Criminals, Cops, and Politicians: The Dynamics of Drug Violence in Colombia and Mexico A dissertation presented by Angélica Durán-Martínez B.A. in Political Science, Universidad Nacional de Colombia M.A. in Latin American and Caribbean Studies, New York University To the Department of Political Science In partial fulfillment of the requirements for the Degree of Doctor of Philosophy in the subject of Political Science Brown University Providence, Rhode Island May 2013     © Copyright 2013 by Angélica Durán-Martínez All rights reserved   ii     This dissertation by Angelica Duran-Martinez is accepted in its present form by the Department of Political Science as satisfying the dissertation requirement for the degree of Doctor of Philosophy Date ___________ __________________________________ Richard O. Snyder, Advisor Recommended to the Graduate Council Date ___________ __________________________________ Peter Andreas, Reader Date ___________ __________________________________ Pauline Jones-Luong, Reader Approved by the Graduate Council Date ___________ __________________________________ Peter Weber, Dean of the Graduate School   iii     Curriculum Vitae Angelica Duran-Martinez was born in Bogotá, Colombia, on May 14, 1979. She obtained a B.A. in Political Science from Universidad Nacional de Colombia and completed an M.A. in Latin American and Caribbean Studies at New York University thanks to a Fulbright Scholarship. She has been a Fulbright Fellow at the United Nations Secretariat and a consultant for the UN Office on Drugs and Crime, the UN Development Program, and Global Integrity. In Colombia she worked for the Foundation Ideas for Peace (FIP) and for several research projects about the armed conflict. Her dissertation research received funding from the Jennings Randolph Peace Scholarship of the United States Institute of Peace (USIP), the International Dissertation Research Fellowship of the Social Science Research Council (IDRF-SSRC), and the Drugs, Security, and Democracy Fellowship of the SSRC, Open Society Foundation, and IDRC. Her research also received support from the Graduate Program in Development, the Center for Latin American and Caribbean Studies, the Graduate School, and the Watson Institute for International Studies at Brown University. She has authored “Presidents, Parties and Referenda in Latin America”, (Comparative Political Studies, 2012) and coauthored "Does illegality breed violence?: Drug trafficking and state-sponsored protection rackets" with Richard Snyder (Crime, Law and Social Change, 2009). She has also coauthored “The politics of drugs and illicit trade in the Americas” with Peter Andreas, in Kingstone and Yashar, eds., Handbook of Latin American Politics, Routledge (2012).   iv     Acknowledgements In 1989, when I was 10 years old, I was walking with my mom two blocks away from our home in Bogotá, and a bomb went off at a nearby mechanic shop. It was the time of the narcoterrorist terror. Only a few seconds saved us from being victims of that bomb. As people rushed into the scene, my mom ran away in the opposite direction, confused, scared, and thankful that we were safe. That was my closest experience to violence while growing up; living in Bogotá, I was relatively protected from the violence that besieged other rural areas, but in that period narcoterrorism overwhelmed the major cities of Colombia. Over the years, I have come to believe that such experience generated my interest in the topic of violence, drug trafficking, and crime. However, such interest would have never been materialized without years of discussion, mentoring, support, and valuable teachings, both at the core of family discussions, and during my years studying Political Science at the Universidad Nacional de Colombia. Thus it would be impossible for me to thank appropriately all the people that have been crucial in my intellectual formation, and thus, essential for the successful completion of this project. I have to limit for those who directly contributed to making this project possible. Since I started a Ph.D. at Brown, the intellectual guidance and personal support of Richard Snyder, Pauline Jones Luong, and Peter Andreas were crucial both in crafting out my research proposal, and in producing the dissertation. As the chair of my dissertation committee, Richard went out of his way to mentor me, and a combination of encouragement and thoughtful criticism always pushed me in new directions and forced   v     me out of my comfort zones. Peter always posed sharp questions that illuminated all the crucial holes in my thinking and thus encouraged me to sharpen ideas, to be careful with words, and to think broadly. Pauline’s comments always forced me to reflect on crucial methodological questions, on broader theoretical debates, and on key conceptual dilemmas. Barbara Stallings has provided feedback over the years and supported my project. My special thanks also go to Melani Cammett, Matthew Guttman, Kay Warren, and the entire faculty at Brown that at different points encouraged my work. I spent 15 months conducting fieldwork in Colombia and Mexico, and before the project actually took shape I carried out short trips to Cali, Medellin and Mexico City. All these trips were possible thanks to the generous funding from the United States Institute of Peace USIP, the Social Science Research Council SSRC, the Drugs, Security, and Democracy DSD Fellowship of the SSRC, Open Society and IDRC, and grants from the Graduate Development Program, the Center for Latin American and Caribbean Studies, and the Watson Institute for International Studies at Brown University. This research would have not been possible without the generosity of those who shared their time, thoughts, and experiences with me while I was in the field. I cannot thank all of them directly because of anonymity concerns, but every person who shared experiences with me made this project what it is today. I am forever indebted to Maria Eugenia and Lucila in Medellin; Mireya and Blanca in Cali; Luz Maria, Pedro, and Pedro Pablo in Culiacan; Alan and Laura in Ciudad Juarez, who generously opened the doors of their houses and their hearts for me, and became a family in each of my research locations, making me feel safer and welcomed. Luis Astorga, both an intellectual guide and a friend, graciously introduced me to his family and provided key contacts and   vi     insights. My special thanks go to Angelika Rettberg at the Universidad de los Andes in Bogotá, Arturo Alvarado at Colmex in Mexico City, Alberto Hernandez at COLEF, Mauricio Romero at the Universidad Javeriana in Bogotá, Roddy Brett at the time at the Universidad del Rosario, and David Shirk at the Transborder Institute at the University of San Diego, for having provided key institutional support and space for me to share the preliminary findings of my research while in the field. The research assistance of Mundo Ramirez, Rocio Duran, and Luis Cañas was invaluable, thanks to them for their hard work, insights and patience to go through hundreds of news papers, in most cases dusty records, to create the dataset on drug violence. The work I present here reflects the knowledge that all my interviewees shared, and many discussions with friends and colleagues. The usual disclaimer is always necessary that my omissions are my sole responsibility. Colleagues at fellowship workshops with USIP, SSRC, and the DSD program, provided crucial insights, criticisms, and encouragement at different stages of the project. The intellectual community of the DSD program was essential in the latter part of fieldwork and has remained a highly motivating network for research on topics of security, criminality, and violence in Latin America. Special thanks go to Desmond Arias, Ana Arjona, Susan Brewer, Yanilda Gonzalez, Javier Osorio and Reynaldo Rojo for their feedback on the general ideas of this dissertation. My colleagues and friends at Brown were crucial and my special thanks go to Sinem Adar, Erin Beck, Jennifer Costanza, Susan Ellison, Eduardo Moncada, and Dikshya Thapa, who read earlier drafts of chapters and made great comments on them. Susan Hirsch and Suzanne Brough always provided essential administrative support throughout my years at Brown.   