits from their criminal activities, namely narco-trafficking. Many other sources also indicated that during this time the ‘business model’ of the Oficina de Envigado and Los Urabeños changed. This new model was based on co-operation between the two groups, rather than using violence for territorial control (McDermott 2014). The new model was often referred to as the ‘gun pact’ or the pacto de fusil, an agreement signed between the Oficina de Envigado and Los Urabeños in July 2013. While some government officials denied the existence of the pact, other sources explained how the pact divided territory in the city and lifted any invisible borders (fronteras invisibles) which had previously marked out the territory between the different criminal actors. The pact also enforced the ‘death penalty’ for any rule breakers and ordered that if any assassinations took place, they needed to be conducted more discreetly, such as by making a corpse disappear (Semana 2013a & 2013b; Jaramillo & Gil 2014, 146; González 2014, 99). One community leader (CL2) interviewed in 2014 provided a detailed description of the pact:

‘There is a pact between the different illegal groups. Groups you can call either bandas or paramilitaries. The pact is not public, but it does exist. This pact means that they can’t take the territory of another group and whilst this respect is kept, there will not be confrontations… This pact means that no one is permitted to assassinate people and if they do, they must do it outside of the neighbourhood or outside of Medellín or dismember the body and throw them into the Medellín river. Importantly there is no authorisation to kill in the neighbourhood or in Medellín. You have to ask permission from the armed group to assassinate someone’.

An NGO personnel (NGO1) interviewed in 2014 also spoke of this pact and how it can explain the reduction of homicides in the city:

‘…You are talking with the person who first started using the term pacto de fusil. It is real! Donbernability is related to the pacto de fusil. If you do not have anyone to fight with, then the homicides will reduce. However, when there are confrontations, homicides and massacres are noted. Before the pact, there may have been around 40 or 60 homicides in a weekend, but when there is ‘peace’, we are talking about 5 or 6 on a weekend… Before 13 July 2013 there were armed confrontations constantly and a lot

137

of homicides… But around the end of July we [the NGO] noticed certain changes.. We received information that the criminal groups had signed a pact…’.

An immediate reduction of homicides from 90 (in May 2013) to 60 (in August 2013) and a 30 percent reduction by the end of 2013 gave weight to these respondents’ views that a pact between criminal groups was a major reason for falling homicide rates. The existence of this pact not only indicates the flexibility but also the resilience of these groups (Gil 2013, 12). The decision to negotiate as a preference to using violence shows how the criminal actors in Medellín in 2015 have learnt from the failures of the Medellín cartel in using open violence against the state, such as car bombs and selective assassinations against political and state authorities. This escalation of violence attracted increased attention from the authorities and directed the cartel’s energies into violence and thus reduced the profits cartel members could obtain from their criminal activities, namely narco-trafficking. Given that homicide rates are used as an indicator of levels of violence, a reduction in homicides demonstrates that there has been a reduction of violence and thus suggests policies are working and that armed criminal gangs are no longer a threat to public security. This suits the gangs as it enables them to focus on criminal activities largely undisturbed.

The existence of the pacto de fusil suggests that any reduction in homicide rates in the city is ‘partial and fragile’ and ‘subject to outside factors’ (Giraldo & Preciado 2015, 12; EAFIT-CAP 2014, 28). For example, the reduction in homicide rates was commented on by participants interviewed for this study in 2014 and 2015. A former government official (FGO5) interviewed in 2014 spoke of how a reduction in homicides in an area in the city does not indicate the absence of criminal actors:

‘There are places that are not on the police radar and then they ‘explode’... For example Barrio Triste was relaxed and then one week ago, there were eight homicides. The police did not think it was a dangerous place. But then a gibara, which is similar to the drug dealer in the neighbourhood who sells bazuco and marijuana, had some problems and then he turned up dismembered and because of the effect of this dismemberment there were eight homicides’.

Another finding from the Centre of Political Analysis is that homicides continue to be ‘organised, selective, have specific patterns and are directed towards specific populations’ (Giraldo & Preciado 2015, 12; EAFIT- CAP 2014, 28). One example of this was during data collection there were three separate grenade attacks in July 2014 which were linked to criminal

138 groups fighting over territory (El Tiempo 2014). Other examples of selective and organised assassinations include four dismembered bodies found in September 2015 (Analisis Urbano 2015), eight dismembered bodies found in 2015 and a triple homicide in November 2015, with evidence showing that the victims were tortured before being assassinated (Minuto30 2015b). An official from a research centre (OF1) interviewed in 2014 spoke of how an increase in homicides is directly related to the fighting between the bandas:

‘Something is happening. I have noticed the difference in the homicides in the centre of the city and Castilla [a comuna in Medellín]. They have almost multiplied since last year in Castilla! It could be because of the supposed pacto del fusil, supposed because there is not concrete evidence… But you know what I think is dangerous? When there are ruptures in the pacts, because that means people are shot and this is the worst, because it doesn’t matter who is killed. It is like an act of terrorism, to create terror and show people that they are here…’.

A worker from an NGO (NGO2) interviewed in 2014 spoke of how an increase in homicides was only the ‘tip of the iceberg’:

‘In the last two weeks there has been an increase of homicides in an area called Castilla…This is because there is fighting between three bandas over the control of drug outlets…This shows that every now and then when there are violent episodes there is not an underlying calm, there is not tranquillity and there is not security, there is a subtle control….The armed groups have control over the people in their communities, which is a very violent control…Because they are the owners of the lives of people and they decide who lives and dies on a daily basis…This is an invisible type of violence which cannot really be captured…The homicides are just the tips of the icebergs’.

Given that the criminal actors appear to have changed how they use violence to obtain and/or maintain control of territories, other indicators, such as forced urban displacement might better show their territorial presence and level of control in the city. Forced urban displacement is the practice of threatening residents to leave their territory or face severe consequences such as death and it has been acknowledged by the Medellín government as a more effective ‘mechanism of territorial control [by the criminal actors] as it does not leave the same physical evidence as homicides’ (Alcaldía de Medellín 2011, 7). This practice of violence has increased ‘considerably’ since 2008 with 5395 victims in 2014 according to Medellín Como Vamos (2015, 65). Another finding from Medellín Como Vamos is that despite the reduction in

139 homicides, citizens do not feel safer. For example, the perception of citizens who feel safe in the city has actually decreased from 55 percent in 2012 to 44 percent in 2015 (Medellín Como Vamos 2015). In acknowledging the limitation of relying on homicide rates, alone, to measure levels of violence, an ex-government official observed that:

‘We shouldn’t be proud of having a homicide rate of 25 per 100,000, if we had this rate 35 years ago, we must have been in a very bad place. We have to recognise that we have not arrived, we have only just begun!’ (Field Diary, July 2015).

Despite the reduction of homicides in Medellín there has not been a reduction in the distrust residents have of local authorities. Residents continue to distrust these authorities and are reluctant to talk to them about local security issues because of possible threats to their lives or becoming victims of forced displacement (Abello-Colak 2013, 103). There is also an indication of fear of retaliation from the bandas due to possible collaboration ‘between public security forces and illegal armed groups’ (Franco & Caputo 2013, 21). This possible collaboration was best explained during participant observation by a journalist (Field Diary, May 2015):

‘The police and the bandas live in the same neighbourhood, their kids go to the same school, they go to the same supermarket. The bandas will threaten to kill the families of the police if the bandas think the police are going to arrest them, so they don’t really have a choice’.

These perceptions or knowledge of collaboration between public security forces and illegal armed groups therefore reinforces the failing in the legitimacy of public security forces. This also reduces the likelihood of reporting the criminal bandas to the authorities and also raises levels of impunity (Franco & Caputo 2013, 22). One example of the type of retaliation from the bandas was provided by an NGO (NGO1) worker interviewed in 2014:

‘In the Comuna 13 they dismembered a boy...The mother knew who did it and she made the complaint to the public prosecutor’s office…After making the complaint, she found a note at her house, which said ‘You daughter of a bitch, if you keep making the complaint, we are going to cut up your daughter, your grandson, your granddaughter’...So of course she stopped making the complaint’.

140

The Bandas have Captured the Medellín Miracle In 2015, Latin American policy-makers, urbanists and academics acknowledged progress in Medellín, in terms of the reduction in homicide rates and innovative urban development practices (Abello-Colak & Pearce 2015, 202). However, the data collected for this research strongly indicates that criminal bandas continue to have a presence in the city. If policy makers in Medellín are to address levels of urban violence they should obtain in-depth understanding of the dynamics and processes of the non-state armed groups in Medellin, such as the criminal bandas. Thus, the final section of this data chapter will discuss some of the characteristics and methods of these criminal actors. Given the often-illicit nature of the activities of these groups, and their reluctance to entertain academic researchers, it is difficult to obtain precise information on them. However, if urban violence reduction and prevention policy makers are to effectively address levels of urban violence, they need to have access to valid and reliable information about these groups, information which can only be obtained from researchers taking a context-specific approach and collecting data on-the-ground, as in this thesis.

While there is limited evidence about the criminal bandas, that which does exist suggests that the criminal bandas are hired from ‘charge offices’ (oficinas de cobro) to carry out specific tasks, such as extortion, car theft and local drug dealing. They are then paid by the wider transnational structures of the bandas criminales (BACRIM) and have to report to them (McDermott 2014). These suggestions were confirmed during participant observation in 2015, when an individual with knowledge on the criminal structures commented on how ‘the city is controlled by bandas, the situation is never going to change, there are too many structures in the city and all the bosses of the bandas live in El Poblado [the elite neighbourhoods]’ (Field Diary, July 2015). Two community leaders (CL2 and CL3) interviewed in 2015 provided specific examples of the contemporary nature of the bandas:

‘Today we are living in a situation where legal and illegal activities are combined. Today the group controls the territory and the commercial and local economy… They charge extortion and they control the local businesses like eggs, bread, arepas [a corn cake which is a staple food of the paisas], meat…They capture this local economy and protect it so that other groups do not take control of their territory’. [CL2]

‘Selling mobile phone calls in public, street sellers in the centre of the city are all controlled by the armed groups …Here is one example. A person steals some petrol

141

from a pipe, but this person needs security, they call the security ‘carritos’ who warn you when the police are approaching… This person also needs a driver. All the people involved in this group are paid and they all need access to weapons, just in case someone else wants to take over control of this group.. There are a lot of these types of groups in Medellín…’. [CL3]

A business leader (BL1) interviewed in 2014 also expressed his thoughts on the levels of organisation of these groups:

‘In Medellín there is a mafia that has control. This mafia controls the extortion, micro trafficking, the casinos, the slot machines, chance [a type of lottery]. More than anything they control the extortion. A bus driver has to pay ten thousand pesos [US$3], a shop owner has to do the same… This mafia is very difficult to control, a leader may be captured or killed and then they are just replaced… But you know the police are also involved in this, it is a multi-million-dollar business…’.

An official from a research centre (OF1) even explained how the structures of these criminal groups have changed significantly from the modus operandi of the Medellín cartel in the early 1990s:

‘The structures stay in the city. You can kill the boss of whatever combo and they just reorganise and more and more rise up… It was easier in the 1990s to identify the boss of a particular group because it was Pablo Escobar and his sicarios.. But now these actors have transformed and it is more difficult to identify them.. The only people who can identify members of a combo are the ones who live in the community, but for people who do not live in that community it is very difficult to identify them...’.

As indicated by this official, rather than the vertically integrated and hierarchal organisations of the 1990s with a clearly defined command structure (such as the Medellín cartel), these groups operate in less visible and looser networks to avoid attracting attention from authorities (McDermott 2014). Additionally, members of these bandas are not normally identified as being banda members outside their host community. This has been linked to the choice of these criminal actors to operate in less visible networks.

Research by Abello-Colak & Guarneros-Meza (2014) even showed how the criminal actors have enjoyed success in capturing the innovative social development policies. Their research cited examples of bandas using intimidation to force community leaders to contract their

142 services and also demand a share of grants (Abello-Colak & Guarneros-Meza 2014, 9). A former government official (FGO4) interviewed in 2014, with knowledge on the bandas confirmed this manipulation of participatory budgeting by the bandas:

‘The participatory budgeting was captured by the bandas. The bandas started to be elected on the community boards for participatory budgeting. They [the banda leaders] then used these funds for their own income’.

Local media sources have cited examples of a banda demanding payments (vacuna) from locals to use the outdoor escalators in Comuna 13 (El Tiempo 2012). Any resistance from the community and/or community leaders can lead to threats of violence or assassinations. One example was the brutal assassination of a community leader in May 2015, known for making complaints against the bandas’ use of intimidation in the participatory budgeting process. The assassination of this community leader, along with two other leaders within a month demonstrates the continued policy of swift retribution to silence anyone that speaks out against the criminal actors (El Espectador 2015b).

In 2015, there is no precise understanding as to how many bandas exist in the city or how many individuals are involved. Personnel from NGOs (NGO1 and NGO2) interviewed in 2014 provided their estimates which gives some idea of the scale of the phenomenon:

‘In Medellín we could have about 350 bandas which are related to either the Urabeños or the Oficina…There could be about 13,500 members and that is just in Medellín’. [NGO1]

‘Medellín has a network of violence which is very complex... We could say there are about 200 different violence structures in the city in which between 8,000-10,000 people are involved... These structures have been maintained in the city’. [NGO2]

Members of these bandas continue to be young males from the poorer neighbourhoods. For example, data collected by Baird (2011) in 2009 and 2010 from interviews with former and current gang members found young males from the poorer comunas were ‘seduced’ by the banda lifestyle because any alternative and dignified livelihood opportunities are hindered. Baird (2011) found that for these individuals, there is a perpetual lack of legitimate opportunities and social mobility with half of the members interviewed explicitly referring to

143 limited opportunities born out of poverty as a causal factor leading them to gang membership. This lack of legitimate opportunities in the formal society compared to the social and economic benefits obtained by being a member of banda was emphasised by participants. For example, a community leader (CL3) interviewed in 2015 said:

‘Young people [from the poorer communities] finish school and realise they don’t have access to employment opportunities.. For this reason, young people don’t want to study anymore, there is no motivation to stay in school.. They would have liked to have done something after studying, but they realise that in this society, they can’t….’.

A former government official (FGO3) interviewed in 2015 also spoke of the lack of legitimate opportunities in the formal society and how membership in an armed group provides a mechanism to redress this:

‘If you are 14 years old and you see 50 percent or more of 20 year olds in your neighbourhood unemployed and you realise that the possibility of getting into university is about 1 or 2 out of every 100, which really in your perception is zero… Then this lack of opportunities for many young people is what pushes them into violence… The armed group pays you money and you get some social recognition which you don’t get from society’.

An official from a research centre (OF1) interviewed in 2014 spoke of how membership in an armed group provides a mechanism to redress the lack of economic opportunities for young men from the poorer neighbourhoods:

‘The Secretary of Education did a study about the abandonment of studies and they concluded that there is abandonment because the parents [from the poorer communities] don’t value studying as a way of becoming someone. The parents don’t pressure the kids to go to school to study or continue learning, so that they go to university so that they are able to earn money honestly and become someone in life. There could be some families where they do recognise the importance of working hard, but well, when the son arrives home with a new television or bread or whatever, the mother doesn’t care where the money came from, as the most important thing is that he brought something into the house’.

144

This relationship between structural inequalities and motivation to join the bandas is related to Medellín being one of the most unequal cities in Latin America, with a very high Gini coefficient of 0.506 in 2013 and rising to 0.526 in 2014, despite the increased investment in the poorer comunas (Abello-Colak & Guarneros-Meza 2014, 7). This growing inequality as reflected in the Gini coefficient reveals the limitations of social urbanism, suggesting it was ‘only able to symbolically deliver sensations of social inclusion’ or was just ‘a branding exercise’ (Coupé et al. 2012, 45). There is also high unemployment (around 20 percent) and underemployment rates in the poorer comunas. These high rates depress household income and present significant obstacles to social mobility for individuals from the poorer comunas (Abello-Colak & Guarneros-Meza 2014, 17). Qualitative research from Abello-Colak & Guarneros-Meza (2014) found that the bandas take advantage of this lack of access to employment by offering young males employment with higher wages than the formal sector. These banda jobs range from the local distribution of drugs, to demanding vacunas, or in trades that are often part of broader money laundering operations, such as bakers, private security and credit-loan collectors. The economic benefits offered by being in a banda were commented on by a community leader (CL3) interviewed in 2015:

‘The armed groups have taught the community how to have access to money. But they taught the community a lifestyle in which money is more important than people, than family, than friends… Money gives me everything and gaining this money does not necessarily have to be legal, because if you earn you money legally, you will have to work a lot and will earn very little…’.

An official from a research centre (OF1) interviewed in 2014 also spoke of the economic benefits which could be obtained from membership of an armed group:

‘Money is part of the criminal structures and these salaries are divided between the members of the combo. This money is then used to give food to the family, dress better, give a nice present to your girlfriend… Working in a combo is a job and their salary permits them to have this life. It is easier to be in a combo where your salary is over one million pesos [US$300] rather than killing yourself working in a labouring job where they pay you a minimum salary of six hundred thousand pesos [US$190 a month] … Because from a plaza de vicio [drug dealing corner], the minimum that they will make is four million pesos [US$1200] – which is more than five times the minimum 145

salary! This easy money from the mafia culture has been in the city since the 1970s, when all this narco-trafficking started and we have not been able to stop it…’.

Similar to the comments made by the official from the private institutional organisation, an NGO (NGO1) worker also commented on the high salary offered to banda members and how this compares to the formal market when he said:

‘In a banda there are some people who earn eighty thousand pesos a week [US$240] … Others will earn two million pesos [US$600] or five million pesos [US$1,500] a month. The boss of a large combo will earn up to thirty million pesos [US$9,000] per month… This is without counting their monthly bonuses…’.

Recent research which has attempted to explain the continuing existence of these bandas in the city has focused on how the groups are perceived as part of their communities (Giraldo & Preciado 2015; Casas-Casas & Giraldo 2015; EAFIT-CAP 2014; Abello-Colak & Guarneros- Meza 2014; Abello-Colak 2013; Franco & Caputo 2013). For example, qualitative research from Abello-Colak & Guarneros-Meza (2014, 12) found that residents may resort to the bandas to resolve matters ranging from ‘small quarrels to serious offences’ in homes, neighbourhoods, workplaces or schools and/or rely on them for ‘illegal credit systems.’ When residents were asked why they rely on these criminal actors, they replied that ‘the formal and legal mechanisms are perceived as time-consuming, more expensive, less effective or inaccessible’ (Abello-Colak & Guarneros-Meza 2014, 12). From qualitative research in the poorer comunas, the Medellín-based Centre of Political Analysis, found that residents prefer ‘the bandas existence because of their efficiency and effectiveness in regulating and resolving security and co-existence problems despite the awareness of the negative effects of this relationship’ (Casas- Casas & Giraldo 2015, 70). These negative effects could relate to the criminal actors’ use of violence, including sexual violence, forced urban displacement and extortion. These findings indicate how the residents’ perceptions of inadequate access to government services is taken advantage of by these criminal actors (Casas-Casas & Giraldo 2015, 69). This enables these criminal actors to become integral elements in the daily routines of residents. Examples of this reliance on these criminal actors were also noted during participant observation for this study. Many individuals commented on how they ‘used the services of the Oficina because they would be able to quickly resolve my problem and the police wouldn’t be bothered with my small problem’ or that the community needed ‘mas policia y mas combos’ (more police and more

146 criminal gangs) (Field Diary, May 2014 & June 2015). A community leader (CL3) interviewed in 2015 spoke of the reliance on these groups to resolve conflicts:

‘We [this neighbourhood] have lived through various phases of the violence. We have found that the presence of the armed actors in our community is justified by the same people. The people in our community find in these groups a way of resolving their conflicts’.