vii     Brown University was my intellectual home, and the friends that I made over the years became my family away from home, to all of them my special thanks because their trust and friendship made my journey special and unforgettable. I also thank my longtime friends from Universidad Nacional and New York University because they have always accompanied me, despite distance and time. Astrid, Jessica, and Paola will always be my accomplices. None of this could have ever been possible without the unconditional support of my family. Of my mom, whose endless love and trust in me keeps me going in moments of doubt, and whose sacrifices and guidance made me the person I am today. To Nancy and Rocio, my precious sisters, who never stop encouraging me, motivating me to go further, and teaching me about life. To my loving husband Jack, who came to my life when this project was just starting and has supported me through every single step of it, enduring moments of physical separation, editing chapters, inspiring me through long political conversations, reminding me about the value of my work, but also about the real important things in life. Jack you have been a light in my life and I can never express in words how grateful I am to have you in my life. Finally, to my dad, my hero, who left this world too early to see how his sacrifices pay off, but who, I am sure, still looks out for me from heaven. To them, I dedicate this dissertation.   viii     TABLE OF CONTENTS Acknowledgements ………………………………………………………………… v List of Tables ……………………………………………………………………….. x List of Figures ………………………………………………………………………. xi Glossary of terms and abbreviations………………………………………………… xii Chapter 1. State power, criminal competition, and drug violence: A systematic comparison between Colombian and Mexican cities………………………………… 1 Chapter 2. Silent traffickers or brutal criminals: defining and assessing the frequency and visibility of drug violence ……………………………………………………………. 50 Chapter 3. The state capacity to enforce the law and state autonomy from criminal influence in contexts of institutional transformation …………………………………. 98 Chapter 4. From a “perfect storm” of violence to unstable pacification: the case of Medellin ………………………………………………………………………………..157 Chapter 5. The “businessmen” and the “gentlemen”: low visibility violence and state- criminal symbiosis in Cali and Culiacán ………………………………………………212 Chapter 6. Beyond a border tale: The divergent violent trajectories of Ciudad Juárez and Tijuana …………………………………………………………………………………259 Conclusion ……………………………………………………………………………..331 Bibliography …………………………………………………………………………...352 Appendix ………………………………………………………………………………386   ix     List of Tables Table 1.1. Definition and components of each variable Table 2.1. Indicators of high visibility and low visibility violence Table 2.2. Dimensions of visibility in Colombian and Mexican cities Table 4.1. Methods used in violent attacks, Medellin 1984 and 1989 Table 4.2. A political approach to the evolution of drug violence in Medellin Table 5.1. Cali and Culiacán: Basic statistics Table 5.2. Trajectories of Drug Violence in Cali and Culiacán Table 6.1. Methods used in violent attacks, Tijuana Table 6.2. Methods used in violent attacks, Ciudad Juárez Table 6.3. A political economy approach to the evolution of drug violence in Ciudad Juarez and Tijuana   x     List of Figures Figure 1.1. Types of drug related violence and location of cases Figure 1.2. A dynamic depiction of drug violence Figure 1.3. Trajectories of violence in each city 1984-2010 Figure 2.1. Types of drug related violence and location of cases Figure 2.2. Homicide rates 1985-2010, Mexican cities Figure 2.3 Homicide rates 1984-2010, Colombian cities Figure 2.3. Percentage of violent acts by method Figure 3.1. Coca cultivation and potential cocaine production in Colombia Figure 3.2. Number of new drug users in Mexico 1984-2008 Figure 4.1. Homicide rates in Medellin 1984-2010 (per 100,000 inhabitants) Figure 5.1. Evolution of homicide rates in Cali Figure 5.2. Occupation of victims of violence in Cali 1984-1989 Figure 5.3. Homicide rates Culiacán 1985-2010 Figure 5.4. Proportion of homicides by age group 1985-1996 Figure 6.1. Homicide rates in Tijuana 1985-2010 Figure 6.2. Homicide rates in Cd Juárez 1985-2010   xi     Glossary of Terms and Abbreviations AFI Federal Investigation Agency, Mexico AUC United Self Defense Forces of Colombia BACRIM Criminal Gangs CENDRO Center for Drug Control Planning, Mexico CISEN Intelligence and National Security Center, Mexico DANE National Statistics Department, Colombia DAS Intelligence Department, Colombia DEA Drug Enforcement Administration DIJIN Director of Criminal Investigation and Interpol, Colombia DTO Drug Trafficking Organization ELN National Liberation Army ERPAC Popular Revolutionary Anti-insurgent Army of Colombia EZLN National Liberation Zapatista Army FARC Revolutionary Armed Forces of Colombia FEADS Special Attorney for Crimes against Health, Mexico INCD National Institute for the Combat of Drugs, Mexico INEGI National Statistics and Geography Institute, México M-19 April-19 Movement, Colombia MAS Death to Kidnappers   xii     MOE Electoral Observation Mission NGO Non governmental organization PAN National Action Party, Mexico Pepes Persecuted by Pablo Escobar PF Federal Police PFP Federal Preventive Police PGR Attorney General’s Office, Mexico PIN National Integration Party, Colombia PJF Federal Judicial Police PRD Democratic Revolution Party, Mexico PRI Revolutionary Institutional Party, Mexico SEDENA National Defense Secretariat, Mexico SEGOB Government Secretariat, Mexico SEMAR Marine Secretariat, Mexico SINAIS National System of Health Information, Mexico SNSP National System of Public Security, Mexico SSP Public Security Secretariat, Mexico UEDO Specialized Unit in Organized Delinquency, Mexico   xiii   CHAPTER 1. STATE POWER, CRIMINAL COMPETITION, AND DRUG VIOLENCE: A SYSTEMATIC COMPARISON BETWEEN COLOMBIAN AND MEXICAN CITIES. “After April 30th 2008 a collective psychosis started. On May 5th five federal police were killed. We went from deaf violence to high impact violence, in which there were shootouts downtown.” (Local journalist, Culiacán, March 2011) “The situation today is not as bad, because of what was done back then [in the 90s] the Mafiosi changed their profile. One does not feel scared of going to 70th Street as before. The problem of the narco can even be worse now, but the fact is that the bombs made it a more aggressive period [in the 1980s].” (Local councilwoman, Medellín, October 2010) “When I arrived to Ciudad Juárez in 1988, there was coordination in everything, there was only one [trafficking] group. There were executions, but the bodies disappeared, they were not thrown out in the streets.” (Former police, Mexico City, June 2011) The former statements exemplify many of the responses I received when I asked people to describe drug violence in their cities. To my surprise, my interviewees in many different sectors not only identified an increase or a decrease in violence, but most importantly, changes in the criminal’s willingness to expose the evidence of their attacks. In Culiacán, a city long besieged by drug trafficking violence, people identified 2008 as a crucial year in which violence became more visible. By contrast, in Medellín people talked about the more hidden character of criminal violence in recent years. This variation presented several puzzles, such as why would criminals decide to expose violence, if doing so can attract attention that can be detrimental for their business? And why would criminals hide violence if violence can be used to show power and scare away   1   enemies? In this dissertation I aim to explain these puzzles and the enormous variation in patterns of violence within countries similarly afflicted by drug trafficking. The relationship between illegal markets and violence has been a crucial concern both in the study of criminal organizations and in the study of civil wars and conflict. Illegal markets like drug trafficking are expected to fuel violence, either as a result of greed in civil wars, or as a result of the increasing power and extension of organized crime. This expectation has limited our understanding of the conditions driving the dramatic changes in violence that countries experience over time, such as the upsurge of drug violence in Mexico, a country that has been home to trafficking organizations for over six decades, or the decrease of violence in Colombia despite the persistence of trafficking in the country. Furthermore, we lack analytical tools to capture puzzling variation within countries with similar levels of drug trafficking. For example during the 1980s, a highly violent period in Colombia, the city of Medellín was significantly more violent than the city of Cali. This dissertation explains variation in types of drug violence by comparing five cities that have experienced contrasting patterns of drug violence over the past three decades (1984-2012): Cali and Medellín in Colombia, and Ciudad Juárez, Culiacán, and Tijuana in Mexico. I argue that the interaction between two critical factors -the state security apparatus and the structure of illegal drug markets- determines the incentives and opportunities of drug traffickers to employ violence. Thus I build and test a novel political economy framework that treats drug violence not as a phenomenon that emerges in the margins of the state, but rather as one constructed through interactions between the state and criminal actors.   2   To characterize the varied types of violence that exist in drug markets, and to understand the impact of, and reactions to, drug violence, this study unpacks violence not only looking at the frequency, but also including a novel dimension in the understanding of violence: visibility. The case studies will show that the frequency of violence increases as the structure of the illegal market shifts from monopolistic to competitive, and the visibility of violence increases as the structure of the state security apparatus shifts from cohesive to fragmented, because cohesion enables local authorities to be able to make a credible commitment to protect or, alternatively, to persecute criminals. Criminal actors, in turn, may refrain from using visible violence that can trigger state reactions when they receive credible protection from the state and fear losing it, or when they believe that the state has the capacity to dismantle them. Thus, distinct configurations of frequency and visibility reflect particular interactions between the state and the drug market. While advancing this argument, I bring the state and law enforcement into the center of understanding dynamics of violence by unpacking power relations within the state security apparatus and their impact on the ability of the state to credibly protect, or alternatively, effectively attack criminals. This study also explores an often-overlooked aspect of organization in criminal markets that has crucial implications for violence: the type of armed coercion criminals employ, and more specifically whether violence is “insourced” (that is, vertically integrated into the organization) or “outsourced” (that is, contracted out to youth gangs). I find that when criminals outsource violence they lose the ability or willingness to discipline their armed forces. This in turn, explains why some competitive illicit markets experience extreme spikes of violence.   3   1.1. The multiple dimensions of drug related violence Drug violence, defined as acts of lethal violence that emerge in the functioning of drug markets, can involve civilians, state officials, and criminals alike. It is a crucial phenomenon that constitutes an increasing source of instability and a threat to governance and citizen security in places as diverse as Afghanistan, Bolivia, Brazil, El Salvador, Guatemala, Guinea Bissau, Honduras, India, Iran, Italy, Myanmar, Pakistan, Peru and Russia. Colombia and Mexico represent paradigmatic examples of this phenomenon as they have been at the frontline of the war on drugs and suffer the deleterious consequences of violence. In Mexico drug violence caused more than 60,000 deaths between 2006 and 2011,1 and in Colombia drug trafficking contributes to intensifying the longest standing armed conflict in the world. Existing knowledge provides contradictory predictions about how much violence we should expect in drug markets. On the one hand, violence can be seen as a rare and irrational strategy because it attracts enforcement attention, and criminals thus may prefer to limit its use. On the other, it can be expected to be frequent, because in illegal markets that lack state regulation, violence is deployed strategically to gain market share, solve contractual problems (Andreas and Wallman 2009, Friman 2009) and to address disciplinary and succession issues (Reuter 2009). These two expectations in turn, raise two interrelated but opposite puzzles: if violence is inherent to the drug trade, when can drug traffickers control their use of violence? And why do drug traffickers decide to use                                                                                                                 1 This calculation of drug homicides in Mexico is based on Government estimates of homicides associated with drug trafficking. The definition of a “drug related homicide” is controversial, and thus this number presents a conservative estimate of the death toll in Mexico. Overall homicide statistics add up to 120,000 deaths between 2006 and 2012 (Molzahn, Rodriguez, and Shirk 2013).   4   visible violence if doing so can attract unnecessary enforcement and media attention, which can be detrimental for their business? To address these puzzles I argue it is necessary to evaluate the frequency of violence, that is, the rate at which violent events occur, because violence can sometimes be a rare event, and other times be a frequent occurrence in drug markets. But I also argue that the full “repertoire” of violence includes another dimension, its visibility. Visibility refers to whether criminals publicly display the evidence and/or claim responsibility for their attacks.2 Visibility thus encompasses the methods employed by killers, the type and number of victims per attack, and the location of violent attacks. The importance of considering visibility was depicted by an intelligence official I interviewed, while describing the violence deployed by current drug trafficking organizations in Colombia: “the BACRIM3 dismember bodies with machetes, of course, they are not as sophisticated as before when they used chain saws because they don’t want to give a message, it’s simply brutality. In the past they used techniques such as impalements, but now we see another type of violence, a new profile of the killer, there is more interest in hiding, but the new structures are absolutely violent.4” Intelligence officers interviewed in Colombia echoed this idea claiming that as a response to                                                                                                                 2 Visible violence is more likely to attract enforcement and media attention, but attracting such attention is not a necessary objective for criminals. 3 Acronym used by the Government of Colombia to refer to criminal structures devoted to drug trafficking, stands for Criminal Gangs (Bandas Criminales). The term is controversial as these groups are linked to old paramilitary structures, and thus many consider the term neoparamilitaries to be a better way to label these groups. 4 Author’s interview conducted with intelligence official, General Attorney’s Office (CTI-Fiscalia). Bogotá, September 13, 2010. The emphasis is mine because the key in this statement is not the official’s assumption that there is no message in violence, but that criminals try to hide its use.   5   enforcement actions, traffickers have learned that it is better for them to be “traffickers in silence.”5 Visibility thus captures crucial, but overlooked changes in the use of violence by drug traffickers, which are determined by the latter’s interactions with the state;6 this is illustrated in the testimony of a paramilitary commander in Colombia when he was asked why paramilitaries used methods like mass graves, cremations, or throwing bodies in rivers so they could not be found: “Judge: Who made the decision to get rid of the people? Why get rid of the people? Why not simply kill them? Paramilitary: He [the commander] said, “if dead bodies start pilling up, it would make the authorities look bad. And it could be detrimental to us as well”.”7 The five cities analyzed in this dissertation have experienced contrasting changes in both the frequency and visibility of drug violence. In Colombia, between 1984 and 1993, during a period known as narcoterrorism, violence in Medellín was more frequent than in Cali. However, over time frequency in Medellín decreased whereas it remained stable in Cali8 even though the large Drug Trafficking Organizations (DTOs) of the 1980s had been dismantled in both cities. Likewise, since 2006 violence in Ciudad Juárez became more visible as criminals claimed responsibility for beheadings, threatened and killed public officials, and carried out massacres in a pattern that                                                                                                                 5 Expression used by intelligence official in author’s interview, Bogotá, September 20,2010. 