An NGO worker (NGO1) interviewed in 2014 said that while the community may be scared of these actors they still rely on them for their daily survival:

‘The criminal actors have a social base in their communities… The community is either scared of these actors or they have some affection towards them… They either live with them or they are obligated towards their control… There are people from the neighbourhood that come and fight against the police when they want to arrest them… These groups provide a large service… They help to regulate everything… It has become normal to live with the two groups. One group helps me to study and another group provides me with security so that I am able to go and study. There are public roads, public services and the other group provides me with security’.

Another NGO worker (NGO2) interviewed in 2014 spoke of the community reliance on these actors for the provision of security. This worker mentioned how the community may not support them and may be scared of them, but that these actors provide services which are integral to their daily survival:

‘Whatever neighbourhood you go to, you will find the police, the local government, the health centre and these armed actors… With actors that have been around for so long, they have a nexus with the community, with the politicians, with the legality. It is almost like an invisible pact. With so much history of illegal actors this means that for many people the security provided (the person controlling the territory) does not matter if they are state forces or paramilitary or militia or a banda related to narco-trafficking. Whatever actor guarantees their security, the principle reason for using these actors is because we feel safe... For example, I was recently in a meeting with a researcher and the researcher was telling the mayor, this model of public investment is getting old, you

147

need to keep it, but also offer other things, because you are fighting against these actors who offer their community services like security, loan sharking, local justice. You have to fight with the state, fight against these sets of services…’.

A government official (GO3) interviewed in 2014 also spoke of the reliance on these actors in providing security services when he said:

‘The armed groups charge about five thousand pesos [US$1.70] and they say ‘while these boys are in the neighbourhood nothing will happen.’ It is almost like ‘more oficinas and more police’ – the oficina has their law and the police has their law. You could go to one of north-eastern neighbourhoods and they will say ‘wow, we live really well, there are no gun fights, there are no homicides, but we are paying for this service...’.

The insights into the characteristics and methods of the criminal bandas in Medellín were all provided by local and international qualitative researchers, including this researcher, who have spent time on-the-ground interviewing an array of local actors in Medellín. These insights show the need for urban violence reduction and prevention policy makers to take into account the processes of non-state armed group formation and maintenance in the reduction and prevention of urban violence.

Conclusion

This chapter discussed the patterns of violence in Medellín between 2000 and November 2015. The discussion showed how the city experienced a 90 percent reduction in homicides from the high rates of 1991 and the city was no longer listed among the ‘50 Most Dangerous Cities in the World.’ This reduction was firstly contextualised within the social urbanism policies implemented by the Medellín government. The discussion showed how these policies were widely cited as the reason for this significant reduction in violence as measured in the reduction in homicides following the implementation of the policies. This was because the social urbanism policies appeared to follow the assumptions the most influential urban violence reduction and prevention approaches make about urban violence and how to reduce it. The discussion then contextualised violence in terms of how the non-state armed groups use violence to obtain and/or maintain control of territories. The evidence showed that despite the significant reduction in homicides, non-state armed groups continue to have a presence in the

148 city and engage in other forms of violence, such as forced urban displacement, to maintain their control of territories. Members of these groups continued to be young poor males who see membership as providing them with social and economic benefits which they cannot obtain from the wider society.

149

Chapter Seven: Explaining the Reduction of Urban Violence in Medellín

Medellín was one of the world’s most violent cities in 1991, but by 2015 Medellín had experienced a remarkable 90 percent reduction in its homicide rate. The main aim of this thesis is to explain the factors that led to this reduction. This chapter provides this explanation by applying the analytical framework identified in the literature review to the empirical data presented in the three empirical chapters. This framework involves firstly drawing from the conflict and development literature to create a multifaceted approach to explain the emergence of urban violence. Secondly, the framework employs concepts used in conflict and insurgency literature to analyse the different non-state armed groups in Medellín. Thirdly, the framework uses the most influential policy approaches for urban violence reduction and prevention to guide the discussion on the reduction of violence in Medellín. The insights provided in this chapter are drawn from the triangulation of primary, secondary and tertiary data collected for this thesis. These insights provide an important contribution to the urban violence literature as they allow a deep and rich understanding of the complexities of urban violence. The chapter utilises the reduction of homicides in Medellín as the prime indicator of the reduction of violence as this is the most widely used measure for assessing levels of violence in urban environments.

The chapter commences with an explanation for the growth of violence in Medellín and an analysis of the policy approaches employed to reduce and prevent the violence. This analysis involves examination and explanation of the assumptions policy makers made about why violence emerged in the city and the motivations of the perpetrators. Both are drawn from the framework in the literature review of explanations of the emergence of violence and conflict in urban environments. In this framework were the grievance perspective, greed theory and the structural violence approach. The analysis will then apply the most influential violence reduction and prevention approaches identified in the literature review - social disorganisation theory, broken windows theory and social capital - to explain the reduction of violence in Medellín. Lastly, the chapter will draw from concepts used in conflict and insurgency literature as identified in the literature review to provide an analysis of the perpetrators of violence focusing on their changing use of violence. This chapter concludes that in accounting for the

150 reduction of urban violence, policy makers must take account of the processes of non-state armed group formation and maintenance and the dynamics of non-state inter-group relations. This conclusion has implications for Latin American urban violence reduction and prevention policy makers, as these policy makers should widen the scope of assumptions they make about the perpetrators of violence.

The Emergence of Violence in Medellín

In 1991, Medellín had a major problem of violence. It was acknowledged as the most ‘dangerous city in the world’ with a homicide rate of 385 per 100,000. All research participants, including former and current government officials, academics, community leaders and members of NGOs, interviewed for this study in 2014 and 2015, spoke of the city’s unregulated growth and the lack of state attention to the new neighbourhoods. The increasing levels of violence were linked to this growth and the deprivations experienced by the recently settled citizens. As one community leader recalled, the migrants that arrived in the city ‘had no way of finding food, of finding housing and they started doing anything to survive’ (Interview CL1, 2015). Chapter Four traced the rapid urbanisation Medellín experienced and how the local government was unable or unwilling to establish services for this burgeoning migrant population. The state failed to provide the new migrants with essential services such as sewage, water, schools and health care (Drummond et al. 2012). The state explicitly neglected these new residents, even routinely demolishing their settlements and omitting consideration of them in planning decisions. Such exclusion and deprivation provided the context in which widespread violence could occur.

Chapter Five identified the establishment of a special Presidential Council for Medellín and the Metropolitan Area (Consejería para el Área Metropolitana de Medellín) as a key event in solving the city’s problem of violence. The purpose of this Council was to understand why the violence had emerged, who were the perpetrators and the best approaches to reduce the violence and to prevent any future outbreaks. Numerous political and other stakeholders, such as community leaders, NGOs, academics and the UNDP, were invited to contribute to seminars and reports to propose solutions to the violence. Their solutions were presented in the report commissioned by the Council called ‘Medellín: Re-encounter with the future’ (Medellín Re- encuentro con el futuro). The report concluded that the main explanation for the high levels of violence in the city was because of ‘structural reasons’, such as high levels of poverty and

151 inequality, high levels of unemployment, particularly youth unemployment, low educational and health levels, high levels of informal housing, with many areas at risk of disaster and a significant lack of public space. All of these factors were related to the mass wave of urbanisation the city had experienced from the 1950s. In forming these assumptions, the stakeholders utilised a grievance perspective, which explained individuals engaging in violence as a result of their neglect by the state (Sambanis 2002, 223; Gurr 1970). They also acknowledged Galtung’s 1969 structural violence, in that the severe inequalities limited or proscribed an individual’s life experiences and thus led to violent conflicts.

However, such perspectives are inadequate on their own to explain Medellín’s experience with rising violence. These perspectives identified an environment that would be conducive to violence but are unhelpful in identifying what triggered the violence and why it spiralled out of control. Insight into the answers to these questions was provided by participants interviewed in 2014 and 2015. These participants recalled how the rise of the illicit narcotics business from cultivation to retail sale, or narco-trafficking, was a crucial reason for the emergence of violence in Medellín. One government official interviewed in 2014 best summarised the effect narco-trafficking had on the city when he said ‘it had an effect on all spheres of life, such as politics, institutions, the private sector and the media’ (Interview GO3, 2014). The ‘Medellín: Re-encounter with the future’ report also concluded that the high levels of violence were related to the prevalence of narco-trafficking. Chapter four of this thesis traced the development of narco-trafficking in Medellín, demonstrating how the business had grown as a result of corrupt officials and businessmen smuggling ‘everything from stolen cars and appliances to emeralds and cigarettes’ from the 1950s (Lamb 2010, 39). This smuggling was transformed by the global expansion of the marijuana trade and the take-over and aggressive protection of smuggling routes. The individuals involved in this business then started profiting by gaining control of coca production and its distribution and taking advantage of Medellín’s breakdown in law and order, a breakdown to which they had greatly contributed. Therefore, violence escalated not only as a result of state neglect but also because of the greed motivations of those individuals involved in narco-trafficking. This perspective on Medellín’s violence uses the greed hypothesis to explain the rising incidence of violence in the city (Collier & Hoeffler, 1998; Collier 2000; Collier & Hoeffler 2000).

From the analysis of primary, secondary and tertiary data, this thesis concludes that there is no one reason or causal factor which can explain the emergence of the high levels of violence in Medellín, but rather there is a need to take a multi-faceted approach. This approach

152 encompasses the structural reasons linked to the rapid and uncontrolled growth of the city and the resulting marginalisation and deprivation experienced by the migrants together with the growth of narco-trafficking.

Policy Approaches to Urban Violence Reduction in Medellín

As identified in the literature review, there is a wide range of policy approaches to contain, prevent and reduce violence in urban environments, with the most influential ones being, social disorganisation, broken windows and social capital. In measuring the success of a violence reduction or prevention policy, policy makers often measure progress in terms of the reduction of homicides. The objective of policy makers in Medellín was to find approaches to reduce the violence and prevent any future outbreaks. A reduction in homicide rates would indicate success.

In Medellín, the alarming escalation of violence, the findings of the Presidential Council and interest among international organisations led to a search for policy solutions to arrest the violence. Thus, the internationally funded policy program, Subnormal District Improvement in Medellín Program’ (Programa Integral de Mejoramiento de Barrios Subnormales en Medellín - PRIMED), was launched with the objective of addressing the structural causes of violence as identified by the stakeholders in the Presidential Council. The PRIMED initiatives included improving public services and public spaces in the poorer neighbourhoods and encouraging state-civil partnerships (PRIMED 1996; Bahl 2012). There were also repressive militarised actions against individuals involved in the Medellín cartel. They included the internationally funded (by the CIA and DEA) Search Block (Bloque de Búsqueda), which had the primary objective of capturing Pablo Escobar, dead or alive. This military-style approach took the fight into the settlements, where the non-state perpetrators of violence lived and which they controlled.

As shown in Chapter Five, these solutions did not initially achieve their objectives. This was because Medellín’s local government did not support the nationally led Presidential Council. This resulted in the ‘fragmented’ implementation of the policies of the Presidential Council in Medellín (Restrepo et al. 2012). Data were also presented to show that these policies to address urban violence encountered a major problem. The activities of the drug trafficking cartel extended their influence beyond those directly involved in the trafficking of narcotics.

153

Interviewees spoke of how Escobar was perceived positively by the inhabitants of the poorer neighbourhoods as his actions compensated for the previous official inaction in the settlements. Thus, a former government official explained how Escobar ‘was seen as an idol’ because ‘he invested in many things for these poorer people’ (Interview FGO5, 2014) while a current government official said ‘he [Escobar] ended up looking like a Robin Hood’ (Interview GO3, 2014). Furthermore, the repressive use of state violence in the 1990s contributed to even higher levels of violence and greater distrust of state authorities, with one NGO official commenting on how the use of violence was perceived by many settlement residents as a ‘type of dirty war from the public forces’ (Interview NGO1, 2014).

Despite the initial disappointment with the results of policies to reduce the violence, the search continued to find the right measures. Chapter Six showed how the involvement of non-political party stakeholders in the Presidential Council provided them with an opportunity to form a coalition and their leader, Sergio Fajardo, to be elected to the position of mayor in December 2003. During participant observation and interviews, many citizens spoke of their respect for Fajardo. They commented on how his administration provided a ‘breath of fresh air’, with many perceiving Fajardo as an honest person who did not follow the clientelistic political patterns of the past (Field Diary, May 2015). The Fajardo administration had an interest in collaborating with professionals from different backgrounds, including academics, community leaders, urbanists and social workers in designing policy. One academic who was involved in the Fajardo administration recalled in 2014 how a ‘more civic government arrived, which allowed a space of trust between different actors’ (Interview A1, 2014). A community leader who was also involved in the administration spoke in 2015 of the respect he had for the Fajardo administration when he said ‘it was about recognising that these community organisations had knowledge’ (Interview CL3, 2015). Fajardo also had support from the influential business elites because of his professional background. For example, a prominent business leader interviewed in 2014 said how ‘Fajardo showed people that without corruption there is money for everything’ (Interview BL1, 2014).

Fajardo ran an inclusive administration in which a wide range of stakeholders contributed to the policy mix. Chapter Six traced the development of these policies, which looked to reduce inequalities through ‘Integral Urban Projects’ (Proyectos Urbanos Integrales - PUIs). These projects involved increasing public investment and extending public services into the poorer neighbourhoods and consisted of improving public spaces, encouraging community involvement and participation in the development of policy, and collaboration between

154 agencies with the objective of raising the welfare levels of the poor and marginalised thereby reducing inequalities (Alcaldía de Medellín 2007). The PUIs soon became known as ‘Social Urbanism’ (Urbanismo Social), which was the umbrella term for investment in the physical infrastructure in the poorer areas of the city. The projects included the construction of the public transport links system to connect the poorer neighbourhoods with the economic centre of the city (the MetroCable) and the library park (parque bibloteca), which provided the community with a library, community centre and childcare facility. The PUIs were seen as a success. Over the period 1997-2007, there was a 13-point increase in the quality of life index in the intervention neighbourhoods, compared to an average of eight points in other districts of the city (Medellín Como Vamos 2007).

Chapter Six also showed how the local administration took concentrated steps to improve the international image of the city which had been severely damaged from over 30 years of high levels of violence and conflict. To pursue this objective, the Fajardo administration was able to gain support from both the wider public and the political elite. For a former Fajardo administration official, ‘it was about doing a lot of work to find an exit from the story people knew about the city’ (Interview FGO1, 2015). The policy initiatives of Fajardo were seen as a success both domestically and internationally and by 2007 the city was perceived as a ‘model for social urbanism’ and an ideal hub for business, tourism and investment (Hylton 2007).

Despite these efforts of the Fajardo administration, when data was collected for this thesis in 2014, Medellín was the most unequal city in Colombia with a Gini coefficient of 0.526. Many participants interviewed in 2014 and 2015, from community leaders to government officials, pointed to these inequalities and how they limited an individual’s opportunities in the wider society. For example, a community leader said that young people from the poorer communities were not motivated to finish school because ‘they would have liked to have done something after studying, but they realise in this society, they can’t’ (Interview CL3, 2014). Thus, many poorer residents of Medellín perceived that the international marketing of the city has been awarded more importance than addressing the day-to-day issues of the poor and marginalised. This sentiment is expressed in the phrases ‘la innovación es maquillaje’ (the innovation is just a cosmetic change) or ‘el alcalde solo quiere maquillar la cuidad’ (the mayor just wants to show off cosmetic changes to the city). These comments could be heard on a daily basis in both the elite and poorer neighbourhoods during data collection in 2014 and 2015.

155

The data presented in chapter Six showed there had also been another social improvement under the Fajardo administration. Most notably there had been a 32 percent reduction in homicides, from 58 per 100,000 in 2004 to 37 per 100,000 in 2007, which translated to almost a 90 percent reduction from the high rates of 1991. The objectives of the policies under the Fajardo administration were not specifically about reducing the violence, although it was certainly a consideration, but because the government had improved the built environment, promoted public spaces, improved participatory citizen planning and exchanges with the public sector, international organisations started to make claims about how these policies had caused the reduction in homicides. These claims were made because the interventions appeared to address the assumptions about the perpetrators of violence made by the most influential approaches to reducing violence. These approaches, as set out in the literature review and analytical framework in Chapter Two, include social disorganisation theory, broken windows theory and social capital, make about perpetrators and levels of violence. All these approaches support investments of the type used by the Fajardo administration to address the deprivations and low levels of social capital that lead to escalating violence. Thus, according to these perspectives, improving the welfare of the poor and marginalised reduces the incidence of violence. The structural impediments to leading a better life are removed and crime, including violence, is reduced.

Claims made by international organisations linked the policy approaches of the Medellín government to homicide reduction and of crime more generally. For example, the UN- Habitat’s Report on Enhancing Urban Safety and Security: Global Report on Human Settlements 2007 compared Medellín very favourably to other Colombian cities when it commented on how ‘a strong mayor and civil society were determined to overcome drug violence and dominance of drug lords’ (Un-Habitat 2007, 38). In 2009, the Inter-American Development Bank argued that the ‘massive infrastructure and social investment’ and ‘modern architecture and green spaces’ caused the homicides rates to reduce (Inter-American Development Bank 2009). In 2013, The Wall Street Journal, Citigroup and the Urban Land Institute even awarded the city the title of ‘The Most Innovative City in the World’ and argued that the construction of ‘public libraries, parks and schools in poor hillside neighbourhoods’ and ‘community organisations, health centres and youth groups… empowering citizens to declare ownership of their neighbourhoods’ had caused the homicide rate to decline substantially. A specialist in security and development, Robert Muggah (2014) also advanced the same argument. He said: ‘Social cohesion matters. Mobility matters in our cities. We’ve

156 got to get away from this model of segregation, exclusion and cities with walls. My favourite example of how to do this comes from Medellín.’ These claims were reiterated in a 2015 Journal of International Affairs publication when he discussed successful crime prevention policies and said ‘the most stunning case is Medellín… A succession of mayors…turned things around by devoting more attention to tackling the poorer and most dangerous comunas’ (Muggah 2015, 30). These statements reveal a dominant discourse that drew its explanations for the reduction of violence in Medellín from elements of social disorganisation, broken windows and social capital.

Quantitative studies cited earlier also argued that events in Medellín provide support for social disorganisation theory and broken windows theory. For example, research by Colombian and US epidemiologists (Cerdá et al. 2011) found the decline in homicides was 66 percent greater in the intervention neighbourhoods than in control neighbourhoods and thus, they concluded that physical and social integration of informal urban neighbourhoods can reduce criminal acts. Their findings have been widely cited, notably in the Inter-American Development Bank Report (2015) Closing Knowledge Gaps: Toward Evidence-Based Crime Prevention Policies in Latin America and the Caribbean (Inter-American Development Bank 2015). A more recent study by Medellín-based researchers at the researchers in spatial economics group and Spanish researchers from the geo-environmental cartography and remote sensing group, gives further support to the argument linking the declines in crime and homicides to the social urban policies of the Medellín government (Patiño et al. 2014). They tested ‘which land cover, structure and texture descriptors were significantly related to intra-urban homicide rates in Medellín, while controlling for socioeconomic confounders.’ They argued that their findings were ‘expected and supported by social disorganisation theory’… ‘and broken windows theory’ by concluding ‘…areas with higher homicide rates tended to have higher local variation and less general homogeneity…’ (Patiño et al. 2014, 48).

During data collection in 2014 and 2015, the current Medellín government, of Gaviria, added to the chorus singing the praises of successive government policies and their causal relationship to the reduction of crime and violence. Government officials argued that the investment in the poorer neighbourhoods, the improvements to the built environment, the promotion of public spaces and improvements in state-civil relations were the integral reasons why Medellín was ‘experiencing the lowest homicides rates in 35 years.’ One government official interviewed summed up this view when stating that ‘addressing inequalities, improving access to education, capacity building, development of essential infrastructure, improving public transport and

157 social inclusion [led] to the declining rates of homicides’ (Interview GO2, 2015). There were numerous government-sponsored advertisements throughout the city celebrating that the reduction in homicides was the ‘first success of [the government’s] 100 successes’ (Hecho No 1 de 100 hechos).

Policy analysts know it is hard to identify, let alone quantify the impact of policies with a reasonable degree of accuracy, but the international organisations and researchers cited above claimed the social urbanism policies caused the reduction of homicides because the homicide rate declined following the implementation of the policies. There is certainly a temporal link with the reduction in homicides occurring at the same time as the various social urbanism initiatives. However, there is limited to no analysis elucidating the causal links to demonstrate that the policies did in fact cause the reduction. There are even hints that other forces may be at work. For example, Muggah’s (2015) work seems to be paradoxical. On the one hand, he argued that the policies caused the reduction in homicides and on the other hand, he provides evidence from residents who perceive the homicides have reduced because of a pax mafiosa or pax criminale (Christian Science Monitor 2014; The Guardian 2014). We will return to this matter later.

There is also a methodological issue relating to how urban violence can be measured. The quantitative studies are limited in their explanatory value because they relied on homicide rates, alone, to measure levels of violence. In the literature review, I identified the limitations of using these objective indicators alone because of their reliance on the accuracy of institutions in recording homicides, the lack of attention to other practices of violence and also because they fail to take into account citizens’ perceptions of violence. Muggah (2012, 62) also acknowledged such limitations when discussing interventions, which have been based on assumptions made by social disorganisation and broken windows approaches, when he wrote (in a footnote) ‘these measures [homicide rates and HDI ranking] were insufficient to capture more subtle changes in safety and security, particularly in marginal areas where data collection capacities were weakest.’ Thus, when scrutinising the claims made by international and national researchers about Medellín and the reduction of violence, researchers often confuse correlation with explanation. They note the temporal correlation between the implementation of the social urbanism policies and the reduction of homicides but generally fail to explicitly link the two phenomena through causal analysis. Data collected for this thesis strongly indicate that these claims, which relied on assumptions made by social disorganisation, broken windows

158 and social capital approaches do not tell the whole story of the over 90 percent reduction of homicides from 1991 to 2015 for two reasons.

The first reason why these policies do not tell the whole story is because the policies were devised on certain assumptions about cause and effect in human behaviour. While the policy makers did not explicitly state they were drawing on particular theoretical perspectives, they do appear to have borrowed from social disorganisation, broken windows and social capital approaches to violence reduction and prevention. However, as identified in the literature review, these approaches have been developed from research in Western Europe and North America. In these environments, there is an assumption of more trusting relations between citizens than may be the case in some developing country cities. In Medellín these same assumptions cannot be made. The literature review identified how in societies with high levels of violence or a history of violence there is a climate of fear and distrust (World Bank 2011a). Medellín has this climate of fear and distrust, which has stemmed from its history of violence. Examples of this distrust were provided in the data chapters of this thesis. During data collection in 2014 and 2015, I observed the perception the middle and upper class had of the dangers of the poorer neighbourhoods, despite not visiting them in recent years, if ever. These individuals were not interested in visiting the iconic improvements to the built environment in these neighbourhoods, for which the city had become so famous (Field Diary 2014 & 2015). Incorporated into social disorganisation, broken windows and social capital are assumptions of general trust in a broader institutional framework, such as the government and local authorities. In Medellín, this assumption cannot be made. The data chapters provided examples of the distrust citizens have of the government and local authorities. For example, given the history of state violence in the poorer neighbourhoods, residents have generally only experienced the state as an agent of repression. In 2014, Medellín-based researchers found from qualitative and quantitative studies that citizens (from all sectors of society) continue to perceive the police as ‘corrupt, ineffective, abusive and do not trust them’ with the police even acknowledging these problems (CAP-EAFIT 2014, 31; Abello-Colak & Pearce 2015, 209).

The second reason why the social urbanism policies do not tell the whole story about the reduction of homicides in Medellín is because of the assumptions the policy makers made about the perpetrators of violence. The literature review identified the difficulties of obtaining reliable information on perpetrators of violence generally. Such difficulty applies in Medellín. In Medellín, policy makers assumed individuals engaged in violence in response to their neglect by the state. The Fajardo administration set the agenda of addressing the ‘historical social debt’

159 owed to the poorer neighbourhoods, assuming that addressing this ‘social debt’ would reduce the violence. For example, an academic interviewed in 2014, who was involved in the Fajardo administration, spoke of how the social urbanism policies ‘help[ed] in making violence more difficult. It was about ensuring a young person would have other alternatives’ (Interview A1, 2014). Evidence was also provided from Medellín-based researchers who, in 2015, ‘intuitively suggested two out of every three homicides correspond to diffuse spontaneous violence and citizen intolerance’ (Giraldo & Preciado 2015, 11). While these researchers did not provide an explanation of ‘spontaneous violence’ and ‘citizen intolerance’ they are assumed to be initially non-violent conflicts which turn into violent acts, such as disputes noted on Mother’s Day, the most violent day in all of Colombia. One NGO worker interviewed in 2014 spoke of these levels of violence and reasons for it when he said ‘People drink a lot and there are a lot of street fights and fighting in families and since there is both a large presence of guns and knives, people end up stabbed and die as a result.’ A community leader spoke of these intolerances after soccer matches when he said ‘they leave the game and attack each other and end up being stabbed, just because they are wearing the shirt from the other team…’ (Interview CL3, 2015).

However, Mother’s Day is only one day a year and soccer matches occur on a weekly basis and only during the season. On other days of the year, many homicides do not support the assumption of the perpetrators being linked to spontaneous violence and citizen intolerance. For example, homicide victims included four dismembered bodies found in September 2015 (with eight dismembered bodies found in 2015) and a triple homicide in November 2015, with victims found with evidence of torture before they were assassinated (Minuto30 2015b; Analisis Urbano 2015). These victims indicate that the perpetrators are related to organised violence, and not the spontaneous variety. The assumptions policy makers in Medellín made about the perpetrators of violence did not acknowledge that such individuals could act rationally to make decisions that provide them with the greatest personal benefits, which could be manifested in making profit through involvement in organised crime, such as narco- trafficking, or being gang members. The policy makers did not acknowledge that the behaviour of these individuals might be independent of social urbanism policy and be subject to other influences that can play a role in changing their behaviour. The next section of this chapter will analyse these individuals, focusing on the following questions: Who are they? What are their methods? Have their methods changed? What have been the motivations for these changes? How have these changes impacted on the success of their activities? How does this ultimately contribute to explaining the reduction of violence (homicides) in Medellín?

160

The Perpetrators of Violence (Homicides) in Medellín

This section analyses the perpetrators of violence in Medellín, with particular emphasis on how patterns of violence (homicides) have changed over the period studied. As was shown in the data chapters, there is a complex history of the rise and fall of the groups, which all used violence to maintain and/or obtain territorial control and often to engage in crime, notably narco-trafficking. Government officials and the media have generally classified these groups into three types: criminal gangs (bandas), left-wing militias and right-wing paramilitaries. Thus, this first section analyses these different groups, focusing on their use of violence to maintain and/or obtain territorial control. The second section then analyses if a change in how the groups use violence to maintain and/or obtain territorial control can explain the reduction in violence in Medellín.

Chapter Four traced the emergence of the criminal bandas. These groups worked for the hierarchically-structured, Medellín cartel, which was led by Pablo Escobar. An individual’s employment in the cartel ranged from working as a bodyguard to being hired as an assassin (sicario), with their victims often political and social leaders who criticised the Medellín cartel. By 1990, there were an estimate 150 different bandas in Medellín. Of this number, 45 were linked directly to the cartel and the remainder controlled the local drug market and used homicidal violence to maintain and/or obtain territorial control (Salazar & Jaramillo 1992; Bedoya Marin & Jaramillo Martinez 1991, 93-97). Following the removal of Escobar in 1993, practices in and among bandas changed. The first change was that some former banda members joined the militia groups as membership in such groups provided them with status and possibilities for upward mobility, which they could not obtain from the wider society (Melguizo & Cronshaw 2001, 115). Banda members who did not join the militia started working as ‘contractors’ for the criminal organisation, the Oficina de Envigado, led by Don Berna. This Oficina was a ‘network’ which connected customers to the Oficina’s violent services, such as selective assassination and kidnapping, and charged a ‘finder’s fee’ (Hylton 2007, 84; Lamb 2010, 168; Gutiérrez & Jaramillo 2004, 27). The bandas continued to use homicidal violence to obtain territorial control. Invisible borders (fronteras invisibles) marked out a group’s territory and the death penalty was enforced for any individual crossing into another’s territory. However, banda members had also acknowledged the importance of needing the support of the community to maintain territorial control and thus they offered

161 security and social services to residents (Gutiérrez & Jaramillo 2004, 25; Melguizo & Cronshaw 2001, 125). Banda members continued to be employed by Don Berna, a major leader of these groups, even when he joined the right-wing paramilitary group, the AUC, in the late 1990s.

Chapter Four also traced the emergence of the left-wing militia in Medellín. These groups emerged around 1984, as a result of national negotiations with the left-wing guerrilla group, the M-19. These negotiations had brought Peace Camps (Campamentos de Paz) to Medellín (and other major cities) as a way to build popular support for the M-19 for when it would become a legitimate political organisation once the peace agreements were concluded (Lamb 2010, 86; Gutiérrez & Jaramillo 2004, 21). However, just nine months after the peace talks, the peace camps were outlawed and some, but not all, participants were arrested. Given that not all participants were arrested, ‘outlawed’ militia groups remained in Medellín and provided their host communities with ‘services’ such as confronting, and often socially cleansing, banda members and common criminals threatening the security of their communities (Gutiérrez & Jaramillo 2004, 21-22). This provision of service meant that the militia were generally supported by the community, with one official describing them as ‘saviours’ (Interview FGO2, 2014). By the early 1990s, the militia were broadly distinguished between groups linked to the M-19 and groups supported by the ELN, and other independent groups with no political affiliation. In the early 1990s, militia membership increased and the militia expanded throughout the city. However, with this expansion came increased territorial conflicts between the militia groups. These conflicts arose because ex-banda members had joined the militia and these ‘new’ members were more violent than the ‘old’ ones. Victims of social cleansing by the militia now included drunks, bad sons, bad parents and homosexuals, with violent acts against such persons losing the militia the support of their host communities (Lamb 2010, 117; Gutiérrez & Jaramillo 2004, 21). The ‘new’ members also increasingly became involved in narco-trafficking, which escalated conflicts between the different groups not only because of territorial conflicts but also because not all militia members supported this involvement in narco-trafficking (Lamb 2010, 114). These conflicts were noted particularly before, during and after the 1994 Media Luna negotiation between the different militia groups in Medellín. Throughout the 1990s, scattered militia groups remained in the city with links to the left-wing FARC and ELN.

The final group that used homicidal violence to obtain and/or maintain territorial control in Medellín was comprised of the right-wing paramilitaries. On a national level, these

162 paramilitaries had the primary objective of countering left-wing guerrillas. By the late 1990s, most of the right-wing paramilitary groups throughout the country were disunited but the United Self-Defence Forces of Colombia (Autodefensas Unidad de Colombia - AUC) united them. The AUC was commanded by the Castaño brothers and was made up of different groups or blocks across the country, to provide a ‘civil-defence service’ against the left-wing guerrillas (Lamb 2010, 179). By 1997, the media and government reported that the AUC had entered Medellín under the Metro Block (Bloque Metro - BM), led by the man who used the alias Doble Cero. The BM targeted communities in Medellín where residents were complaining about the ‘abuses’ of the militia and often engaged in ‘bloody social cleansing campaigns’ against them (Verdad Abierta 2016; Lamb 2010, 183; El Colombiano 2002a). By 1999, there were two blocks of the AUC in Medellín, the BM and the Block Cacique Nutibara (Bloque Cacique Nutibara) (BCN), with the former led by Don Berna. Both the BCN and the BM used homicidal violence to obtain and/or maintain territorial control in removing the left-wing militia from the city.

At first, there were clear distinctions between the criminal bandas, the left-wing militia and right-wing paramilitaries. However, as shown in the data chapters, their practices and even their identities became ‘blurred.’ The bandas, the militia and the paramilitaries all provided their host communities with security services. While the community may have feared, rather than supported, the group which controlled them, the armed groups did provide social and security services which the state was unable or unwilling to provide for the communities. This provision of services was also beneficial to the group as it meant they could rely on the loyalty of their host community. Escobar went so far as to provide the poorer neighbourhoods with social services such as houses, soccer fields and schools. Given these valued gifts, the local newspaper, Semana (1983) described Escobar as ‘Antioquia’s Robin Hood’ and ‘one of the most popular personalities in Antioquia.’ A former government official recalled how Escobar was seen ‘as an idol’ (Interview FGO5, 2014) and another how ‘he became a local popular hero’ (Interview GO3, 2014). Given this provision of services, Escobar was able to rely on the support of residents from the poorer communities during his conflict against the Colombian state. As shown in the data chapters, similar practices were noted with the militia and the paramilitary.

Another similarity between the militia, the paramilitaries and the bandas was the characteristics of their members. As shown in the data chapters, members were largely young males from the poorer neighbourhoods. These young males joined a group because of the social and economic

163 benefits membership provided them. Benefits included providing ‘a form of structure’ to their lives (Interview FGO4, 2014), being perceived as ‘someone important in the community…with a sensation of power’ (Interview CL3, 2015) or as a ‘dandy’ (Interview CL2, 2015). Examples were shown throughout the data chapters to illustrate how membership provided these young males with a mechanism of informal socialisation and that they were more attracted to the benefits this membership offered them rather than the political ideologies of a particular group. For example, Baird (2011) found from 40 life-history interviews with youths who sometimes called themselves gang members and sometimes paramilitaries, not one interviewee claimed they joined gangs for political purposes. The changing allegiances of group members also indicated their attraction to the benefits that were provided to them rather than any attachment to specific political ideologies. For example, in the early 2000s, the November 6 and 7 Militia (Milicias 6 y 7 de Noviembre) (M-6&7), affiliated with the ELN, changed their allegiances to the right-wing BM, with one interpretation for this change being because ‘the urge was always to have a motorcycle, to get girls, to be in a gang’ (Lamb 2010, 195).

The ‘blurred practices’ of the criminal bandas, the militia and the paramilitaries were also noted by research participants during data collection. Interviewees interchangeably used labels such as criminal gangs (bandas), delinquent gangs (combos), the office (la oficina), self-defence groups (autodefesanas), paramilitaries (paramilitares), the mafia (la mafia), demobilised paramilitaries (neo paramilitares), armed illegal groups (grupos ilegales armados) or the armed actors (los actores armados) to label the previous and current perpetrators of violence in the city. In some interviews participants even used all of these labels when making reference to one group. This, showed that for residents it was not important to classify these groups into specific categories, but, rather there have been and continue to be simply groups which engage in organised acts of violence to obtain and/or maintain territorial control.

Changing Patterns of Violence in Medellín

This section analyses how the rise and fall of the different non-state groups, the militia, the paramilitaries and the criminal bandas, contributed to the changing patterns of violence in the city. This section also analyses how a change in the manner in which these groups used violence to obtain and/or maintain territorial control contributes to explaining the 90 percent reduction of homicides Medellín has experienced.

164

In 1991, Medellín was classified as the ‘most dangerous city in the world.’ Chapter five showed how these high levels of violence were largely linked to conflicts the Medellín cartel had with the Cali cartel and with the Colombia state. These conflicts resulted in high levels of selective homicides against political, social and judicial leaders, as well as state violence to weaken the criminal bandas (Lamb 2010, 60; Human Rights Watch 1994, 31; Riaño Alcalá 2006, 40). Following the death of Escobar in 1993, homicide rates did decrease. However, throughout the 1990s, the city still had a serious problem of violence, with an average homicide rate of 266 per 100,000. These high rates were because of conflicts between the militia and bandas to obtain control over territories in the city. In 1998, there was a short-lived reduction to 156 per 100,000 (from 184 per 100,000 in 1997), but homicides increased to 166 per 100,000 in 1999. This increase was largely linked to the entry of the paramilitaries to Medellín, who also used homicidal violence to obtain territories.

From 1999 to 2002, rather than decreasing, homicides continued to increase (from 166 per 100,000 to 179 per 100,000). This increase was a result of three conflicts between the non-state armed groups in the city. The first conflict was between Banda La Terraza and the BCN, which started around 1999. This banda had previously been employed by Don Berna to deliver services, such as selective assassinations and kidnappings, and there are inconclusive reasons as to why the conflict started. More importantly, this conflict resulted in high levels of violence including two car bombs in elite neighbourhoods of Medellín in 2001, with Banda La Terraza being the number one suspect (Semana 2001b; Semana 2014). By 2003, La Terraza and other rival bandas were either defeated by the BCN or they agreed to work for Don Berna under the Oficina de Envigado. Another conflict in the city around this time was between the militia and the AUC. The majority of the militia (namely the Peoples Armed Commandos (Comandos Armados del Pueblo - CAP) were present in the north-eastern neighbourhood, the Comuna 13. The conflict between the AUC blocks (the BM and the BCN) and the CAP had escalated by 2002, with 442 homicides in the first six months in this comuna alone. They included targeted assassinations of community and religious leaders (Martin 2014, 445). There were even two military operations, Operation Mariscal and Operation Orion, to remove the militia from the Comuna 13, with the latter seen as a success when the militia left the city shortly following the 2002 military operation. However, the power vacuum left by the militia was soon filled by the BCN. The third reason for the increase in violence was because of the conflict between the two AUC blocks in Medellín, the BCN and the BM. The conflict had started when Don Berna became involved in the AUC. The leader of the BM, Doble Cero, disliked Don Berna’s

165 involvement in the AUC, as Doble Cero refused to allow narco-trafficking to corrupt his work and was perceived as the ‘purest of the political paras’ (Lamb 2010, 182; Gutiérrez & Jaramillo 2004, 27). Given that the leaders of the AUC obtained 70 percent of their revenue from narco- trafficking, they provided support to the BCN to defeat the BM which they did around 2003 and Doble Cero was assassinated in 2004 (Amnesty International 2005, 8).

By late 2003, there was no more territory to be conquered in the city. The militia had been removed, the BM had been defeated by the BCN and the majority of the criminal bandas worked for Don Berna in the Oficina de Envigado. Given that Don Berna was the leader of the BCN and the Oficina de Envigado he had obtained control over all territories and thus there was no need for territorial conflicts. Given that there was no need for territorial conflicts, homicidal violence fell dramatically from 2002 to 2003 (179 per 100,000 to 98 per 100,000). This reduction was mentioned by research participants during data collection with one official recalling ‘if you do not have anyone to fight with, then the homicides will reduce’ (Interview NGO1, 2014) and another recalled how ‘even though there were different structures of criminal groups, there was a control of the city and this was because of the pacification of the city’ (Interview NGO2, 2014). The pacification that was so extensive that some traded the term, governability, for ‘Donbernability’ (Donbernabilidad) as Don Berna managed the illicit networks so closely that not one assault, robbery or assassination could take place without his permission (Semana 2005; Lamb 2010, 238). This reduction in homicides continued, reaching a low of 37 per 100,000 in 2005, largely attributed to the lack of territorial disputes between the non-state groups with a capacity for violence.