6 I acknowledge that visible violence is not always instrumental or rational, and that visible violence can generate dynamics of self-reproduction. I explore some of these issues in Chapter II, but the emphasis of this dissertation is on explaining why criminals purposefully use visible violence in the first place. 7 Confession hearing of military commander Jorge Laverde. Footage on documentary Impunity, by Juan Lozano and Hollman Morris, 2011. 8 Medellín’s average homicide rate was 228 between 1984 and 1993 and fell to an average of 154 between 1994 and 2005, with a low point of 34 in 2005. By contrast, in the same periods Cali’s average was 68 and increased to 91, with a high peak of 104 in 1994.   6   contrasted with the silent and less visible disappearance of people that prevailed in the early 1990s. Figure 1.1 illustrates the configurations that emerge from combining the two dimensions of violence and the wide variation experienced both across cities and within each city over time. This study focuses on analyzing drug violence at the city level for several reasons: First, in the past two decades cities across the developing world have become a crucial setting for criminal violence (Davis 2010, Koonings and Kruijt 2007, Moser 2004). Second, cities concentrate economic and human resources that make them attractive for criminal actors. In cities criminals can connect a wide range of illegal activities (from informal economies to local drug markets and extortion), and opportunities for money laundering in large formal economies that are not readily available in the countryside. Due to urbanization and developmental processes, cities can host large numbers of impoverished populations that might be employed by criminals. Third, cities are strategic hubs to connect production areas, trafficking routes and key markets. Fourth, local actors have become increasingly crucial in security policies (Moncada 2011), and thus local dynamics of violence cannot be simply analyzed through the lens of national governments. Figure 1.1. Types of drug related violence and location of cases Low Frequency (LF) High Frequency (HF) Cali 84-88 Cali 89-10, Medellín 07-10 Low visibility (LV) Cd Juárez 84-94, Tijuana Cd Juárez 94-07, Tijuana 88- 84-87/10 95/98-08 Medellín 03-07 Culiacán 84-08 High visibility Tijuana 96-97 Cd Juárez 08-10 (HV) Cd Juárez 05-06 Medellín 84-93 Culiacán 08-10   7   1.1.1. The Aims of the Study Existing research about the drug trade is not well equipped to explain the sharp variation described above because it expects violence to be a natural byproduct of drug trafficking. Analyses that emphasize the increasing threat of violent transnational organized crime (Naim 2005),9 studies that link the existence of lootable goods (like drugs) to the duration of civil wars (Fearon 2004, Ross 2004), and analyses that include illegal profits as a distinctive component of new wars (Kaldor 2001, p.102) tend to create a direct connection between illegal markets like drug trafficking and violence, thus preempting the possibility of explaining variation of violence in illegal markets.10 Recently, there has been an increase of studies unpacking the connection between illegality and violence (Andreas and Wallman 2009, Garzon 2008, Snyder and Duran- Martinez 2009) and partially as a response to the increase of violence in Mexico there has been an upsurge of analyses explaining why violence in drug markets can reach extreme levels (Astorga and Shirk 2010, Lessing 2012a, Osorio 2012, Rios and Shirk 2010, Rios 2012, Special Issue Crime, Law and Social Change 2009). This dissertation builds on and expands this emerging body of research in several ways. First, I analyze systematically, cases of extreme violence, cases of relative peace, and the most common, but often under analyzed intermediate forms of frequent and low visibility violence which do not elicit much media or enforcement attention. Thus, I get                                                                                                                 9 Andreas (2011) presents an excellent critique of the misconceptions attached to the idea of transnational or globalized organized crime, one of which is the idea that as the power of global organized crime increases, so does the violence associated with it. 10 The relationship between markets, natural resources and civil war is still an ongoing issue of debate in the literature on civil wars. Within this debate many acknowledge that the effect of drugs, as a distinctive type of commodity, on war, is still far from being clarified (Ross 2005) and that the mechanisms explaining the connection are complex and not automatic (Cornell 2005). Yet very few of these debates explore the non-instances of the relationship, for example when rebel groups have not recurred to drug trafficking as a funding source, or when drug markets generate less violence.   8   beyond the selection bias in existing studies that infer the causes of violence using a truncated sample of the independent variable, looking only at cases of extremely high violence. I capture a fuller range of variation in drug violence, by studying five cities over almost three decades (1984-2010). Second, I introduce an often-overlooked political logic into the study of drug markets by considering violence as the product of the interaction between the state and criminal organizations. Third, by analyzing the frequency and visibility of violence I offer a nuanced but systematic analysis that unpacks the enormous variation that exists in illegal markets and tackles the most fundamental puzzles of violent action. To present the building blocks of my political economy approach to drug violence the second section of this chapter presents alternative explanations to drug violence and their limitations. The third section lays out the theory and the mechanisms that explain how the state security apparatus and the structure of illegal markets determine types of violence. The third section presents conceptual clarifications to define the structure of the illegal market and the structure of the state security apparatus and describes the configurations of violence produced by the interaction between security apparatus, criminal market, and type of armed coercion criminals employ. The last section describes the research design, methods, and the road map of the dissertation. 1.2. Beyond unidimensional and economically-focused explanations of drug violence Existing research about the drug trade does suggest potential explanations that offer important insights for explaining violence and that can be labeled as 1) economic oriented, 2) focused on international policy, and 3) focused on structural economic and   9   social processes. While these explanations provide crucial insights to understand violence, they do not suffice to explore systematically the puzzling variation countries experience over time, and within the same territory. 1.2.1. Economic approaches to drug violence Economic approaches contend that certain aspects of the drug trade chain (e.g., distribution), certain goods (e.g., cocaine), and higher volumes of trade generate more violence. The mechanism that links these aspects to violence is the level of profits. Yet, one of the limitations of an economic approach is precisely that the effect of profits on violence is indeterminate, and prices tend to be very volatile (Caulkins and Reuter 1998). In some explanations the rationale is that markets that generate more economic profits are likely to generate more violence as they create incentives for criminals to access the market and to protect their turf violently, as in the case of cocaine markets that are assumed to be more profitable and more violent than marijuana markets (Reuter 2009). Yet, in other explanations the rationale is the opposite, that lower or shrinking profits spur conflict because in order to maintain profits criminals need to invade rival territories (Goldstein 1997). Thus, economic approaches analyze differences across drug markets and account for associations such as the increasing incidence of cocaine trafficking in Mexico and increasing violence. Yet, they are limited to explain why violence can be associated both with high and low market prices, and both with shortages and oversupply of drugs. These approaches do not explain, for example, why cities that witnessed at the same time the presence of cocaine distribution organizations such as Medellín and Cali, still experienced strikingly different forms of violence.   