However, this reduction was only short-lived, as homicides started to increase again, and by 2009 homicides were almost double those of 2008 (49 per 100,000 to 94 per 100,000). Insight into this increase was provided by research participants who recalled how following the extradition of Don Berna to the US in 2008, for narco-trafficking charges, there was no overarching control of the groups with violent capabilities in the city. Thus, one NGO worker recalled how ‘shooting started again in Medellín, because these criminals were left without a defined structure, without a leader. There was no one to pay their salaries’ (Interview FGO4, June 2014) and another official recalled ‘it [the extradition] had left these structures free to do whatever they wanted’ (Interview OF1, 2014). Conflicts started within the Oficina de Envigado for a new leader and a new group, Los Urabeños, used this power vacuum as an opportunity to expand their growing drug trafficking network with both groups using homicidal violence to obtain and/or maintain territorial control. The extremes of these conflicts were noted with the

166 massacre of eight individuals in July 2010 with the Oficina de Envigado the number one suspect and the massacre of nine Oficina de Envigado members in December 2012 (Semana 2010; Insight Crime 2013; El Tiempo 2013).

By 2013, homicide rates had reduced again. They had returned to those of 2007 (38 per 100,000) and by 2015 the city was experiencing the ‘lowest homicide rates in 35 years.’ Insight into this reduction was provided by research participants. These participants spoke of the ‘business model’ of the Oficina de Envigado and Los Urabeños and explained how from around July 2013 they had a new model. This new model was based on cooperation between the two groups, rather than using violence for territorial control. It was often referred to as pacto de fusil. Its effect on the groups with capacities for violence was clearly explained by one community leader:

‘No one is permitted to assassinate people and if they do, they must do it outside of the neighbourhood or outside of Medellín or dismember the body and throw them into the Medellín river. Importantly, there is no authorisation to kill in the neighbourhood or in Medellín. You have to ask permission from the armed group to assassinate someone’ (Interview CL2, 2015).

This pacto de fusil agreement has been acknowledged by some Medellín-based researchers and media outlets (Semana 2013a & 2013b; Jaramillo & Gil 2014, 146; Gonzalez 2014, 99; McDermott 2014). The immediate reduction of homicides from 90 in July 2013 to 60 in August 2013 and a 30 percent reduction from 2012 to 2013 adds further weight to this argument.

Insight into the change of ‘business model’ of the Oficina de Envigado and Los Urabeños was provided by research participants. These participants explained how homicidal violence draws the attention of the authorities to the perpetrators and this reduces the effectiveness of the perpetrators business, i.e. maximising their profits from their criminal activities, such as narco- trafficking. For example, one government official said ‘it was not convenient for them [the criminal actors] to have confrontations. They need to run their business [narco-trafficking] effectively’ (Interview GO3, 2014). Another official said ‘they [the criminal actors] have understood that if they use such obvious violence [such as homicides], then it will call attention to their presence…’ (Interview NGO2, 2014). When participants did explain the change in the ‘business model’ of the Oficina de Envigado and Los Urabeños they did not link this change to the policies instituted by the Fajardo administration and/or subsequent ones. This strongly suggested that the decisions and actions of these non-state actors were largely independent of

167 the social urbanism policies that others have claimed as being responsible for urban violence reduction.

Another theme that emerged from the triangulation of data was the ‘pockets of insecurity’ throughout the city. For example, the third data chapter discussed the advances in the collection and analysis of homicide data in Medellín. These improvements provided the tools for local researchers to map homicides throughout the city. This mapping revealed that some comunas had homicide rates higher than ‘the most violent states in Mexico’ and others ‘lower than Boston, US’ (Giraldo & Preciado 2015, 9). While these researchers were unable to explain why these differences, they did provide important baseline statistical information for urban violence reduction and prevention policy makers. However, during data collection, interviewees also commented on how even though a particular area of the city has experienced a reduction in homicides, and thus ceases to be a problem for local authorities, this reduction may only indicate that a non-state armed group has obtained control of the territory. For example, a former government official said ‘there are places that are not on the police radar [because of low homicide rates] and then they explode [increase in homicides]’ (Interview FGO5, 2014) while another stated that ‘every now and then when there are violent episodes [it shows] there is not an underlying calm…there is a subtle control…The homicides are just the tips of the icebergs…’ (Interview NGO2, 2014). Looking at the submerged parts of the icebergs reveals other practices of violence to maintain control, such as forced urban displacement. This practice of violence involves threatening residents to leave their territory or face severe consequences, such as death. This practice of violence has increased considerably since 2008 with 5,395 reported victims in 2014 and the Medellín government has recently acknowledged this practice is more ‘effective’ than homicidal violence as it does not leave ‘physical evidence’ (Medellín Como Vamos 2015, 65). There are also selective assassinations against ‘rule breakers’ such as the four dismembered bodies found in September 2015 and a triple homicide in November 2015, with victims found with evidence of torture before they were assassinated (Minuto30 2015b; Analisis Urbano 2015). There are also selective assassinations against individuals who criticise the control these bandas have, such as the assassination of three human rights leaders in less than a month in 2015, including one of the potential participants for this research, who was shot 25 times (El Espectador 2015a & 2015b). The insights provided through this analysis of data show the importance of conducting on-the-ground research and interacting with a wide range of actors when studying patterns of urban violence. Government officials used the reduction of homicides to justify the success of the social urbanism policies, but on-the-ground

168 research, such as the data collection methods used for this thesis, provide a more in-depth and localised explanation of changing patterns of homicides. Not only has the change in the ‘business model’ of the Oficina de Envigado and Los Urabeños led to a reduction of homicides but also their practices address some of the limitations of the social urbanism policies. While the social urbanism policies were identified by the local government as being successful in addressing inequalities, Medellín continues to be regarded as one of the most unequal cities in Latin America. Research participants identified that unemployment and underemployment in the poorer neighbourhoods’ limit residents’ social mobility, which can have the effect of ‘pushing’ them into the employment opportunities offered by the bandas. For example, one former official said ‘the armed group pays you money and you get some social recognition which you don’t get from society’ (Interview FGO3, 2015) and a current official said ‘it is easier to be in a combo where your salary is just over one million pesos [US$300] rather than killing yourself working in a labour job where they pay you a minimum salary of six thousand pesos [US$195 a month]’ (Interview OF1, 2014). Research participants estimated there were ‘between 8,000-10,000 people…’ (Interview NGO2, 2014) or ‘about 13,500 members’ (Interview NGO1, 2014) involved in the bandas and providing ‘employment opportunities’ for the young males in the poorer neighbourhoods. Therefore, while the community may fear, rather than support these actors, they do furnish them with services which residents perceive the state is unable or unwilling to provide.

Conclusion The main objective of this research was to identify the explanation for the over 90 percent reduction of homicides Medellín experienced from 1991 to 2015. This chapter has provided this explanation by applying the analytical framework identified in the literature review to the data presented in Chapters Four to Six. The first part of this chapter analysed the policy approaches used in Medellín to reduce or prevent violence. The policy makers assumed violence emerged for two reasons or causes. First was local grievances related to the massive wave of urbanisation the city experienced and second, individuals found they could profit by gaining control of the trafficking of narcotics and take advantage of the breakdown in law and order. This thesis concludes there is no one reason or causal factor which explains the emergence of the high levels of violence in Medellín, but rather there is a need to take this multi-faceted approach.

169

The second part of the chapter analysed the policy approaches to reduce the violence. It showed how social urbanism investments were used to address the deprivations and the low levels of social capital that are held to lead to escalating violence, according to the assumptions of social disorganisation theory, broken windows theory and social capital on urban violence reduction and prevention. These assumptions led to the initiatives by the Medellín government to address deprivation, marginalisation and social inequality. They were widely perceived as being successful and having the effect of reducing violence. However, given the assumptions these approaches make about cause and effect in human behaviour and the perpetrators of violence, it was concluded they are limited in explaining the over 90 percent reduction of homicides. The insights provided in this chapter, and throughout the thesis, make an important contribution to the urban violence academic literature and policy-focused analysis on urban violence. The findings of this research shows the importance of taking into account both objective indicators, such as homicide data, and subjective indicators, such as perceptions of insecurity, in measuring the success of urban violence reduction and prevention policies. These findings highlight the importance of studies on urban violence taking a context-specific approach and researchers spending time on-the-ground collecting data, such as was done for this thesis. The final part of the chapter analysed the perpetrators of violence and revealed additional important factors leading to the reduction of violence (homicides) over the period studied. These factors concerned the changing dynamics of inter-group rivalry and control of territory and illegal business in Medellín. This analysis showed how the non-state groups in Medellín - the militia, the paramilitaries and the criminal bandas - all used homicidal violence to maintain and/or obtain control of territories and often to engage in crime, notably narco-trafficking. This inter- group rivalry contributed to the changing patterns of violence in the city over the period studied. It was shown how a more plausible conclusion in explaining the over 90 percent reduction of homicides in Medellín must consider changes in how these groups use violence to maintain and/or obtain control of territories. The motivation for the reduction in violence can be interpreted as a strategy for the perpetrators to increase the effectiveness of their criminal activities, notably narco-trafficking, and was largely independent of the social urbanism policies that others have claimed as being responsible for the reduction of urban violence in Medellín. These conclusions have implications for Latin America urban violence reduction and prevention policy makers. These policy makers should widen the scope of assumptions they make about perpetrators of violence. Homicides in Medellín have largely been related to conflicts between non-state armed groups over territorial disputes. These non-state armed groups are not disorganised delinquents who engage in ‘spontaneous acts of violence’ but 170 rather they are organised and adopt a rational approach as to how they use violence. The members of these groups may come from the poorer neighbourhoods but they are integrated into larger organised networks whose business is narco-trafficking. These findings therefore show how in accounting for the reduction of urban violence, policy makers must take account of the processes of non-state armed group formation and maintenance, the dynamics of non- state inter-group relations and the relations between state and non-state groups.

171

Chapter Eight: Conclusion

What explains the reduction of violence in an urban environment? This was the question set out at the beginning of this research to examine through the case study of the over 90 percent reduction of homicides Medellín has experienced from 1991 to November 2015. Given the decline of internal armed conflicts and the perceived rise of urban violence, the latter has increasingly attracted attention in academic circles, particularly in the last 30 years. There is considerable agreement in this literature on the adverse impacts, costs and consequences of urban violence. There has also been considerable testing and analysis of explanations of violence in urban environments in high-income countries. However, only in the last 20 years has research turned to better understanding violence in the cities of middle- and low-income nations. Generating the evidence base on violence in these environments is important because of the predicted growth of these cities in the coming years and that they could become the future sites of national insecurity (Patel & Burkle 2012; Kilcullen 2013; Moser & McIlwaine 2014; Winton 2014; Kalyvas 2015; De Boer & Bosetti 2015; Muggah 2012). High levels of violence and conflict are already being noted in Latin American urban environments. The region is currently the world’s most violent one, with only 9 percent of the world’s population but 33 percent of global homicides and, in 2015, 47 out of the world’s 50 most dangerous cities (Muggah & Giannini 2015).

This thesis makes a significant contribution to the growing literature on urban environments in middle- and low-income countries through the analysis and explanation of the reduction of urban violence through the empirical case of Medellín, Colombia. This case study was selected for this analysis and explanation for three reasons. First, no other city in the world has experienced such critical and high levels of violence outside of a civil war. Second, Medellín represents an extreme case of complex perpetrators of violence including drug cartels, gangs, youth assassins, urban guerrillas, community militia, and paramilitaries. Lastly, no other city has undergone such a dramatic reduction in homicides in less than a generation. Medellín experienced over a 90 percent reduction in homicides from 1991 to 2015. International organisations, such as UN-Habitat, Inter-American Development Bank and the World Bank have praised the policy approaches used in the city for this reduction of homicides (UN-Habitat 2007; Inter-American Development Bank 2009; World Bank 2011a). After data collection was finished for this thesis (post-November 2015), the city continues to be seen as a benchmark for

172

Latin American urban violence reduction policy. For example, in June 2016, the World Bank stated how Medellín ranks as one of the ‘most liveable and innovative cities in Latin America’ and how the security situation had improved because of ‘the dismantling of violent drug cartels and a series of innovative urban and social development projects that helped integrate impoverished hillside neighbourhoods with the rest of the city’ (World Bank 2016b). In July 2016, the city was awarded the most significant award given to cities for planning, urban renewal and transformation, the Lee Kuan Yew World City Prize. In awarding the city the prize, the Chairman of the Nominating Committee, spoke of how Medellín can ‘become a mecca of learning’ for cities in developing nations because of its transformation from one of the world’s most dangerous cities into ‘a liveable and innovative one’ (Lee Kuan Yew World City Prize 2016). To provide theoretical and conceptual context, Chapter Two reviewed contemporary literature that dealt with urban violence. This included leading explanations of the emergence of violence in urban environments. It was shown that there is not a single approach to explain this emergence because of the recency of theorising on urban violence. This emphasises the importance of taking a multifaceted approach to explain the emergence of urban violence. It was found that writers drew on existing frameworks designed to explain civil conflicts. Notable among these were the grievance perspective, greed hypothesis and structural violence. Attention was also paid to the individual different causes or circumstances of urban violence. These included population increase and density, weak or ineffective state institutions and inequality and poverty. It was concluded that only by drawing from these multi-layered and inter-related approaches that a holistic explanation of the emergence of violence in Medellín can be provided. A review of the perpetrators of violence in Latin American urban environments identified the principal perpetrators as gangs, although other non-state armed groups were also found to exist. The need to understand the specific characteristics of these perpetrators was emphasised and involved phenomena, such as their organisation, motivation and use of violence and their relationship to the state. The last part of the literature review identified the most influential approaches to urban violence reduction and prevention - social disorganisation, broken windows and social capital. It was noted how urban violence reduction and prevention approaches need to be designed with a nuanced understanding of the historical, geographical, spatial and structural complexities of specific urban contexts. This study demonstrated how such a methodology provides the data for a context-specific analysis and a deep and rich understanding of the complexities of violence in an urban environment. A final point discussed in the literature review was that in measuring the success of a particular approach, policy makers often rely on the reduction of homicide rates, as these indicators are 173 the most widely used measure for assessing levels of violence in urban environments and the prime indicators for the reduction of violence. After reviewing the contemporary literature on urban violence and identifying five key analytical questions, this thesis focussed on the following questions to guide the research:

1. What explains the emergence of violence in Medellín? 2. What factors explain the decline of violence in Medellín? 3. Who are the perpetrators of violence? 4. How do these perpetrators use violence? 5. What implications does the Medellín experience have for Colombia and addressing urban violence more generally?

Chapter Three outlined the methodology to pursue these research questions. The methodology included the use of a single case study, documentary and archival analysis and fieldwork in Medellín. This fieldwork combined semi-structured interviews with an array of actors including government officials, former government officials, academics, business leaders, and members of community and civil society organisations. These actors all had valuable knowledge about the emergence and decline of violence as well as of the different perpetrators of violence in Medellín. Fieldwork also used participant observations methods. These included observing, and participating in academic, public policy, community forums and social events. The experience of living in Medellín provided a holistic insight into the dynamics affecting perceptions of security and an appreciation of the complexities of violence in an urban environment from both a public policy perspective and from the viewpoints of citizens. The methodology of this thesis makes an important contribution to the urban violence literature as during fieldwork I was able to obtain a deep and rich understanding of the complexities of urban violence. The insights provided in this thesis will assist policy makers seeking to design and implement urban violence reduction and prevention policy approaches in the urban environments of low-and-middle-income countries. The empirical data collected for this thesis were organised into three chapters. The first data chapter (Chapter Four) provided a historical narrative of the emergence of increasing levels of violence in Medellín in the 1980s. There was a discussion of the themes and concepts which help to explain the emergence of violence in Medellín; notably the rapid urbanisation the city experienced, the high levels of inequality and deprivation, and the neglect of the growing population by the state. All of these circumstances created an environment conducive to violence. But it was shown how the growth of narco- trafficking in Medellín played a major role in the rising tide of violence. In tracing the growth

174 of narcotrafficking this chapter focused on the Medellín Cartel and the criminal bandas which worked for this organisation. In order to control the drug business, criminal groups fought amongst each other, and with the state, to assert their claims and protect their gains. These individuals took advantage of the breakdown in law and order, a breakdown which they greatly contributed to. The chapter showed the complexity of the situation and that no single event, actor or situation throughout the 1980s can be considered in isolation when analysing the emergence of the high levels of violence in the city. The data presented in this chapter were used to provide multifaceted explanations of the emergence of violence in Medellín, ranging from the effects of the rapid urbanisation to the rise of narco-trafficking and the criminal bandas. No one single approach can explain why the levels of violence in the city spiralled out of control and it only by drawing from these multi-layered and inter-related approaches can a holistic explanation of the emergence of violence in Medellín be provided.

The second data chapter, Chapter Five, provided an overview of the rising levels of violence in Medellín in the 1990s. In the early part of the 1990s, homicides reached appallingly high levels with the city portrayed as the ‘most dangerous city in the world’ in 1991. The chapter identified the Presidential Council as a turning point in efforts to solve the city’s problem of violence. Stakeholders involved in this Council had concluded the main causes for the high levels of violence were ‘structural’ by which they meant high levels of poverty and inequality, high levels of unemployment, low educational and health levels, high levels of informal housing, a significant lack of public space and the prevalence of narco-trafficking. Solutions were developed to address the structural problems, such as PRIMED and the Search Block. However, these approaches did not achieve their objective of reducing the violence. One reason for this was tension between the nationally led Presidential Council and the local government in implementing the policies. Also, the activities of the Medellín Cartel extended their influence beyond those directly involved in the trafficking of narcotics. The chapter also showed how by the end of the 1990s, there were three types of non-state armed groups active in the city: the militia, the criminal bandas and the right-wing paramilitaries. All of these non-state armed groups used homicidal violence to obtain and/or maintain control over territories. The identities of group members were however ‘blurred’, as young male members appeared to be attracted to a group for its social and economic benefits, rather than a commitment to particular political ideology.

The third data chapter, Chapter Six, provided an overview of the decline of levels of violence in the city in the 2000s up to November 2015. In this period, there was a reduction from the

175 appallingly high levels of homicides in the 1990s to moderate ones and by 2015 Medellín even departed from the list of the ‘50 Most Dangerous Cities in the World’. The reduction of homicides was firstly contextualised within the social urbanism policies introduced by the city government. The social urbanism policies had the primary objective of improving the welfare of the poor and marginalised and were seen as a success. The discussion showed how these policies were widely cited as the reason for the reduction of homicides, given that the reduction occurred during and after the implementation of these policies. But this total explanatory reliance on social urbanism was questioned and research attention turned to the non-state armed groups in the city and their use of violence. The insights provided in this chapter from an array of local actors in Medellín show the importance of being on-the-ground and obtaining a deep and rich understanding of the complexities of urban violence. The reduction was then contextualised in terms of how the non-state armed groups use violence to obtain and/or maintain territorial control. The discussion showed how despite this significant reduction in homicides, these non-state armed groups continued to have a strong presence in the city. It was found that these non-state armed groups were using other forms of violence, such as forced urban displacement, to maintain and/or obtain territorial control. Members of these groups continued to be young poor males who see their membership as providing them with social and economic benefits which they cannot obtain from the wider society. In accounting for the reduction of urban violence policy makers must take account of the processes of non-state armed group formation and maintenance, the dynamics of non-state inter-group relations and the relations between state and non-state groups. Chapter Seven analysed the factors which explain the reduction of violence in Medellín applying the theoretical frameworks identified in the literature review to the data presented in the empirical chapters to answer the research questions.