10   Another variant of economic approaches contends that the diversification of markets and business portfolios (e.g. trafficking organizations getting involved in extortion, kidnapping, or local drug dealing) can increase violence. Here the rationale is that these highly localized markets force organizations to fight more fiercely and systematically for territorial control, as illustrated by strategies used by mafia organizations to establish protection rackets (Gambetta 1993, Skaperdas 2001). The problem with this reasoning is that diversification is more often a product, rather than a cause of violence. For example in the 1980s the Medellín DTO and its leader Pablo Escobar specialized in cocaine trafficking even though a myriad of other illegal markets existed in Medellín at the time. This situation changed as law enforcement pressure on Escobar mounted and his resources shrunk, making it necessary for him to tap into other illegal activities such as kidnapping and the control of local drug distribution to fund his war against the Colombian state (Medina Franco 2006). Similarly in Mexico the unprecedented increase in extortion has occurred after, not before, violence exploded in cities like Ciudad Juárez (Guerrero 2011). A political economy approach to violence does not discount the importance of economic aspects, and in fact considers market organization as a crucial variable to understand violence. Yet it provides a stronger basis to understand systematically variations in violence, and to explain why, under similar markets circumstances, criminals recur to very different violent strategies.   11   1.2.2. Explanations focused on international policy These explanations argue that violence in drug markets is a function of the global drug prohibition regime and the US led War on Drugs, which emerged in the early prohibitions of 1914 and consolidated with the 1961 UN Single Convention on Narcotic Drugs and the 1969 Declaration of the War on Drugs by US President Richard Nixon. This international prohibition framework has a tremendous impact on the content of antinarcotics policies and their emphasis on enforcement and militarized responses to drug trafficking (Youngers and Rosin 2005; LACDD 2012). As the US enforcement pressures associated with prohibition increase, the monetary stakes, competition, and consequently the violence associated with drug markets increase too (Gootenberg 2012). More specifically, explanations focused on foreign policy argue that the level of pressure associated with prohibition varies across countries and over time (Gonzalez 1989, Tickner 2003); as pressure mounts, so does conflict, as when relatively lenient US narcotics policy towards Mexico shifted in the mid 2000s triggering militarized responses towards traffickers (Serrano 2007). Even though explanations based on foreign policy and the prohibition framework pay crucial attention to the political implications of the drug market, in the end they are based on an economic rationale to link foreign policy and drug violence: as prohibition and enforcement drive the drug market underground, it becomes more violent, because illicit markets are more likely to be violent than licit ones. The arguments linking the global prohibition regime with an increase in black markets, human rights violations, and higher levels of insecurity and violence have become more public and widespread, as public opposition to the regime experienced a historic turn since 2009 (Open Society Foundations 2012; UNODC 2008; LACDD 2012).   12   These arguments have made increasingly clear that the prohibition regime does constitute the backdrop against which the overall criminalization and violence in drug markets flourishes. Yet, a focus on international policy does not suffice to explain variation in violence for three main reasons. First, even within the same country, antinarcotics policies can be implemented differently. The Mexican government’s decision to deploy military troops in its declaration of war against trafficking organizations in 2006 had differing impacts across cities: a larger fragmenting effect in the security apparatus which led to more violence in Ciudad Juárez, than in Tijuana where the security apparatus was more cohesive and violence less extreme. Second, once drug trafficking becomes violent and impacts citizen security, it obliges local governments to intervene; thus, local actors become crucial in determining the success of antinarcotics policies and enforcement. Third, the strategies of international actors vary depending on the structure of local enforcement. For example, according to a Colombian former Minister of Defense, in the 1990s the creation of an elite police group with joint military-police commands facilitated coordination and the channeling of US aid to Colombia, which up to that point had been fragmented across different enforcement agencies.11 As a result of this variation in the implementation of the global prohibition regime, a political economy approach to drug violence proves to be better equipped to understand the cross national and subnational variation that, within the same prohibition regime, leads to differing forms of violence.                                                                                                                 11 Author’s interview with Rafael Pardo, Bogotá, January 13 2011.   13   1.2.3. Explanations focused on processes of economic and social change Research on urban violence links broader processes like globalization, rapid urbanization, and economic transformation to criminal violence. This research has been especially useful to understand the upsurge of urban violence in Latin America in the 1990s (Kruijt and Koonings 1999, Moser 2006, Moser and Rodgers 2005, Davis 2010) but overlooks how illegal markets can be relatively non-violent and variation exists even among cities undergoing similar structural processes. Furthermore, a focus on broad structural processes is limited to explain rapid changes and fluctuations12 for example the increase of over 700% in homicide rates that Ciudad Juárez experienced between 2007 and 2008. By contrast, a political economy approach can capture the volatility of violence. Even if processes of transformation in states and markets can be slow, there can also be sharp variation in relatively short periods of time, and drastic changes of direction: the increase of violence in Ciudad Juárez occurred as the state security apparatus became more fragmented due to the massive and sudden deployment of the military in antinarcotics operations, and due to heightened competition in the illegal market. 1.3. A political economy approach to drug violence A political economy approach to violence goes beyond economic based explanations that link specific aspects of the drug trade or the size of profits in illegal markets to variation in violence, or that simply assume violence as a result of the                                                                                                                 12 To be precise, scholars of urban violence do recognize the complex and changing character of violence and relate it to rapid socio-economic transformations (Moser and Rodgers 2005). Yet while focusing on the highly violent situations they tend to overlook the socio-economic changes that do not produce criminal violence, or the contexts in which the criminality produced by these changes is not openly violent.   