In answering the first research question of this thesis (what explains the emergence of violence in Medellín?) the first part of the chapter analysed the emergence of violence in Medellín using the grievance theory, greed hypothesis, structural violence and the different specific causes of urban violence: population increase and density, weak or ineffective state institutions and inequality, and poverty. The explanation focused on the rapid growth of urbanisation the city experienced from the 1950s and how the state was unable or unwilling to support this growth. The neglect by the state led to the exclusion of the new migrants from the larger society and their life experiences were limited and opportunities/inequalities increased. But to adequately explain the emergence of violence it was necessary to show how the growth of narco-

176 trafficking in Medellín led levels of violence in the city to spiral out of control. This led to the conclusion that there is no one reason or causal factor which can explain the emergence of the high levels of violence in Medellín, but rather there is a need to take a multi-faceted approach drawing from grievance theory, greed hypothesis and structural violence as well as the actions of individuals and organisations involved in narco-trafficking or championing particular political ideologies.

The second research question which guided this thesis was: what factors explain the decline of violence in Medellín? The second part of the analysis chapter focused on the policy approaches to reduce the violence and prevent any future outbreaks of violence. The explanation showed how the policy approaches in the 1990s did not achieve their objectives because of tensions between the national and local government in the implementation of policy and because the activities of the Medellín Cartel extended its influence beyond those directly involved in the trafficking of narcotics. Consideration was then made of the approaches used by the Fajardo administration of Medellín from 2003. Given the reduction of homicides during this administration, claims were made about how Fajardo’s social urbanism had caused the reduction. This supported assumptions in social disorganisation theory, broken windows theory and social capital about urban violence reduction and prevention. However, in scrutinising the claims made by international organisations, researchers and the local government about the social urbanism’s effect on violence, it was shown that there were limited attempts to clearly delineate the causal links to demonstrate that the policies did in fact cause the reduction. Although the policies appear to have contributed to violence reduction, it was proposed that they did not provide a total explanation for the over 90 percent reduction of homicides the city experienced because of two reasons. The first reason was because of the reliance on assumptions social urbanism makes about cause and effect in human behaviour. The second reason was because of the assumptions the policy makers made about perpetrators of violence. The policy makers did not examine the dynamics of group conflict and its accommodation and how this impinged on the incidence of homicide. Also, there could be simplistic assertions alleging that spontaneous violence was responsible for many of the homicides, a focus that conveniently omitted consideration of the non-state armed groups.

The final part of the chapter analysed the perpetrators of violence, answering the questions: Who are the perpetrators of violence and how do these perpetrators use violence? The chapter revealed additional important factors leading to the reduction of violence (homicides) over the period studied. These factors concerned the changing dynamics of inter-group rivalry and

177 control of territory and illegal business in Medellín. This analysis showed how despite differences between the non-state armed groups in Medellín, the militia, the paramilitaries and the criminal bandas, they all used homicidal violence to maintain and/or obtain control of territories and often to engage in crime, notably narco-trafficking. This inter-group rivalry contributed to the changing patterns of violence in the city over the period studied. The chapter concluded that a more plausible explanation of the over 90 percent reduction of homicides in Medellín can be found in changes in how these non-state armed groups used violence (homicides) to maintain and/or obtain control of territories. When control over the violent groups was centralised or pacts reached between the major non-state players, the incidence of homicide decreased. The monopolistic control of violence and the pacts also entailed the adoption of a business model that saw high homicide rates as bad for illegal business. When homicides were low the state was disinterested in the non-state armed groups’ activities. This disinterest from the state enabled the non-state armed groups to focus on their illegal business as against the costly pursuit of fighting each other or the state as had been the case under Escobar’s Medellín Cartel. This change in the business model of these non-state armed groups was one which was largely independent of the social urbanism policies that others have claimed as being responsible for the urban violence reduction in Medellín.

The final question of this thesis was: What implications does the Medellín experience have for Colombia and addressing urban violence more generally? The first implication of findings from this thesis is that policy makers from Latin America urban environments should widen the scope of assumptions they make about perpetrators of violence. The data collected and analysed for this thesis showed how homicides in Medellín are largely related to conflicts between non-state armed groups over territorial disputes. These non-state armed groups are not disorganised delinquents who engage in ‘spontaneous acts of violence’ but rather they are organised and adopt a rational approach as to how they use violence. The members of these groups may come from the poorer neighbourhoods; however, they are integrated into larger organised networks whose business is narco-trafficking.

Recent events in Medellín confirm that non-state armed groups can commit homicidal violence or at least place pressure on influential officials when they perceive their business opportunities are under threat. For example, in February 2016, over a period of 48 hours three dismembered bodies were found in different parts of the city, with one Medellín-based analyst interpreting these organised acts of violence as the criminals ‘dumping’ them in different areas to avoid drawing the attention of authorities in one area of the city (Analisis Urbano 2016). In the same

178 month, the newly elected mayor, Federico Gutiérrez, received reports that there was a plan to assassinate him in retaliation for police operations in the centre of the city and the Castilla neighbourhood (El Espectador 2016). However, some media cited Medellín-based analysts who spoke of how the bandas did not have the courage or determination to actually assassinate the mayor, as this would be a return to the high levels of violence of the 1990s. The threat, therefore, was made as a form of pressuring the mayor to stop the operations in the centre of the city, where the majority of profits from organised crime are generated (El Tiempo 2016). The assassinations of two foreigners (Israeli and Danish) in less than a month also show how homicides are related to territorial disputes between non-state armed groups. Local media reported that these foreigners were assassinated because of their involvement in organised crime networks in the city, as the foreigners had been renting out apartments to tourists and offering them ‘a package of cocaine and prostitutes’ a business known as ‘narco-tourism’ (Semana 2016). The manner in which these foreigners were assassinated, with the Israeli shot ten times and the Danish shot in the head, show how violence continues to be manifested in an organised manner rather than a disorganised one (Semana 2016).

What this thesis has clearly demonstrated is that in accounting for the reduction of violence in Latin American cities one must take account of the processes of non-state armed group formation and maintenance, the dynamics of non-state inter-group relations, and the relations between state and non-state groups. Each of these factors change, sometimes quite radically, as we have seen in Medellín. The agents of urban violence make strategic choices about what they will attempt to do to achieve their objectives and this can range from the use of extreme violence, as with Escobar, to its much-moderated utilisation under Don Berna. But these strategic choices are made by leaders and their lieutenants with imperfect information and in unstable environments. This appears to be what accounts for Medellín’s transformation from the most dangerous city in the world to a more desirable one. Understanding the processes of these non-state armed groups requires on-the-ground research and interaction with an array of local actors. The methodology of this thesis thus makes an important contribution to the urban violence academic literature and policy analysis literature as it provides deep and rich understanding of the complexities of urban violence. The findings of this thesis lead to a moderation of the claims made by the advocates who saw social urbanism policies as being solely responsible for the reduction of violence in Medellín. It has been clearly demonstrated in this research that the social urbanism policies provide only a partial explanation for the reduction of violence in Medellín. This is not an argument against such policies. Indeed, there

179 needs to be more effort to reduce the high levels of socio-economic and political inequality that are seen in many Latin American cities and which are often associated with urban violence. The techniques of social urbanism have been seen to improve the welfare of the poor and marginalised and lead to more inclusive development. Those who do benefit appreciate such policies especially as they have often suffered years of neglect by the state. Encounters with the latter are often in terms of repression and state violence, occurrences which create high levels of mistrust of the state among poor urban communities. While the inhabitants of the communities may not like the non-state armed groups which take up residence in their midst, at least these groups provide some services that community members want and/or need. It is dynamics of this nature that policy makers need to understand if they are to effectively address urban violence in Latin American cities. It is only by further understanding the relations between non-state armed groups and communities can urban violence be significantly reduced.

180

Reference List

Abadinsky, H. 2012. Organized Crime. Tenth Edition. Belmont: Wadsworth Cengage Learning. Abello-Colak, A. & Guarneros-Meza, V. 2014. The role of criminal actors in local governance. Urban Studies, 51, 3268-3289. Abello-Colak, A. & Pearce, J. 2015. Securing the global city?: An analysis of the ‘Medellín Model’through participatory research. Conflict, Security & Development, 15, 197- 228. A1 (Academic) May 2014 ‘Interview’ in PhD Interviews ed. Caroline Doyle Medellín, Colombia. Aguilar, J. & Miranda, L. 2006. Entre la articulación y la competencia: las respuestas de la sociedad civil organizada a las pandillas en El Salvador. Maras y Pandillas en Centroamérica: Las respuestas de la Sociedad Civil Organizada, 4. Alcaldía de Medellín. 2010. La Empresa para la Seguridad Urbana, ESU, da un Salto Technologico para Fortalecer la Seguridad y la Convivencia el Area Metropolitana [Online] Available: http://www.amuprev.org/noticias/?id=288 [Accessed 21 May 2016]. Alcaldía de Medellín. 2011. Laboratorio Medellín Catalogo de diez practicas vivias [Online] Available: http://www.aciMedellín.org/Portals/0/Images/pdf_publicaciones/laboratorio_Medellín -aci.pdf [Accessed 30 May 2016]. Alcaldía de Medellín. 2015. Hechos de Nuestra Nuevo Medellín: Tasa de homicidios mas bajas en los ultimos 35 anos. [Online] Available: https://www.youtube.com/watch?v=7PQT_SXpqy4 [Accessed 30 May 2015]. Alcaldía de Medellín. 2016. Mapas de la division politico y administrativa de Medellín y sus corregimientos [Online] Available: https://www.Medellín.gov.co/irj/portal/Medellín?NavigationTarget=navurl://640dc96 1eeef4b5fad087a0a73e038c2 [Accessed 30 May 2016]. Altbeker, A. 2009. The Building of the New South African Police Service: The Dynamics of Police Reform in a Changing (and Violent) Country. Policing Developing Democracies, 260-279.

181

Americas Programme. 22 July 2015. Rebuilding Community in Medellín’s Violent Slums [Online]. Available: http://www.cipamericas.org/archives/15613. [Accessed 31 May 2016]. American Security Project. 2013. Narco-trafficking in the Americas. [Online] Available: https://americansecurityproject.org/wp-content/uploads/2013/05/Ref-0107- Narcotrafficking-in-the-Americas.pdf [Accessed 31 May 2016]. Amnesty International. 2005. Colombia: The Paramilitaries In Medellín: Demobilisation or Legislisation? [Online]. Available: https://www.amnesty.org/en/documents/amr23/019/2005/en/. [Accessed 30 May 2016]. Analisis Urbano. 19 September 2015. Dos nuevos casos de desmembramiento en Medellín [Online]. Available: http://analisisurbano.org/dos-nuevos-casos-de- desmembramiento-en-Medellín/. [Accessed 31 May 2016]. Analisis Urbano. 11 February 2016. 48 horas de terror en Medellín y fiscalia mira hacia otro lado. [Online] Available: http://analisisurbano.org/48-de-terror-en-Medellín-y- fiscalia-mira-hacia-otro-lado/ [Accessed 06 October 2016]. Angarita, P, 2003. Conflictos, guerra y violencia urbana: Interpretaciones problemáticas. Revista Nómadas Universidad Central de Colombia, (19), 20. Angarita, P., Gallo, H. & Jiménez, B. I. 2008. Dinámicas de guerra y construcción de paz: estudio interdisciplinario del conflicto armado en la comuna 13 de Medellín. Grupo interdisciplinario e interinstitucional de investigación sobre conflictos y violencias. Medellín: Universidad de Medellín. Área Metropolitana del Valle de Aburrá. 2006. Encuesta domiciliaria origen y destino 2005, Medellín. Archer, R. P. 1990. The transition from traditional to broker clientelism in Colombia: political stability and social unrest. University of Nortre Dame: Helen Kellogg Institute for International Studies. Architectural Review. 2011. AR Report on Colombia. [Online]. Available: https://www.architectural-review.com/archive/viewpoints/ar-report-on- colombia/8610406.article [Accessed 31 May 2016]. Arco Rivas, A. 2005. Ciudadania Armada Aportes a la Interpretacion de Procesos de Defensa y Aseguramiento Comuntario en Medellín: El Caso de la Milicias Populares. Masters Thesis, Universidad de Antioquia: Medellín. Bagley, B. M. 1988. Colombia and the . Foreign Affairs, 67, 70-92. 182

Bahl, V. 2012. Murder Capital to Modern Miracle?: The Progression of Governance in Medellín, Colombia, Development Planning Unit, University College London. Baird, A. 2009. Methodological dilemmas: Researching violent young men in Medellín, Colombia. IDS Bulletin, 40, 72-77. Baird, A. 2011. Negotiating Paths to Manhood: Violence Reproduction in Medellín's Periphery - Thesis submitted for degree of Doctor of Philosophy. [Online] Available: https://bradscholars.brad.ac.uk/bitstream/handle/10454/5246/BAIRD,%20Adam%20F ull%20Thesis.pdf?sequence=1 [Accessed 05 May 2016]. Baker, J. & Schuler, N. 2004. Analyzing Urban Poverty: A Summary of Methods and Approaches. Washingston DC: World Bank Publications. Baker, J. 2008. Urban Poverty: A Global Overview. Washington DC: World Bank Publications. Ballentine, K. & J. Sherman (eds.) 2003. The Political Economy of Armed Conflict: Beyond Greed and Grievance. Boulder: Lynne Rienner Publishers Ballentine, K. & Nitschke, H. 2005. Peace before profit: The challenges of governance. In Ballentine, K & Nitzsche. H (eds) Profiting from Peace: Managing the Resource Dimensions of Civil War. London: Lynne Rienner. Barrero Tapias, J. R. 2001. El anclaje cultural del conflicto político: las milicias populares y el ELN como casos de estudio. Revista Javeriana, 673, 213-238. Bassiouni, M. C. 2008. The new wars and the crisis of compliance with the law of armed conflict by non-state actors. The Journal of Criminal Law and Criminology (1973-), 98, 711-810. Baumer, E. P. & Wolff, K. T. 2014. Evaluating contemporary crime drop (s) in America, New York City, and many other places. Justice Quarterly, 31, 5-38. Becker, J., 1996. The Internet, structural violence and non-communication. Media Development, 43, 11-12. Beall, J., Goodfellow, T. & Rodgers, D. 2011. Cities, conflict and state fragility - Working Paper no. 85 - Cities and Fragile States. [Online]. Available: http://www.alnap.org/resource/7341. [Accessed 31 May 2016]. Beall, J., Goodfellow, T. & Rodgers, D. 2013. Cities and conflict in fragile states in the developing world. Urban Studies, 50 (15) 3065- 3083. Belousov, K., Horlick-Jones, T., Bloor, M., Gilinskiy, Y., Golbert, V., Kostikovsky, Y., Levi, M. & Pentsov, D., 2007. Any port in a storm: fieldwork difficulties in dangerous and crisis-ridden settings. Qualitative Research, 7(2), 155-175. 183

Bedoya Marín, D.A. & Jaramillo Martínez, J., 1991. De la Barra a la Banda.Medellín: Lealón. Bellair, P. E. 1997. Social interaction and community crime: Examining the importance of neighbor networks. Criminology, 35, 677-704. Bellows, J. & Miguel, E. 2006. War and institutions: New evidence from Sierra Leone. The American Economic Review, 96, 394-399. Berdal, M. 2005. Beyond greed and grievance–and not too soon. Review of International Studies, 31, 687-698. Berdal, M. R. & Malone, D. 2000. Greed & Grievance: Economic Agendas in Civil Wars, London: Lynne Rienner Publishers. Betancur, J. J. 2007. Approaches to the regularization of informal settlements: The case of PRIMED in Medellín, Colombia. Global Urban Development Magazine, 3. Bhatia, M. V. 2005. Fighting words: Naming terrorists, bandits, rebels and other violent actors. Third World Quarterly, 26, 5-22. Boulding, K. E. 1977. Twelve friendly quarrels with Johan Galtung. Journal of Peace Research, 14, 75-86. Bourdieu P. 1986. The forms of capital. In Richardson J (eds) Handbook of Theory and Research for the Sociology of Education. New York: Greenwood. Brand, P. 2013. Governing inequality in the South through the Barcelona model: 'social urbanism' in Medellín, Colombia. [Online] Available: http://www.dmu.ac.uk/documents/business-and-law- documents/research/lgru/peterbrand.pdf [Accessed 21 May 2016]. Brezina, T., Piquero, A. & Tibbetts, S. 2002. Assessing the rationality of criminal and delinquent behavior: a focus on actual utility. Rational choice and criminal behavior: Recent research and future challenges, 244-264. Buhaug, H. & Urdal, H. 2013. An urbanization bomb? Population growth and social disorder in cities. Global Environmental Change, 23, 1-10. Burnham, P., Lutz, K. G., Grant, W. & Layton-Henry, Z. 2008. Research Methods in Politics, New York: Palgrave Macmillan. BL1 (Business Leader) May 2014 ‘Interview’ in PhD Interviews ed. Caroline Doyle Medellín, Colombia. Business Insider. 26 January 2016. Most Violent Cities in the World [Online]. Available: http://www.businessinsider.com/most-violent-cities-in-the-world-2016-1?IR=T/#2- san-pedro-sula-honduras-had-11103-homicides-per-100000-residents-49 184

Calderon, C. 2008. Learning from Slum Upgrading and Participation [Online]. Available: http://kth.diva-portal.org/smash/get/diva2:126733/FULLTEXT01 [Accessed 30 May 2016]. Caracol Radio. 27 January 2011. Asesinada un joven en el noroccidente de Medellín [Online]. Available: http://caracol.com.co/radio/2011/01/27/judicial/1296111900_417089.html [Accessed 10 May 2016]. Casas-Casas, A. & Giraldo, J. 2015. Seguridad y Convivencia en Medellín Aproximaciones empiricas a sus atributos y desafios. Medellín: Universidad Eafit & Alcaldía de Medellín. Centre for Evidence-Based Crime Policy. 2016. Broken Windows Policing. [Online] Available: http://cebcp.org/evidence-based-policing/what-works-in-policing/research- evidence-review/broken-windows-policing/ [Accessed 05 October 2016]. Centro de Analisis Politico. 2016. Lineas de trabajo [Online]. Available: http://www.eafit.edu.co/centros/analisis-politico/Paginas/lineas-de-trabajo.aspx [Accessed 10 May 2016]. Cerdá, M., Morenoff, J. D., Hansen, B. B., Hicks, K. J. T., Duque, L. F., Restrepo, A. & Diez-Roux, A. V. 2012. Reducing violence by transforming neighborhoods: a natural experiment in Medellín, Colombia. American Journal of Epidemiology, 175 (10). Christian Science Monitor. 18 July 2014. Who Controls Medellín? Fragile Peace in Colombia’s Model City. [Online] Available: http://www.csmonitor.com/World/Americas/2014/0718/Who-controls-Medellín- Fragile-peace-in-Colombia-s-model-city [Accessed 21 May 2016]. CNN (Cable News Network). 23 October 2013. Medellín reborn: Colombian city moves out of Escobar shadow. [Online] Available: http://edition.cnn.com/2013/10/18/world/americas/colombia-Medellín-regeneration/ [Accessed 21 May 2016]. Cock, J. 2000. Section Three: Firearms and a Culture of Violence in South Africa. Society Under Siege: Managing Arms in South Africa, 3, 75. Cockayne, J. & Lupel, A., 2009. Introduction: Rethinking the relationship between peace operations and organized crime. International Peacekeeping, 16(1), 4-19. Cockayne, J. 2013. Chasing shadows: Strategic responses to organized crime in conflict- affected situations. The RUSI Journal, 158, 10-24. Coffey, A., 1999. The Ethnographic Self: Fieldwork and the Representation of Identity. Sage 185

Collier, P. 2000. Economic causes of civil conflict and their implications for policy. [Online]. Available: http://go.worldbank.org/HBOXCW5CZ0 .[Accessed 10 May 2016]. Collier, P. & Hoeffler, A. 1998. On economic causes of civil war. Oxford Economic Papers, 50, 563-573. Collier, P. & Hoeffler, A. 2004. Greed and grievance in civil war. Oxford Economic Papers, 56, 563-595. Collier, P. & Hoeffler, A. 2012. High-value natural resources, development, and conflict: Channels of causation. Strengthening Post-conflict Peacebuilding through Natural Resource Management, 1. Collier, P. & Sambanis, N., 2005. Understanding Civil War, (2 vols.). Washington: The World Bank. Colombia Reports. 23 January 2013. Medellín mayor invites Hillary Clinton to visit. [Online] Available: http://colombiareports.com/Medellín-mayor-invites-hillary-clinton-to-visit/ [Accessed 31 May 2016]. Community Cohesion Review Team. 2001. The Cantle Report- Community Cohesion: A Report of the Independent Review Team. [Online]. Available: http://resources.cohesioninstitute.org.uk/Publications/Documents/Document/Default.a spx?recordId=96. [Accessed 30 May 2016]. CL1 (Community Leader one) July 2015 ‘Interview’ in PhD Interviews ed. Caroline Doyle Medellín, Colombia. CL2 (Community Leader two) July 2015 ‘Interview’ in PhD Interviews ed. Caroline Doyle Medellín, Colombia. CL3 (Community Leader three) July 2015 ‘Interview’ in PhD Interviews ed. Caroline Doyle Medellín, Colombia. Coupé, F., Brand, P. and Dávila, J.D., 2012. Medellín: contexto institucional y cambio de paradigma urbano. In Davila J. D.(eds) Movilidad urbana y pobreza: Aprendizajes de Medellín y Soacha, Colombia. London: Development Planning Unit and University of Colombia. Creswell, J. W. 2013. Research Design: Qualitative, Quantitative, and Mixed Methods Approaches. Fourth Edition. California: Sage publications. Dávila, J. D. & Daste, D. 2011. The NAERUS Annual Conference ‘The city at a human scale’ Faculty of Architecture, Universidad Politécnica de Madrid 20-22 October 2011. [Online]. Available:

186

https://www.bartlett.ucl.ac.uk/dpu/metrocables/dissemination/Davila-Daste-Naerus- 2011.pdf [Accessed 10 May 2016]. Davila, L. 2016. Violencia urbana, conflicto y crimen en Medellín: una revisión de las publicaciones académicas al respecto. Revista Criminalidad, 58(2), pp.107-121. DEA (Drug Enforcement Agency). 1994. The Cali Cartel: The New King's of Cocaine. [Online] Available: https://www.ncjrs.gov/pdffiles1/Digitization/152436NCJRS.pdf [Accessed 10 May 2016]. De Boer, J. & Bosetti, L. 2015. The Crime-Conflict “Nexus”: State of the Evidence. Occasional Paper. [Online]. Available: http://collections.unu.edu/eserv/UNU:3134/unu_cpr_crime_conflict_nexus.pdf. [Accessed 10 May 2016]. Devlin & Chaskel. 2010. From Fear to Hope in Colombia: Sergio Fajardo and Medellín, 2004- 2007 [Online]. Available: https://successfulsocieties.princeton.edu/sites/successfulsocieties/files/Policy_Note_I D116.pdf [Accessed 11 May 2016]. Dowdney, L. 2003. Children of the drug trade: A case study of children in organized armed violence in Rio de Janeiro. [Online]. Available: nisat.prio.org/misc/download.ashx?file=24336. [Accessed 10 May 2016]. Drummond, H., Dizgun, J. & Keeling, D. J. 2012. Medellín: A city reborn? FOCUS on Geography, 55, 146-154. Duffield, M. R. 2001. Global governance and the New Wars: The Merging of Development and Security. London: Zed books. EAFIT Universidad - Centro de Analisis Politico (CAP). 2014. Informe final del Acompañamiento al Diseño de la Política Pública de Seguridad y Convivencia para el Municipio de Medellín. Medellín: Universidad EAFIT- Secretaría de Seguridad- Alcaldía de Medellín. Eide, K., 1971. Note on Galtung's Concept of'Violence'. Journal of Peace Research, 8(1), 71- 71. El Colombiano. 2002a. ACU aprueba el proyecto urbano en el Valle de Aburra. El Colombiano, 17 June. El Colombiano. 2002b. Un mes despues, la 13 respire otro aire. El Colombiano, 15 November. El Colombiano. 4 June 2015a. Por disturbios de encapuchados Universidad de Antioquia suspendio actividades [Online] Available: 187

http://www.elcolombiano.com/antioquia/seguridad/universidad-de-antioquia- suspendio-actividades-por-disturbios-de-encapuchados-FN2078007 [Accessed 10 August 2016]. El Colombiano. 9 July 2015b. Asesinan a un hombre en el poblado [Online] Available http://www.elcolombiano.com/antioquia/seguridad/asesinan-a-un-hombre-en-el- poblado-XL2278189 [Accessed 10 August 2016]. El Colombiano. 10 May 2015c. Fin de semana de Madres dejó 4 muertos [Online]. Available: http://www.elcolombiano.com/antioquia/seguridad/fin-de-semana-de-madres-dejo-4- muertos-CA1893985 [Accessed 30 May 2016]. El Espectador. 16 March 2015a. Ana Cristina Restrepo, la última periodista que los violentos intentan amedrentar [Online]. Available: http://www.elespectador.com/noticias/judicial/ana-cristina-restrepo-ultima-periodista- los-violentos-i-articulo-549726 [Accessed 10 May 2016]. El Espectador. 29 May 2015b. Están asesinando a nuestros defensores de derechos humanos. [Online]. Available: http://www.elespectador.com/noticias/nacional/estan-asesinando- nuestros-defensores-de-derechos-humano-articulo-563430 [Accessed 10 May 2016]. El Espectador. 03 March 2016. Si amenazan el alcalde de Medellín que decir a los ciudadanos. [Online]. Available: http://www.elespectador.com/noticias/nacional/antioquia/si-amenazan-al-alcalde-de- Medellín-decir-de-los-ciudada-articulo-620120 [Accessed 06 October 2016]. El Pais. 03 December 1993. El narcotraficante Pablo Escobar, abatido a tiros por la policía colombiana en Medellín [Online]. Available: http://elpais.com/diario/1993/12/03/internacional/754873211_850215.html [Accessed 31 May 2016]. El Tiempo. 17 September 1991. Milicias populares: sangriento juicio [Online]. Available: http://www.eltiempo.com/archivo/documento/MAM-155834 [Accessed 10 May 2016]. El Tiempo. 06 May 1999a. Paras siguen huella de el mexicano [Online]. Available: http://www.eltiempo.com/archivo/documento/MAM-910444 [Accessed 30 May 2016]. El Tiempo. 1999b. A piedad se la llevó el polvorero [Online]. Available: http://www.eltiempo.com/archivo/documento/MAM-903382 [Accessed 30 May 2016].

188

El Tiempo. 28 May 2001a. Carro bomba en Medellín [Online]. Available: http://www.eltiempo.com/archivo/documento/MAM-510591. [Accessed 31 May 2016]. El Tiempo. 11 January 2001b. Noche de terror en Medellín [Online]. Available: http://www.eltiempo.com/archivo/documento/MAM-654652 [Accessed 31 May 2016]. El Tiempo. 08 July 2001c. Sembrando paz [Online]. Available: http://www.eltiempo.com/archivo/documento/MAM-429694 [Accessed 31 May 2016]. El Tiempo.25 January 2012. Bandas cobrarían extorsión por usar escaleras eléctricas de Comuna 13 [Online]. Available: http://www.eltiempo.com/archivo/documento/CMS- 10998261 [Accessed 10 May 2016]. El Tiempo.06 January 2013. Modelo convidó a más mujeres a narcofiesta que terminó en masacre [Online]. Available: http://www.eltiempo.com/archivo/documento/CMS- 12493811 [Accessed 30 May 2016]. El Tiempo. 29 July 2014. Nuevo ataque con granada en Medellín deja nueve heridos [Online]. Available: http://www.eltiempo.com/colombia/Medellín/atentado-con- granada-en-el-centro-de-Medellín/14318918 [Accessed 10 May 20120]. El Tiempo. 18 August 2015. Medellín es refernte en tecnologias para la seguiridad en Colombia [Online]. Available: http://www.eltiempo.com/colombia/Medellín/Medellín-es-referente-en-tecnologias- para-la-seguridad-en-colombia/16249306 [Accessed 10 September 2015]. El Tiempo. 29 February 2016. Los capos que estarian planeando atentar contra el alcalde de Medellín. [Online]. Available: http://www.eltiempo.com/colombia/Medellín/amenazas-contra-federico-Gutiérrez- alcalde-de-Medellín/16522316 [Accessed 06 October 2016]. Estrada, W. & Adriana Gómez, B. 1992. Somos Historia. Comuna Nororiental. Medellín: No publisher. European Forum for Urban Security. 2013. Euro Social Program Para La Cohesion Social En America Latina [Online]. Available: https://efus.eu/en/about-us/our- partners/partners/8554/ [Accessed 30 May 2016). Fajardo, A. & Andrews, M. 2014. Does successful governance require heroes? The case of Sergio Fajardo and the city of Medellín: A reform case for instruction. WIDER Working Paper. [Online]. Available: 189

https://www.wider.unu.edu/sites/default/files/wp2014-035.pdf. [Accessed 12 May 2016]. Fajnzylber, P., Lederman, D. & Loayza, N. 1998. Determinants of crime rates in Latin America and the world: an empirical assessment. Washington DC: World Bank Publications. Fajnzylber P., Lederman, D. & Loayza N. 2000. What causes violence crime? European Economic Review 46, 1323-1357. Farah, D. 01 March 2016. The Nature of the Maras: Douglas Farah - a Rebuttal [Online]. Available: http://www.insightcrime.org/news-analysis/the-nature-of-the-maras-doug- farah-rebuttal [Accessed 10 May 2016]. Field Diary May 2014. Caroline Doyle. Medellín, Colombia. Field Diary June 2014. Caroline Doyle. Medellín, Colombia. Field Diary July 2014. Caroline Doyle. Medellín, Colombia. Field Diary May 2015 Caroline Doyle. Medellín, Colombia. Field Diary June 2015. Caroline Doyle. Medellín, Colombia. Field Diary July 2015. Caroline Doyle. Medellín, Colombia. Fijnaut, C. 2000. Transnational crime and the role of the United Nations in its containment through international cooperation: a challenge for the 21st century. European Journal of Crime Criminal Law and Criminal Justice, 8, 119-128. Filippone, R. 1994. The Medellín Cartel: Why we can't win the drug war. Studies in Conflict & Terrorism, 17, 323-344. Forbes. 2012. Forbes History: The Original 1987 List Of International Billionaries. [Online]. Available: http://www.forbes.com/sites/seankilachand/2012/03/21/forbes-history-the- original-1987-list-of-international-billionaires/#29ba47d4696e [Accessed 31 May 2016]. Fortune. 2014. Fortune 5: The biggest organized crime groups in the world. [Online] Available: http://fortune.com/2014/09/14/biggest-organized-crime-groups-in-the- world/ [Accessed 05 October 2016]. Foster, J. 1995. Informal social control and community crime prevention. British Journal of Criminology, 35, 563-583. FGO1 (Former Government Official one), July 2015 ‘Interview’ in PhD Interviews ed. Caroline Doyle Medellín, Colombia. FGO2 (Former Government Official two), June 2014 ‘Interview’ in PhD Interviews, ed. Caroline Doyle, Medellín, Colombia. 190

FGO3 (Former Government Official three) (FGO3), June 2014 ‘Interview’ in PhD Interviews ed. Caroline Doyle Medellín, Colombia. FGO4 (Former Government Official four), June 2014 ‘Interview’ in PhD Interviews, ed. Caroline Doyle, Medellín, Colombia. FGO5 (Former Government Official five), July 2014 ‘Interview’ in PhD Interviews, ed. Caroline Doyle, Medellín, Colombia. Franco L. & Caputo, C. 2013. Urban violence and humanitarian action in Medellín. Available: http://www.alnap.org/resource/8742 [Accessed 30 May 2016]. Franco V. 2004. Conflictos Urbanos en las Comunas 1, 3, y 13 de la Ciudad de Medellín, Medellín: Universidad Autónoma Latinoamericana. François, M. & Sud, I. 2006. Promoting stability and development in fragile and failed states. Development Policy Review, 24, 141-160. Fukuyama Francis, C. S. 25 April 2011. Half a Miracle [Online]. Available: http://foreignpolicy.com/2011/04/25/half-a-miracle/ [Accessed 10 May 2016]. Galtung, J. 1969. Violence, peace, and peace research. Journal of Peace Research, 6, 167- 191. Galtung, J. 1985. Twenty-five years of peace research: Ten challenges and some responses. Journal of Peace Research, 22, 141-158. Galtung, J. 1990. Cultural violence. Journal of Peace Research, 27, 291-305. Galtung, J. 1991. Peace with peaceful means, Oslo: PRIO. Gay, R. 2010. Popular Organisation and Democracy in Rio de Janeiro: A tale of Two Favelas. Philadelphia: Temple University Press. Gazdar, H. & Mallah, H. B. 2013. Informality and political violence in Karachi. Urban Studies, 50, 3099-3115. Geneva Centre for Democratic Control of Armed Forces. 2015. Armed Non-State Actors: Current Trends and Future Challenges. [Online] Available: http://www.dcaf.ch/Publications/Armed-Non-State-Actors-Current-Trends-Future- Challenges.[Accessed 21 May 2016]. Ghani, A. & Lockhart, C. 2009. Fixing failed states: A Framework for Rebuilding a Fractured World. USA: Oxford University Press. Gil, M.Y. 2013. Medellín 1993-2013: Una ciudad que no logra encontrar el camino para salir definitivamente del laberinto. [Online] Available: https://www.wilsoncenter.org/sites/default/files/Gil%20Ramirez%20- %20Colombia%20-%20Paper_1.pdf [Accessed 21 May 2016]. 191

Gimode, E. A. 2001. An anatomy of violent crime and insecurity in Kenya: the case of Nairobi, 1985-1999. Africa Development: A Quarterly Journal of CODESRIA, 26, 295-335. Giraldo, J. A. (2008). Conflicto armado urbano y violencia homicida. El caso de Medellín. Urvio Revista Latinoamericana de Seguridad Ciudadana (5), 99-113. Giraldo, J. & Mesa, J. P. 2013. Reintegration without Demobilization: The Case of Urban Militias in Medellín. Colombia Internacional, 217-239. Giraldo, J. & Preciado, A. 2015. Medellín, from Theater of War to Security Laboratory. Stability: International Journal of Security and Development, 4. Global Affairs Canada, 2013. Map- Colombia. [Online] Available: https://www.international.gc.ca/cil-cai/country_insights-apercus_pays/map- carte_co.aspx?lang=eng [Accessed 25 May 2016]. Goffman, E. 1989. On fieldwork. Journal of contemporary ethnography, 18(2), 123-130. Goldstein, D.M. 2014. Qualitative research in dangerous places: Becoming an ‘ethnographer’ of violence and personal safety. Social Science Research Council, DSD Working Papers on Research Security, 1, 1-18. González, J. 2014. La democratizacion del narcotrafico. Revista Encrucijada Americana-Año, 7, 2. GO1 (Government Official one), June 2014 ‘Interview’ in PhD Interviews ed. Caroline Doyle Medellín, Colombia. GO2 (Government Official two), July 2015 ‘Interview’ in PhD Interviews ed. Caroline Doyle Medellín, Colombia. GO3 (Government Official three) June 2014 ‘Interview’ in PhD Interviews ed. Caroline Doyle Medellín, Colombia. Graham, C. & Chaparro, J.C. 2011. Inseguridad salud y bienstar una exploracion inicial basada en encuentas sobre la felicidad en America Latina y el Caribe. Banco Interamericano de Desarrollo. Washington. Gurr, T. R. 1970. Sources of rebellion in Western societies: Some quantitative evidence. The Annals of the American Academy of Political and Social Science, 391, 128-144. Gutiérrez, F., Pinto, M., Arenas, J. C., Guzmán, T. & Gutiérrez, M. 2013. The importance of political coalitions in the successful reduction of violence in Colombian cities. Urban Studies, 50 (15) 3134- 3151. Gutiérrez F. & Jaramillo, A. M. 2004. Crime,(counter-) insurgency and the Privatisation of Security-the Case of Medellín, Colombia. Environment and Urbanization, 16, 17-30. 192

Guzman-Barney. 2014. La planeación policial focalizada y el trabajo coordinado reducen el crimen. Evidencia en ciudades colombianas. Bogotá: Fundación Ideas para la Paz, BID, Policía Nacional. Hagedorn, J. M. 2007. Gangs, institutions, race, and space: The Chicago School revisited. Gangs in the Global City: Alternatives to Traditional Criminology, 13-33. Hauck, P. & Peterke, S. 2010. Organized crime and gang violence in national and international law. International Review of the Red Cross, 92, 407-436. Hazen, J. 2010a. Force multiplier: Pro-government armed groups. Small Arms Survey, 255- 76. Hazen, J. 2010b. Understanding gangs as armed groups. International Review of the Red Cross, 92, 369-386. Hernandez-Garcia, J. 2013. Guest Article: Slum tourism and city branding in Medellín. [Online] Available: https://blog.inpolis.com/2013/01/28/guest-article-slum-tourism- and-city-branding-in-Medellín/ [Accessed 12 May 2016]. Human Rights Watch. 1994. Violence in Colombia [Online]. Available: https://www.hrw.org/reports/1994/colombia/gener2.htm [Accessed 12 May 2016]. Hylton, F. 2007. Medellíns makeover. New Left Review, 44, 66-85. Insight Crime. 13 May 2011. Medellín's turbulent comuna 13 [Online]. Available: http://www.insightcrime.org/investigations/Medellíns-turbulent-comuna-13 [Accessed 12 May 2016]. Insight Crime. 02 January 2013. New Year massacre heralds more violence for Medellín [Online]. Available: http://www.insightcrime.org/news-analysis/new-year-massacre- violence-Medellín [Accessed 30 May 2016]. Insight Crime. 18 November 2015a. Diego Fernando Murillo, alias "Don Berna" [Online]. Available: http://www.insightcrime.org/colombia-organized-crime-news/don-berna [Accessed 12 May 2016]. Insight Crime. 17 November 2015b. Oficina de Envigado [Online]. Available: http://www.insightcrime.org/colombia-organized-crime-news/oficina-de-envigado- profile [Accessed 30 May 2016]. Insuasty, A., Balbin, J., Bastidas, W.,Carrión, S., Jaime, P., Judith, E. & Mejía, W. 2010. Las Victimas en Contextos de Violencia e Impunidad Caso Medellín. Medellín: Fundación Forjando Futuros.