14   illegality of the market. A political economy approach does consider the structure of the market to be a crucial determinant of violence but analyzes it in its interaction with the state.13 To understand drug violence as the result of power relations between the state and illegal actors I create an analytical framework that bridges literature on violence in civil wars and literature on drug trafficking and illicit markets. Recent literature on civil wars that focuses on the dynamics of violence rather than on the causes of war, finds that variation in levels, types, and practices of violence results from changes in local monopolies of violence (Arjona 2010; Kalyvas 1999, 2006) and from changes in the balance of power between the state (incumbents) and insurgencies that challenge state sovereignty. Research on civil wars thus highlights that in the context of civil war, when local monopolies of violence change, technologies of violence can vary. I build upon this crucial insight and expand it in two main ways: First, even though some analyses of civil war explicitly exclude criminal actors as an object of study because criminals arguably leave sovereignty intact, I argue that the violent behavior of criminal organizations can be illuminated by the work done on civil wars because criminal organizations can in fact challenge state’s sovereignty even if they do not explicitly aim at controlling the state. Second, I argue that the literature on civil wars has overlooked how the state security apparatus and the power relationships within the state can affect the violent behavior of non-state actors, either by allowing the protection                                                                                                                 13 I follow here the analytical traditions of looking at the state as an autonomous entity that can have an impact on political and social processes through its policies and its patterned relationships with social groups (Skocpol 1985, p.3) and through the creation of political opportunity structures (Tarrow 1994). I expand it by looking at the impact of the state security apparatus on violence and it its interaction not with legal but with illegal actors. As recent studies in anthropology of the state have pointed out, one limitation of the state-society framework is that it conceptualizes two completely separate arenas when in fact state and society overlap in many ways and are not unified actors (Hansen and Stepputat 2001). This is especially important in the case of state-illegal actor relations because corruption may blur completely the limits between the state and society. I acknowledge this evanescent boundary but assert the validity of analyzing state and illegal actors as two separate fields.   15   of these actors, or by facilitating the attack on and effective policing of them.14 This omission stems from the fact that non-state violent actors are seen as naturally competing with the state, yet the state can sometimes tolerate, promote, or bargain with non-state violent actors, and these patterns of collaboration affect dynamics of violence (Staniland 2012). The possibility of collaboration is even more important for criminal organizations that always require collaboration within states in order to operate (Arias 2006, Lee 1999). Thus, my study expands research on civil wars by acknowledging the importance of power distribution within the state and by getting beyond the conceptualization of the state as a unified actor. The literature on illicit markets provides the keys to disaggregate the state and bring law enforcement to the front of understanding violence because it analyzes how enforcement actions aimed at dismantling illegal organizations can shape the dynamics of illegal markets (Andreas 2011, Friman 2004) and also generate violence because enforcement may dislodge organizations by killing or capturing leaders, thus creating internal power struggles (Caulkins et al 2006; Friman 2004; Goldstein et al, 1989; Rios 2012). These studies create an analytical link between law enforcement and violence, but they usually focus on individual enforcement actions rather than on the context in which these actions take place, thus ignoring that not all enforcement actions lead to the same level of violence. I build upon and expand this literature by examining different forms of                                                                                                                 14 Some analysis of ethnic violence and social mobilization provide insights about the importance of unpacking power within the state in order to understand violent outcomes. For example the collapse or fragmentation of central authority has been linked to increases of violence in the case of the Soviet Union (Beissinger 1998; Brubaker 1996), and political competition and democratic transition have also been linked to increases in violence (Brancati and Snyder 2011, Brass 1999, Snyder 2000). More recently elections have also been linked to variation in violence (Dunning 2011) as they reveal political-ideological preferences that can lead to targeting (Balcells 2001, Steele 2011), contribute to political polarization or help advancing the control of armed groups .Yet in these cases the mechanisms connecting state and violence are limited to the opportunities of groups to access state resources, to mobilize their identities or interests through political parties, but speak less to the actions the state can take as an autonomous actor.   16   violence, and considering how different state security apparatuses affect enforcement actions and determine criminals’ incentives to use violence. 1.3.1. The logic of the argument In illegal markets, criminals may use violence as a way to solve disputes given the absence of legal forms of mediation. Violence can also signal toughness, and thus the more violent an organization is, the less likely it is that competitors will try to overpower it or that members of the organization will cheat or defraud. The more visible violence is, the more likely that the toughness and power of the organization will be communicated and signaled to many people, which is advantageous for criminals. Yet, violence also has drawbacks such as scaring away “non-violent” partners, but especially, attracting police attention (Gambetta 2009, p. 34-35). Since criminals always need some collaboration within states, but states at the same time have the power to prosecute criminals, the structure of the security apparatus is crucial to understand violence precisely because it determines the ability of the state to credibly enforce the law or, alternatively, to protect criminal actors by non-enforcement of the law or enforcing it selectively against rivals. If criminals believe that the state can effectively attack them, they may prefer to hide violence, to minimize the risk of being targeted and then effectively prosecuted by police action.15 Thus using less visible violence can be a tactical adaptation to avoid state attention. Criminals may also prefer to hide violence if they are receiving state sponsored protection because such protection can be dependent on their “peaceful” behavior                                                                                                                 15 Of course criminals can also use violence to scare away enforcers or to retaliate when they have been attacked, as occurred in Colombia when the Medellín DTO declared a war against the government’s decision to enforce extradition laws for drug traffickers. Yet, if criminals fear state action enough, they may refrain from using visible violence against the state.   17   (Snyder and Duran-Martinez 2009a). True, state sponsored protection can also guarantee impunity when criminals commit violent acts, but even a weak captured state may be forced to act in the face of highly visible violence. Therefore, if criminals believe that the state can protect them effectively, they lose incentives to use visible violence that can undermine protection by obliging enforcers to respond to the potential public outrage generated by violence. By contrast, if criminals do not have credible protection, or do not fear state action, they lose incentives to avoid media and enforcement attention and gain incentives to signal their power and pressure the state and their rivals through visible violence. In the absence of predictable protection, refraining from using visible violence does not guarantee that criminals will be protected against enforcement. Thus criminals may decide to use visible violence, or can simply abandon the extra-steps necessary to hide it. In sum, by default criminals prefer visible violence as it can signal and create a reputation of power, and the incentive to use it increases as they face more competitors; yet if they believe that using visible violence will undermine the relationship with the state, or will likely lead them to jail or kill them, they will have less incentives to engage in public displays of violence. Given these incentives and mechanisms of state action, I argue that the frequency of violence depends on the structure of the illegal market, and the visibility of violence depends on the structure of the state security apparatus. The organization of the drug market16 refers to the number of organizations competing over the control of drug                                                                                                                 16 For analytical purposes the emphasis of this dissertation is on drug trafficking, yet it is crucial to note that illegal markets are complex, and one organization can engage in several illegal activities at the same time. For example in Medellín organizations today fight not only for drugs, but also for controlling extortion and   18   markets in a given city, and thus it can be monopolistic or competitive. A monopolistic market is likely to generate less violence than a competitive market where there are disputes for turf and territory among several illegal firms, because illegal firms drive out competitors not by lowering prices but by using force against them (Friman 2009; Skaperdas 2001, p.187; Konrad and Skaperdas 2005).17 The state security apparatus refers to the set of elected authorities and enforcement agencies (police, army, and intelligence agencies) that determine security and antinarcotics policies in any given city, and it can be cohesive or fragmented. The capacity to enforce the law depends critically on the ability to coordinate enforcement actions and thus may increase as power within the security apparatus is more cohesive. Likewise, the ability to provide reliable and stable protection to criminals depends on the ability of actors within the security apparatus to guarantee that no other state authority will enforce the law against criminals, and this is also likely to happen when power within the security apparatus is cohesive. Conversely, as power within the security apparatus fragments, coordination and successful enforcement become more difficult, and protection becomes less reliable. A fragmented security apparatus can make enforcement less effective18 because it makes coordination more complicated, and thus enforcement actions can either leave DTOs untouched or disrupt them without eliminating their fire capacity, making them prone to initiate violent struggles for leadership within and among them. In a fragmented security apparatus even if criminals                                                                                                                 gambling. This was the case even during Pablo Escobar’s period in Medellín. See El Espectador. 2008.“Cerrojo a la Oficina de Envigado.” El Espectador, July 27. 17 However for economists like Skaperdas competition constitutes an equilibrium in illegal markets and thus, he argues, violence is more often present than not. 18 There might be cases when fragmented enforcement apparatuses can be effective as is the case with U.S. law enforcement. This might occur when fragmentation generates overlapping jurisdictions that increase checks and balances and different enforcement actors have the ability to coordinate their actions.   19   have more opportunities to bribe public officials, they also face more possibilities of prosecution and greater transaction costs for corrupt deals. Paradoxically, a cohesive state that has more capacity to enforce the law is also a state that can successfully protect criminals. The results are the same (less visible violence) but the mechanisms are different: a cohesive state complicit with criminals deters the use of visible violence because it allows the formation of state sponsored protection rackets and visible violence may force the state to act thus undermining the benefits of state protection; a cohesive state that enforces the law effectively creates incentives to hide violence by creating a credible equilibrium: criminals know the state can effectively attack them and thus they try to avoid state attention as much as possible: using less visible violence is a tactical adaptation to survive as the state becomes more efficient in enforcing the law. Thus, the visibility of violence depends on the relation of criminals with the state but it does not mean that visible violence is only used against the state: once the incentive to hide violence disappears, criminals may decide, depending on their interests, to use visible violence against their rivals, the state, or in some cases, against civilians. In sum, I argue that when the illegal market is monopolistic violence is likely to be less frequent because criminals do not face turf disputes and thus may recur to violence only as a way to discipline their members or to punish transactions gone wrong. By contrast, when the illegal market is competitive, violence is likely to be more frequent as drug trafficking organizations try to eliminate their competitors. But the structure of the market alone does not explain the changing dynamics of drug violence, thus I further argue that a cohesive security apparatus is likely to reduce the visibility of violence   20   because it can increase the stability of the protection criminals obtain if the state protects criminals. A cohesive state that does not protect criminals can also reduce the incentives for criminals to use violence. Finally, a fragmented security apparatus is likely to increase the visibility of violence because it reduces the predictability of protection and the effectiveness of enforcement, and thus hiding violence may not necessarily decrease the risk of prosecution for criminals. This in turn, increases their incentives to employ visible violence. It is crucial to note that frequency and visibility are interconnected because criminals do not decide on the forms of violence they will use thinking about frequency and visibility as separate dimensions, but simultaneously. In other words, if the market is competitive and market disputes emerge, criminals decide to deploy violence, but they simultaneously decide, depending on the conditions of the state security apparatus, whether that violence could be perpetrated visibly, or should rather be kept hidden. Yet, the disaggregation of the two dimensions and of the factors that affect the decision on each dimension provides a useful way to capture patterns of variation that could not be noticed by using a concept that conflates visibility and frequency. Jointly, the state security apparatus and the structure of the drug market determine distinct patterns of drug violence. It is the interaction between the two that explains violence, because a violent outcome cannot be simply explained by looking at the market or at the state in isolation. A fragmented security apparatus creates conditions for visible violence, but such violence may not emerge if there is no market dispute. Conversely, a   21   market dispute can generate violence, but if the security apparatus is cohesive, such violence is likely to be less visible.19 1.3.2. “Insourcing” and “Outsourcing” violence To fully explain variation in violence across cases, I analyze a third variable that explains extreme spikes of violence, an often overlooked aspect of organization in criminal markets which is the type of armed coercion criminals employ, and more specifically whether violence is insourced or outsourced. Criminals insource violence when they deploy personnel from within the organization to carry out violence, and they outsource when they systematically recur to youth gangs as outside providers of force.20 When criminals outsource their use of violence to youth gangs, they can increase their firepower but at the same time lose the ability or willingness to control their members, because the transaction costs of dealing with many small organizations are high. Thus, better-armed and more violent gang members can spread violence beyond criminal disputes. By contrast, if criminals insource their provision of violence within the organization, criminal soldiers can be easier to discipline and less likely to use violence beyond criminal disputes. Outsourcing can also affect the visibility of violence because hiding violence may not be a strategic priority for gangs as it is for the direct protagonists of the illegal business; in fact showing power can be crucial for the gang’s symbolic                                                                                                                 19 One may argue that in a competitive market, a cohesive state can protect one organization and persecute its rivals, and the rivals in turn may have incentives to use visible violence against the state. However, in the logic posed by the argument a cohesive state can be very effective in enforcing the law, and thus rivals may refrain from using visible violence because they know the state will enforce the law effectively against them. The situation changes if the state security apparatus is fragmented, because it can be less effective and different sectors of the security apparatus can protect different organizations. 