193

Inter-American Development Bank. 2009. Medellín: A City Transformed. [Online]. Available: http://www.iadb.org/en/topics/citizen-security/impact-Medellín,5687.html [Accessed 30 May 2016]. Inter-American Development Bank. 2012. Documento de enfoque: Prevención de la violencia y la delincuencia en América Latina y el Caribe, Evaluaciones operativas y estratégicas de la cContribución del BID, 2012-2013 [Online]. Available: https://publications.iadb.org/bitstream/handle/11319/5516/Documento_de_enfoque__ Documento_de_Enfoque__Prevenci%C3%83%C2%B3n_de_la_Violencia_y_la_Deli ncuencia_en_Am%C3%83%C2%A9rica_%5B1%5D.pdf%3Bjsessionid=41F4101DB 0DD24279C6F49629D8E6C8E?sequence=2 [Accessed 31 May 2016]. Inter-American Development Bank. 2015. Closing Knowledge Gaps: Toward Evidence- Based Crime Prevention Policies in Latin America and the Caribbean. Washington: Inter-American Development Bank. International Business Times. 03 May 2013. Medellín: How Colombia's second city overcame Iits drug cartel past. [Online] Available: http://www.ibtimes.com/Medellín- how-colombias-second-city-overcame-its-drug-cartel-past-1111226 [Accessed 31 May 2016]. International Centre for the Prevention of Crime. 2005. Urban Crime Prevention and Youth at Risk: Compendium of Promising Strategies and Programmes from around the World. 11th United Nations Congress on Crime and Criminal Justice, 18-25 April 2005, Bangkok. International Council on Human Rights Policy. 1999. Ends & means: Human rights approaches to armed groups [Online]. Available: http://www.ichrp.org/files/reports/6/105_report_en.pdf [Accessed 30 May 2016]. Jacoby, T. 2007. Understanding Conflict and Violence: Theoretical and Interdisciplinary Approaches Oxon: Routledge. Jamieson, J., 2002. Negotiating danger in fieldwork on crime. In Lee-Treweek & Linkogle (eds) Danger in the Field: Risk and Ethics in Social Research. London: Routledge. Jaramillo, A. M. (2011). Acerca de los estudios sobreconflicto armado y violencia urbana en Medellín. In J. Giraldo Ramirez, Economía criminal en Antioquia: Narcotrafico (pgs.63-134). Medellín: Universidad Eafit, Fundación Proantioquia y Empresa de Seguridad Urbana.

194

Jaramillo, A. M., Melguizo, R. & Martínez, M. I. V. 1998. En la encrucijada: Conflicto y cultura política en el Medellín de los noventa. Medellín: Alcaldía de Medellín Secretaria de Gobierno. Jaramillo, A. M. & Gil. M. 2014. Ciudades en la encrucijada: Violencia y poder criminal en Río de Janeiro, Medellín, Bogotá y Ciudad Juárez, Medellín: IPC. Jütersonke, O., Muggah, R. & Rodgers, D. 2009. Gangs, urban violence, and security interventions in Central America. Security Dialogue, 40, 373-397. Kalandides, A. & Hernandez-Garcia, J. 2013. Slum tourism, city branding and social urbanism: the case of Medellín, Colombia. Journal of Place Management and Development, 6, 43-51. Kaldor, M., 2013. In defence of new wars. Stability: International Journal of Security and Development, 2(1). Kalyvas, S. N. 2006. The Logic of Violence in Civil War, New York: Cambridge University Press. Kalyvas, S. N. 2015. How civil wars help explain organized crime—and how they do not. Journal of Conflict Resolution, 59, 1517-1540. Kaplan, S. 2009. Identity in fragile states: Social cohesion and state building. Development, 52, 466-472. Keane, J. 1996. Reflections on violence, London: Verso. Keen, D. 2012. Useful Enemies: When Waging Wars is more Important than Winning Them. Cornwell: Yale University Press. Kelling, G. L. & Coles, C. M. 1997. Fixing broken windows: Restoring order and Reducing Crime in our Communities. New York: Simon and Schuster. Kelling, G.L. & Sousa, W.H., 2001. Do police matter?: An analysis of the impact of New York City's Police Reforms. CCI Center for Civic Innovation at the Manhattan Institute. Kilcullen, D. 2013. Out of the Mountains: The Coming Age of the Urban Guerrilla. Oxford: Oxford University Press. Klein, M. W. 2005. The value of comparisons in street gang research. Journal of Contemporary Criminal Justice, 21, 135-152. Klein, M. W. & Maxson, C. L. 2006. Gang structures, crime patterns, and police responses. [Online]. Available: https://www.ncjrs.gov/pdffiles1/nij/grants/188511.pdf. [Accessed 12 May 2016].

195

Koonings, K. & Krujit, D. 1999. Societies of Fear: The Legacy of Civil War, Violence and Terror in Latin America. London: Zed Books. Kruger, T., Landman, K. & Liebermann, S. 2001. Designing safer places: A manual for crime prevention through planning and design. South African Police Service and the CSIR. [Online]. Available www.csir.co.za/Built_environment/docs/Designing_Safer_Places.pdf. [Accessed 30 May 2016]. Kruijt, D. & Koonings, K. 2007. Stability Assessment Colombia for the Conflict Research Unit of the Netherlands Institute of International Relations. The Hague. Johnson, E. 2013. Gangs are the new law in urban Trinidad & Tobago [Online]. Available: http://www.coha.org/gangs-are-the-new-law-in-urban-trinidad-and-tobago/ [Accessed 30 May 2016]. Lamb, R. D. 2008. Ungoverned areas and threats from safe havens [Online]. Available: http://www.dtic.mil/dtic/tr/fulltext/u2/a479805.pdf [Accessed 30 May 2016]. Lamb, R. 2010. Microdynamics of illegitimacy and complex urban violence in Medellín, Colombia [Online]. Available: http://drum.lib.umd.edu/handle/1903/10242 [Accessed 31 May 2016]. La Pastina, A.C., 2006. The implications of an ethnographer’s sexuality. Qualitative Inquiry, 12(4), 724-735. Lee, L. & Stanko, E. 2003. Researching Violence: Essays on Methodology and Measurement. London: Routledge. Lee-Treweek, G. & Linkogle, S (eds). 2000. Danger in the Field: Risk and Ethics in Social Research. London: Routledge. Lee Kuan Yew World City Prize. 2016. 2016 Prize Laureate: Medellín. [Online]. Available: https://www.leekuanyewworldcityprize.com.sg/vid_Medellín.htm [Accessed 06 October 2016]. Leeds, E. 1996. Cocaine and parallel polities in the Brazilian urban periphery: Constraints on local-level democratization. Latin American Research Review, 31, 47-83. Liamputtong, P. & Ezzy, D. 2005. Qualitative Research Methods. South Melbourne: Oxford University Press. Locke, R. 2012. Organized Crime, Conflict, and Fragility: A New Approach. New York: IPI. Lyman, M. & Potter, G. 2007. Organized Crime, Fourth Edition. New Jersey: Prentice Hall.

196

Mazerolle, L., Wickes, R. & McBroom, J., 2010. Community variations in violence: The role of social ties and collective efficacy in comparative context. Journal of Research in Crime and Delinquency, 47(1), 3-30. McDermott J. 2014. The Changing face of Colombian organized crime. Available: http://library.fes.de/pdf-files/bueros/la-seguridad/11153.pdf [Accessed 30 May 2016]. Medellín Como Vamos. 2007. Socialización encuesta de percepción ciudadana - 2007 [Online]. Available: http://www.Medellíncomovamos.org/socializaci-n-encuesta-de- percepci-n-ciudadana-2007 [Accessed 31 May 2016]. Medellín Como Vamos. 2015. Presentación: Informe de calidad de vida de Medellín, 2014. 2015 [Online]. Available: http://www.Medellíncomovamos.org/presentaci-n-informe- de-calidad-de-vida-de-medell-n-2014-2015 [Accessed 11 May 2016]. Maclean, K. 2014. The ‘Medellín Miracle’: The politics of crisis, elites and coalitions [Online]. Available: http://publications.dlprog.org/The%20Medellín%20Miracle.pdf [Accessed 30 May 2016]. Mair, S. 2005. The new world of privatised violence. New Trends in International Politics and Society, 47-62. Manwaring, M. G. 2005. Street gangs: The new urban insurgency. Carlisle, PA: Strategic Studies Institute: The War College. Martin, G. 2014. Medellín, Tragedia y Resurrección: Mafias, Ciudad y Estado, 1975-2013. Medellín: La Carreta Editores. McCord, J. 2003. Cures that harm: Unanticipated outcomes of crime prevention programs. The Annals of the American Academy of Political and Social Science, 587, 16-30. Medina Franco, G. 2006. Una historia de las milicias de Medellín [Online]. Available: http://bibliotecavirtual.clacso.org.ar/ar/libros/colombia/ipc/historiamilicias.pdf [Accessed 12 May 2016]. Melguizo, R. C. & Cronshaw, F. 2001. The evolution of armed conflict in Medellín: An analysis of the major actors. Latin American Perspectives, 28, 110-131. Meth, P. 2004. Using diaries to understand women’s responses to crime and violence. Environment and Urbanization, 16, 153-164. Minuto30. 11 May 2014. "Abrazaton por el respeto a la vida" en el dia de las madres. Fotos [Online]. Available: http://www.minuto30.com/abrazaton-por-el-respeto-a-la-vida-en- el-dia-de-las-madres/239814/ [Accessed 30 May 2016]. Minuto30. 25 June 2015a. En plena Plaza Botero la ciudadania iba linchar a un hombre

197

por ladron. [Online] Available: http://www.minuto30.com/en-fotos-en-plena-plaza- botero-la-ciudadania-iba-linchar-a-un-hombre-por-ladron/355250/ [Accessed 10 August 2016]. Minuto30. 26 November 2015b. EN VIDEO: Fueron Identificadas las victimas del triple homicidio en el barrio Giradot [Online]. Available: http://www.minuto30.com/video- fueron-identificadas-las-victimas-del-triple-homicidio-en-el-barrio-girardot/411696/ [Accessed 11 May 2016]. Minuto30. 10 May 2016. Proponen prohibir el dia de la madres por altos indices de violencia [Online]. Available: http://www.minuto30.com/proponen-prohibir-dia-de-la-madre- por-altos-indices-de-violencia/471225/ [Accessed 30 May 2016]. Miraglia, P., Ochoa, R. & Briscoe, I. 2012. Transnational organised crime and fragile states. [Online]. Available: http://www.crime-prevention- intl.org/fileadmin/user_upload/Publications/Transnational_organised_crime_and_frag ile_states_2012.pdf. [Accessed 30 May 2016]. Moncada, E. 2013. The politics of urban violence: challenges for development in the global South. Studies in Comparative International Development, 48, 217-239. Moser, C. & Holland, J., 1997. Urban Poverty and Violence in Jamaica. World Bank Publications. Moser, C. 2004. Urban violence and insecurity: An introductory roadmap. Journal of Environment and Urbanization, 16, 3-16. Moser, C. & McIlwaine, C. 2004. Encounters with violence in Latin America: Urban Poor Perceptions from Colombia and Guatemala. New York: Routledge. Moser, C. & McIlwaine, C. 2006. Latin American urban violence as a development concern: towards a framework for violence reduction. World Development, 34, 89-112. Moser, C. & McIlwaine, C. 2014. New frontiers in twenty-first century urban conflict and violence. Environment and Urbanization, 26, 331-344. Moser, C & Rodgers, D. 2012 Understanding the tipping point of urban conflict: Violence, cities and poverty reduction in the developing world. [Online] Available: http://siteresources.worldbank.org/EXTGOVACC/Resources/GlobalPolicyReport.pdf [Accessed 30 May 2016]. Muggah, R. 2012. Researching the urban dilemma: Urbanization, poverty and violence, International Development Research Centre. Ottawa.

198

Muggah, R. & Mulli, S. 2012. Rio Tries Counterinsurgency. [Online] Available: http://www.appg-popdevrh.org.uk/Rio%20tries%20counterinsurgency.pdf [Accessed 21 May 2016]. Muggah, R. 2014a. Deconstructing the fragile city: exploring insecurity, violence and resilience. Environment and Urbanization, 26, 345-358. Muggah, R. 2014b. How to protect fast growing cities from failing. [Online]. Available: https://www.ted.com/talks/robert_muggah_how_to_protect_fast_growing_cities_from _failing?language=en [Accessed 21 May 2016]. Muggah, R., 2015. A manifesto for the fragile city. Journal of International Affairs, 68(2), 19 Muggah, R & Giannini, R. 2015. Interactive map tracks murder rate worldwide [Online]. Available: https://theglobalobservatory.org/2015/05/homicide-monitor-brazil- mapping/ [Accessed 11 May 2016]. Naudé, C. M. B., Prinsloo, J. & Ladikos, A. 2006. Experiences of crime in thirteen African countries: results from the International Crime Victim Survey. [Online]. Available: www.unicri.it/services/library.../icvs/.../ICVS_13_African_countries.pdf. [Accessed 31 May 2016]. Neumayer, E. 2005. Inequality and violent crime: Evidence from data on robbery and violent theft. Journal of Peace Research, 42, 101-112. New York Times. 09 Janurary 1991. Medellín Journal. In Colombia's tortured city, elite lives in fear [Online]. Available: http://www.nytimes.com/1991/01/09/world/Medellín- journal-in-colombia-s-tortured-city-elite-lives-in-fear.html [Accessed 30 May 2016]. NGO1 (Non-Government Official one), May 2014 ‘Interview’ in PhD Interviews, ed. Caroline Doyle. Medellín, Colombia. NGO2 (Non-Government Official two), May 2014 ‘Interview’ in PhD Interviews, ed. Caroline Doyle. Medellín, Colombia. Norten, N., 2003, Feral cities. Naval War College Review, LVI (4), Autumn, 97-106. Oberwittler, D., 2004. A multilevel analysis of neighbourhood contextual effects on serious juvenile offending the role of subcultural values and social disorganization. European Journal of Criminology, 1(2), 201-235. O'Donnell, G. 1993. On the state, democratization and some conceptual problems: A Latin American view with glances at some postcommunist countries. World Development, 21, 1355-1369. OF1 (Official from research centre focused on researching quality of life in Medellín), June 2014 ‘Interview’ in PhD Interviews, ed. Caroline Doyle, Medellín, Colombia. 199

Otera, S. Y. 1990. International extradition and the Medellín cocaine cartel: Surgical removal of Colombian cocaine traffickers for trial in the United States. Loyola of Los Angeles International and Comparative Law Review, 13, 955. Patel, R.B & Burkle, F.M., 2012. Rapid urbanization and the growing threat of violence and conflict: A 21 st century crisis. Prehospital and Disaster Medicine, 27(02), 194-197. Patiño, J.E., Duque, J.C., Pardo-Pascual, J.E. and Ruiz, L.A., 2014. Using remote sensing to assess the relationship between crime and the urban layout. Applied Geography, 55, 48-60. Pereira, A. W. 2001. Virtual legality authoritarian legacies and the reform of military justice in Brazil, the Southern Cone, and Mexico. Comparative Political Studies, 34, 555- 574. Perlman, J. 2009. Mega-cities: The politics of urban exclusion and violence in the Global South. In Koonings, K & Kruijt, D (eds) Megacity’s Violence and its Consequences in Rio de Janeiro. London: Zed Books. Pouligny, B. 2010. State-society relations and intangible dimensions of state resilience and state building: A bottom-up perspective. [Online]. Available http://cadmus.eui.eu/handle/1814/13854 [Accessed 31 May 2016]. Presidencia de la Republica. 1991. Medellín: Re-encuentro con el futuro. Presidencia de la Republica: Medellín. PRIMED. 1996. PRIMED A successful experience in urban intervention. [Online] Available: http://unesdoc.unesco.org/images/0012/001297/129776eo.pdf [Accessed 31 May 2016]. Putnam, R. D. 1993. The prosperous community. The American Prospect, 4, 35-42. Putnam, R. D. 2001. Bowling Alone: The collapse and Rrevival of American Community. New York: Simon and Schuster. Rabasa, A. Boraz, S., Chalk, P., Cragin, K., Karasik,. T., Moroney, J., O'Brien, K. & Peters, J. 2007. Ungoverned territories: Understanding and reducing terrorism risks [Online]. Available: http://www.rand.org/content/dam/rand/pubs/monographs/2007/RAND_MG561.pdf [Accessed 30 May 2016]. Rapley, J. 2006. The new middle ages. Foreign Affairs- New York., 85, 95. Reid-Henry, S. & Sending, O. J. 2014. The “humanitarianization” of urban violence. Environment and Urbanization, 26, 427-442.

200

Resilient Cities. 2016. 100 resilient cities [Online]. Available: http://www.100resilientcities.org/#/-_/ [Accessed 11 May 2016]. Restrepo, W. & Velez, J.C. (1998) Estudio sobre las características del homicidio en Medellín. Informe de investigación. Instituto de Estudios Políticos. Medellín: Universidad de Antioquia. Restrepo, W., W. Perez & J. Velez (2012) Las políticas públicas de seguridad ciudadana. In: Alonso, M. A., W. Perez y J. Velez (Eds) Ensayos Sobre Conflicto, Violencia y Seguridad Ciudadana en Medellín, 1997-2007. Medellín Instituto de Estudios Políticos, Universidad de Antioquia. Riaño Alcalá, P. R. 2006. Jóvenes, memoria y violencia en Medellín: una antropología del recuerdo y el olvido. Medellín: Universidad de Antioquia. Richter P. & Miller, M. 2007. Colombia army chief linked to outlaw militias [Online]. Available: http://articles.latimes.com/2007/mar/25/world/fg-colombia25/2 [Accessed 30 May 2016]. Robben, A. & Nordstrom C. 1995. Introduction: The anthropology and ethnography of violence and sociopolitical conflict. In: Nordstrom, C., Robben, A., (eds.) Fieldwork under Fire: Contemporary Studies of Violence and Survival. Berkeley: University of California Press. Rodgers, D. 2001. Making danger a calling: Anthropology, violence and the dilemmas of participant observation [Online]. Available: http://eprints.lse.ac.uk/28314/1/WP6_DR.pdf [Accessed 31 May 2016]. Rodgers, D. 2002. " We live in a state of siege": Violence, crime, and gangs in post-conflict urban Nicaragua [Online]. Available: http://www.lse.ac.uk/internationalDevelopment/pdf/WP/WP36.pdf [Accessed 30 May 2016]. Rodgers, D. 2003. Youth gangs in Colombia and Nicaragua: new forms of violence, new theoretical directions? [Online]. Available http://www.kus.uu.se/poverty&violence/PovertyViolence.pdf. [Accessed 31 May 2016]. Rodgers, D. 2004. “Disembedding” the city: crime, insecurity and spatial organization in Managua, Nicaragua. Environment and Urbanization, 16(2), 113-124. Rodgers, D., 2009. Slum wars of the 21st century: gangs, mano dura and the new urban geography of conflict in Central America. Development and Change, 40(5), 949-976.

201

Rodgers, D. 2010. Urban violence is not (necessarily) a way of life towards a political economy of conflict in cities. [Online] Available: http://purocihle.rrojasdatabank.info/2010-20.pdf [ Accessed 31 May 2016]. Rodgers, D. & Jones, A, J. 2009. Youth violence in Latin America: An overview and agenda for research. In: Jones, A. J. & Rodgers, D. (eds.) Youth Violence in Latin America: Gangs and Juvenile Justice in Perspective. New York: Palgrave Macmillian. pp. 1-24. Rodgers, D. & Baird, A. 2015. Understanding gangs in contemporary Latin America. In: Scott, D. & David P. (eds.) Handbook of Gangs and Gang Responses. New York: Wiley. Rodgers, D. & Jensen, S. 2008. Revolutionaries, barbarians or war machines? Gangs in Nicaragua and South Africa. Socialist Register, 45. Rodgers, D. & Muggah, R. 2009. Gangs as non-state armed groups: The Central American case. Contemporary Security Policy, 30, 301-317. Roguski, M. & Tauri, J.M., 2013. Key issues effecting field researcher safety: A reflexive commentary. New Zealand Sociology, 28(1), 18. Roldan, M., 1999. Colombia: Cocaine and the" miracle" of modernity in Medellín. Cocaine: Global Histories.165-82. Röling, B. V. 1970. The United Nations–A General Evaluation. UN Law/Fundamental Rights. Two Topics in International Law. Alphen aan den Rijn: Sijthoff & Noordhoff International Publishers BV. Rotberg, R. I. 2004. Strengthening governance: Ranking countries would help. Washington Quarterly, 28, 71-81. Rozema, R. 2008. Urban DDR-processes: paramilitaries and criminal networks in Medellín, Colombia. Journal of Latin American Studies, 40, 423-452. Rubio, M. 1997. Perverse social capital—some evidence from Colombia. Journal of Economic Issues, 31, 805-816. Rubio, M. 2013. Colombia: coexistence, legal confrontation and war with illegal armed groups [Online]. Available: http://www.brookings.edu/~/media/press/books/2013/dangerousliaisons/dangerouslias ons_ch4_colombia.pdf [Accessed 31 May 2016]. Safe Street. 2015. Safe St take charge, make change [Online]. Available: http://www.safest.org/en/ [Accessed 30 May 2016]. Salazar, A. & Jaramillo, A. M. 1992. Medellín: las subculturas del narcotráfico. Medellín: Cinep. 202

Salazar, J. A. 1990. No nacimos pa Semilla, Medellín: Cinep. Sambanis, N. 2002. A review of recent advances and future directions in the quantitative literature on civil war. Defence and Peace Economics, 13, 215-243. Samper, S. 2011. Urban Regeneration in a context of violence: The case of the Favela-Bairro in Rio de Janeiro [Online]. Available: http://informalsettlements.blogspot.com.au/2011/06/urban-regeneration-in-context- of.html [Accessed 11 May 2016]. Samper, J. 2012. Urban Resilience in Situations of Chronic Violence Case Study of Medellín, Colombia. Massachusetts Institute of Technology MIT, 1-34. Sampson, R. J., Raudenbush, S. W. & Earls, F. 1997. Neighborhoods and violent crime: A multilevel study of collective efficacy. Science, 277, 918-924. Sampson, R. J., Morenoff, J. D. & Earls, F. 1999. Beyond social capital: Spatial dynamics of collective efficacy for children. American Sociological Review, 633-660. Sampson, R. J. & Raudenbush, S. W. 2004. Seeing disorder: Neighborhood stigma and the social construction of “broken windows”. Social Psychology Quarterly, 67, 319-342. Sampson, R.J. & Wikström, P.O., 2008. The social order of violence in Chicago and Stockholm neighborhoods: A comparative inquiry. Order, Conflict, and Violence, 97- 119. Schneider, S. 2013. Violence, organized crime, and illicit drug markets: a Canadian case study. [Online]. Available: http://spp.revues.org/1028 [Accessed 31 May 2016]. Schwandt, T. A. (1997). Qualitative Inquiry: A Dictionary of Terms. London: Sage. Schroder, I. & Schmidt, B. 2001. Introduction: Violent imaginaries and violence practices. In: Schroder, I., Schmidt, B. (eds.) Anthropology of violence and conflict. London: Routledge. Semana. 13 March 1983. Un Robin Hood Paisa [Online]. Available: http://www.semana.com/gente/articulo/un-robin-hood-paisa/2374-3. [Accessed 31 May 2016]. Semana. 27 June 1994. 'Yo fui el creador de los PEPES.' [Online] Available: http://www.semana.com/nacion/articulo/yo-fui-el-creador-de-los-pepes/22770-3 [Accessed 31 May 2016]. Semana. 29 September 1997. CONVIVIR en blanco y negro. [Online] Available: http://www.semana.com/nacion/articulo/convivir-en-blanco-negro/33785-3 [Accessed 31 May 2016].

203

Semana. 15 Janurary 2001a. Nosotros matamos a Jaime Garzón [Online]. Available: http://www.semana.com/nacion/articulo/nosotros-matamos-jaime-garzon/44785-3. [Accessed 31 May 2016]. Semana. 18 June 2001b. Noche de Horror [Online]. Available: http://www.semana.com/nacion/articulo/noche-horror/46350-3 [Accessed 31 May 2016]. Semana. 12 June 2005. Contra la oficina del terror [Online]. Available: http://www.semana.com/nacion/articulo/contra-oficina-del-terror/73077-3 [Accessed 30 May 2016]. Semana. 09 August 2008. El testamento de Carlos Castaño [Online]. Available: http://www.semana.com/nacion/articulo/el-testamento-carlos-castano/94444-3 [Accessed 30 May 2016]. Semana 17 June 2009. Los reinsertados de Medellín y la ‘donbernabilidad [Online]. Available: http://www.semana.com/opinion/articulo/los-reinsertados-Medellín- donbernabilidad/104236-3 [Accessed 30 May 2016]. Semana 02 July 2010. Masacre en discoteca de Envigado deja al menos muertos [Online] Available: http://www.semana.com/nacion/articulo/masacre-discoteca-envigado-deja- menos-muertos/118774-3 [Accessed 30 May 2016]. Semana. 11 August 2012. 20 meses tras el último capo de la Oficina de Envigado [Online]. Available: http://www.semana.com/nacion/articulo/20-meses-tras-ultimo-capo- oficina-envigado/262782-3 [Accessed 11 May 2016]. Semana. 10 August 2013a. Los ‘pactos del fusil’ [Online]. Available: http://www.semana.com/nacion/articulo/los-pactos-del-fusil/353677-3 [Accessed 11 May 2016]. Semana. 30 November 2013b. La otra paz que vive en Medellín [Online]. Available: http://www.semana.com/nacion/articulo/pacto-de-paz-en-Medellín/366521-3 [Accessed 11 May 2016]. Semana. 08 January 2014. El expediente de la bomba del parque Lleras en Medellín. [Online] Available: http://www.semana.com/nacion/articulo/bomba-en-el-parque-lleras-de- Medellín-expediente-qcontra-miembros-de-la-terraza/370240-3 [Accessed 30 May 2016]. Semana. 09 July 2016. Turismo toxico la muerte de varios extranjeros encendio alarmas en Medellín. [Online] Available: http://www.semana.com/nacion/articulo/turismo-

204

toxico-la-muerte-de-varios-extranjeros-encendio-alarmas-en-Medellín/481119 [Accessed 06 October 2016]. Shapiro, J. N., Fair, C. & Rais, R. B. 2012. Political Violence in Pakistan 1988–2010: Patterns and Trends. London: International Growth Centre. Shaw, C. R. & McKay, H. D. 1942. Juvenile Delinquency and Urban Areas. Chicago: University of Chicago Press. SISC (Sistema de Información para la Seguridad y Convivencia). 2016. Sistema de Información para la Seguridad y Convivencia SISC [Online]. Available: https://www.Medellín.gov.co/irj/portal/ciudadanos?NavigationTarget=navurl://3ba1d de38616fc5bd67d2886146f02d6 [Accessed 30 May 2016]. Simpson, G. 1993. Explaining endemic violence in South Africa. [Online] Available: http://www.csvr.org.za/index.php/publications/1799-explaining-endemic-violence-in- south-africa.html [Accessed 31 May 2016]. Stepputat, F. & Engberg-Pedersen, L. 2008. II. Fragile States: Definitions, Measurements and Processes. Fragile Situations: Background Papers, 31. Stewart, F. & Brown, G. 2010. Fragile States: CRISE Overview 3.1. Centre for Research on Inequality, Human Security and Ethnicity. [Online] Available: http://www.qeh.ox.ac.uk/pdf/pdf-research/crise-ib5 [Accessed 31 May 2016]. Sullivan, J. P. & Bunker, R. J. 2007. Third generation gang studies: An introduction. [Online] Available: http://www.ngcrc.com/journalofgangresearch/jour.v14n4.sullivan.pdf [Accessed 31 May 2016]. Sullivan, J. P. & Bunker, R. J. 2011. Rethinking insurgency: Criminality, spirituality, and societal warfare in the Americas. Small Wars & Insurgencies, 22, 742-763. Tedesco, L. 2007. The Latin American state:―failed or evolving? [Online] Available: http://fride.org/descarga/WP37_LatState_ENG_may07.pdf [Accessed 11 May 2016]. The Economist. 07 May 2014. Medellín's comeback The trouble with miracles [Online]. Available: http://www.economist.com/news/americas/21603432-transformation- colombias-second-city-will-be-hard-copy-trouble-miracles [Accessed 11 May 2016]. The Guardian. 01 August 2013a. Medellín: The fast track from the slums [Online]. Available: https://www.theguardian.com/world/2013/jul/31/Medellín-colombia-fast- track-slums-escalators [Accessed 30 May 2016]. The Guardian. 09 June 2013b. Medellín, Colombia: Reinventing the world's most dangerous city [Online]. Available: https://www.theguardian.com/world/2013/jun/09/Medellín- colombia-worlds-most-dangerous-city [Accessed 30 May 2016]. 205

The Guardian. 17 April 2014. From Murder Capital to Model City: Is Medellín’s Miracle Show or Substance? [Online]. Available: https://www.theguardian.com/cities/2014/apr/17/Medellín-murder-capital-to-model- city-miracle-un-world-urban-forum [Accessed 21 May 2016]. Tilly, C. 1985. War-making and state-making as organized crime. In Evans, P., Rueschemeyer, D. Skocpol, T. Bringing the State Back In. Cambridge: Cambridge University Press. Ugarte, E.F. and Turner, M.M. 2011. What is the ‘Abu Sayyaf’? How labels shape reality. The Pacific Review, 24(4), 397-420. United Nations. 2014. Violence in Central Africa threatening efforts to eliminate Lord's Resistance Army, Security Council Notes. [Online] Available: http://www.un.org/press/en/2014/sc11390.doc.htm [Accessed 05 October 2016]. UNDP (United Nations Development Programme). 2013. Summary Regional Human Development Report 2013-2014 Citizen Security with a human face Evidence and Proposals for Latin America [Online] Available: http://hdr.undp.org/sites/default/files/citizen_security_with_a_human_face_- executivesummary.pdf [Accessed 11 May 2016]. UNDP (United Nations Development Programme) 2015. Human Development Report 2015 Colombia. [Online] Available: http://hdr.undp.org/sites/all/themes/hdr_theme/country-notes/COL.pdf [Accessed 06 October 2016]. UN-Habitat. 2007. Report on Enhancing Urban Safety and Security: Global Report on Human Settlements 2007 [Online]. Available: http://www.poa- iss.org/CASAUpload/Members/Documents/22@Human_Settlements_2007.pdf [Accessed 30 May 2016]. UN-Habitat. 2011. Building Urban Safety through Slum Upgrading [Online]. Available: http://mirror.unhabitat.org/pmss/listItemDetails.aspx?publicationID=3222&AspxAuto DetectCookieSupport=1 [Accessed 30 May 2016]. UN-Habitat. 2014. Propotion of Urban Population living in slums 1990-2014 [Online]. Available: http://unhabitat.org/wp-content/uploads/2014/03/Table-2.2-Proportion-of- urban-population-living-in-slums-1990-2014.pdf [Accessed 06 October 2016]. UNODC (United Nations Office of Drugs and Crime). 2000. United Nations Convention against Transnational Organized Crime and the Protocols Thereto. [Online] Available: https://www.unodc.org/unodc/treaties/CTOC/ [Accessed 30 May 2016]. 206

UNODC (United Nations Office of Drugs and Crime). 2002. Global illicit drugs trends 2002. [Online] Available: https://www.unodc.org/pdf/report_2002-06-26_1/report_2002-06- 26_1.pdf [Accessed 30 May 2016]. UNODC (United Nations Office of Drugs and Crime). 2012. Transnational Organized Crime in Central America and the Caribbean A Threat Assessment [Online]. Available: https://www.unodc.org/documents/data-and- analysis/Studies/TOC_Central_America_and_the_Caribbean_english.pdf [Accessed 30 May 2016]. UNODC (United Nations Office of Drugs and Crime). 2013. Global Study on Homicide [Online]. Available: https://www.unodc.org/documents/gsh/pdfs/2014_GLOBAL_HOMICIDE_BOOK_w eb.pdf [Accessed 30 May 2016]. Uran, O. 2010. Medellín Participatory Creativity in a Conflictive City. In Pearce J Participation and Democracy in the Twenty-First Century City. London: Palgrave Macmillan. pp. 127-153. USAID (United States Agency for International Development). 2016. What works in Reducing Community Violence: A Meta-Review and Field Study For The Northern Triangle [Online]. Available: https://www.usaid.gov/sites/default/files/USAID-2016-What- Works-in-Reducing-Community-Violence-Final-Report.pdf [Accessed 30 May 2016]. Van Dijk, J., Kesteren, J. V. & Smit, P. 2007. Criminal victimisation in international perspective. Key findings from the 2004-2005 ICVS and EU ICS. [Online]. Available: http://www.unicri.it/services/library_documentation/publications/icvs/publications/IC VS2004_05report.pdf. [Accessed 10 May 2016]. Vanderschueren, F., 1996. From violence to justice and security in cities. Environment and Urbanization, 8(1), 93-112. Valdez, A. & Kaplan, C. D. 1998. Reducing selection bias in the use of focus groups to investigate hidden populations: The case of Mexican-American gang members from South Texas. Drugs & Society, 14, 209-224. Van Meter, K. M. 1990. Methodological and design issues: Techniques for assessing the representatives of snowball samples. NIDA Research Monograph, 98, 31-43. Velez, M.B., 2001. The role of public social control in urban neighborhoods: A multilevel analysis of victimization risk. Criminology, 39(4), 837-864.

207

Verdad Abierta. 2016. Bloque Metro. [Online] Available: http://www.verdadabierta.com/victimarios/418-bloque-metro [Accessed 05 May 2016]. Vinci, A. 2006. The "problems of mobilization" and the analysis of armed groups. Parameters, 36, 49. Von Lampe, K.. 17 November 2015. Organized Crime Research [Online]. Available: http://www.organized-crime.de/ [Accessed 12 May 2016]. Wall Street Journal. 2012. Medellín: The innovative city of the year [Online]. Available: http://online.wsj.com/ad/cityoftheyear [Accessed 11 May 2016]. Wesely, J.K., 2006. Negotiating myself Theimpact of studying female exotic dancers on a feminist researcher. Qualitative Inquiry, 12(1), 146-162 Williams, P. 2001. Transational criminal networks. Networks and netwars: The future of terror, crime, and militancy, 1382, 61. Williams, P. 2008. Violent non-state actors and national and international security. Zurich: ETH Zurich. Wilson, W. J. 1996. When Work Disappears: The New World of the Urban Poor. New York, NY: Alfred Knopf. Winton, A. 2004. Young people’s views on how to tackle gang violence in “post-conflict” Guatemala. Environment and Urbanization, 16, 83-99. Winton, A. 2014. Gangs in global perspective. Environment and Urbanization, 26, 401-416. Wirth, L. 1938. Urbanism as a Way of Life. American Journal of Sociology. 44, 1-24. Wolff, M. J. 2015. Building criminal authority: a comparative analysis of drug gangs in Rio de Janeiro and Recife. Latin American Politics and Society, 57, 21-40. WHO (World Health Organisation). 2014. Global status report on violence prevention 2014 [Online]. Available: http://www.who.int/violence_injury_prevention/violence/status_report/2014/en/ [Accessed 31 May 2016]. World Bank. 2011a. Violence in the city: Understanding and supporting community responses to urban violence [Online]. Available: http://documents.worldbank.org/curated/en/2011/04/14831166/violence-city- understanding-supporting-community-responses-urban-violence [Accessed 30 May 2011].

208

World Bank. 2011b. World Development Report [Online]. Available: http://siteresources.worldbank.org/INTWDRS/Resources/WDR2011_Full_Text.pdf [Accessed 31 May 2016]. World Bank. 2016a. Colombia. [Online] Available: http://data.worldbank.org/country/colombia [Accessed 06 October 2016]. World Bank. 2016b. Urban violence a challenge of epidemic proportions. [Online]. Available: http://www.worldbank.org/en/news/feature/2016/09/06/urban-violence-a- challenge-of-epidemic-proportions [Accessed 06 October 2016]. Yin, R. 2003. Case Study Research: Design and Methods 3rd Edition. Thousand Oaks: Sage. Zartman, I. W. 1995. Collapsed Dates: The Disintegration and Restoration of Legitimate Authority. Boulder: Lynne Rienner Publishers.

209

Appendix 1: Semi-structured Interview Questions

1. What factors do you think led to the emergence of violence in Medellín?

a. ¿Qué factores cree usted que explican la aparición de la violencia en Medellín?

2. Do you think there was a particular event that led to emergence of violence in Medellín? (E.g. the assassination of a political leader, riots in a particular area)

a. ¿Cree usted que hay un evento particular que tiene relación con la aparición de la violencia en Medellín? Por ejemplo el asesinato de un líder político o la urbanización de Medellín.

3. What relationship do you think political, social and economic levels have in regards to the emergence of violence in Medellín?

a. ¿Qué relación tienen los diferentes factores políticos, sociales y económicos con la aparición de la violencia en Medellín?

4. Around what year did the violence start to decline in Medellín?

a. ¿Puede proveer un estimado del año en que la violencia empezó a declinar en Medellín?

5. Was there any significant event (e.g. speech by member of government, certain propaganda used, capture of significant person by authorities) during that year which could be linked to this decline?

a. ¿Cree usted que existió un evento durante dicho año (por ejemplo, el discurso de un miembro del gobierno local o nacional, propaganda del gobierno local o nacional, la captura de un individuo importante de los cárteles) que tiene relación con la disminución de la violencia?

6. How important do you think the promotion of the ‘community’ has been in assisting in the reduction of violence?

210

a. ¿Cuán importante cree usted cree que ha sido la promoción y participación de la “comunidad” en la reducción de la violencia en Medellín?

7. Can you provide a particular example of ‘community engagement’ in Medellín?

a. ¿Puede citar un ejemplo en el cual la comunidad se hubiera unido con el objetivo de prevenir o evitar violencia en Medellín?

8. How have you seen success in regards to violence reduction in Medellín?

a. ¿En qué aspectos ha impactado positivamente la disminución de la violencia en Medellín? Por ejemplo, la reducción del número de asaltos o el cambio de la actitud de los ciudadanos.

9. What do you think have been the most important factors/influences in regards to the reduction of violence in Medellín?

a. ¿Cuáles han sido en su opinión los factores/influencias más importantes en la disminución de la violencia en Medellín? ¿Puede decirse que es el resultado de una campaña del gobierno, gracias a la comunidad o algo más?

10. How important has the local government policies and NGO’s been in the reduction of violence in Medellín?

a. ¿Qué importancia tienen los políticas del gobierno local (y/o el gobierno nacional) y de organizaciones como la ONU en la disminución de la violencia en Medellín?

211