20 Criminals can also outsource to other actors such as corrupt enforcement officials or mercenaries, however access to this type of force is prohibitive and usually not massive. Access to gangs is less prohibitive and more massive, therefore has different implications.   22   appropriation of spaces and turf (Venkatesh 1997). Thus, although the primary impact of outsourcing is on the frequency of violence rather than on the purposed deployment of violence as a communicative tool, outsourcing creates spikes of violence because it has a multiplying effect on both visibility and frequency. When traffickers outsource their armed coercion, violence is likely to reproduce to extreme levels because youth gangs acquire abilities that they did not previously have, such as the use of firearms (Salazar 1990, Ceballos 2003). Marginalized youth, with little economic, educational, or social options acquire abilities that end up being used beyond the framework of organized crime and trafficking disputes, thus spreading violence rapidly. Although this process could be characterized as one of professionalization in violence, the issue is precisely that most youth are not professional killers and thus do not necessarily follow the “rules” of professional killers such as mercenaries, enforcement agents at the service of traffickers, or privatized security firms. Although some gang members can engage in a criminal career that takes them to the upper echelons of criminality, most of them only have momentary interactions with the highest criminal ranks. The importance of outsourcing is illustrated by the contrast between Cali and Medellín in the 1980s: while the Cali DTO preferred to insource violence, the Medellín DTO actively hired and deployed youth gangs as paid assassins, and as a result the frequency of violence, as reflected in homicide rates, was far higher in the latter. In the words of a military official closely involved in prosecuting the Cali DTO “the Cali people, because of their style and because they wanted to be different, their sicarios were less preposterous, more sophisticated; they, unlike the sicarios of Medellín, were not paid   23   by their crimes, they were part of the organization and in consequence the sicarios of the Rodriguez Orejuela [leaders of the Cali DTO] could not violate the organization’s parameters, the Rodriguez Orejuela intercepted their sicarios’ communications to keep them controlled.21” The outsourcing of violence resembles the process by which mafia groups were linked to common and juvenile crime as the demand for criminal labor expanded (Arlacchi 1976, p. 221) and met the supply of labor, usually fueled by processes of socio- economic change and crisis that left large masses of young people unemployed. Likewise, in drug trafficking outsourcing occurs when supply of force meets demand. The supply, that is, the mere existence of gangs or large young populations, does not imply that they will be used as criminal labor; outsourcing only occurs if supply meets the demand, created by illegal market disputes. Organizational preferences can also play a role in the decision of criminals to outsource violence: outsourcing can be advantageous while expanding the armed muscle of an organization relatively cheaply, but it also creates risks of exposure. Thus organizations that aim to create a reputation as businessmen or that operate within a cohesive state may try to avoid outsourcing as in the case of the Cali DTO or the Arellano Felix Organization in Tijuana which preferred to recruit youth from the upper classes that came to be known as narcojuniors (Blancornelas 2002). By contrast, organizations can choose to outsource when they enter new territories where they do not have traditional presence as in the case of the Sinaloa DTO in Ciudad Juárez. Before                                                                                                                 21 Author’s interview with military intelligence official, Bogotá January 31 2011.   24   2008 the Sinaloa DTO did not have territorial presence in the city and thus its expansion strategy required both to mobilize soldiers,22 and to find new recruits within the city. Once traffickers outsource violence, they create a situation that induces gang proliferation and perpetuates a cycle of violence reproduction as the expectations of the money that could be gained working for traffickers create incentives for gangs to proliferate and compete for the more profitable jobs (Salazar 1990), even if the expectation of social advancement is rarely fulfilled. In rare occasions criminal organizations can also have a pacifying effect on the behavior of youth gangs, punishing the use of violence and regulating the location and timing of activities like drug distribution, especially when they operate within a cohesive state and have incentives to hide violence. The key is that if criminals actively engage youth gangs in violence, then violence is likely to spike. Outsourcing can be dependent on, but is separated from, the structure of the illegal market. Thus, outsourcing may not be necessary if there is no competition, but if competition emerges, then criminals may decide to insource or outsource. Likewise, to create a market monopoly, criminals may need to insource violence and discipline gangs, but monopolies may also emerge, even if gangs exist, when they are not related to criminals and not violent. To understand the interaction between armed coercion, the structure of the market, and the state security apparatus, Figure 1.2 illustrates the dynamics of the argument. The front panel of the cube depicts the interaction between the state security apparatus and the competition in the illegal market, and the right side of the cube represents the type of armed coercion criminals employ. Outsourcing is more likely to                                                                                                                 22 In fact, on October 4 2011, the Mexican Federal Police arrested Noel Salgueiro who was commissioned by the Sinaloa DTO to invade Juárez with 500 men, entering the city form the South. He later recruited members of large local gangs known as Los Mexicles and Los Artistas Asesinos.   25   emerge when there is competition in the illegal market because criminals may not need to garner additional armed force if they control the market. Thus, the upper right quarter of the cube depicts the possibility that high frequency violence may lead to spikes if there is outsourcing. Criminals that receive predictable protection from cohesive states may prefer to avoid outsourcing as it entails risks of exposure, and thus outsourcing is less likely in the left side of the cube. In this attribute space, the upper right quarter of the backside of the cube depicts the worst possible outcome of violence, when fragmentation in the state security apparatus, competition in the illegal market, and outsourcing lead to high frequency, high visibility, and spikes of violence. Figure 1.2. A dynamic depiction of drug violence 1.4. Conceptual clarifications In this section I address issues that may emerge while defining how an illegal market is organized, and while determining the structure of the state security apparatus.   26   1.4.1.The structure of the market The assumption that drug trafficking activities are highly organized, that criminal organizations are extremely hierarchical, compact, and powerful, has tended to dominate discussions on the drug trade. In response to this assumption scholars of illegal activities have shown that in fact the supply of illegal commodities like drugs tends to be highly disorganized, (Paoli 2002, Reuter 1983, Naylor 1998) that organizations rather than hierarchical tend to operate more like networks, (Kenney 2007) and that trafficking organizations do not have the ability to set prices, and therefore it is inappropriate to label them as cartels23 (Astorga 2005, Dorn, Murji and South 1992, x). These insights have clarified the understanding of criminal markets highlighting that most times depictions of extremely powerful organizations are the result of media or enforcement driven agendas. Yet, there is still wide variation in the way illegal markets are organized, and these differences are consequential for the behavior of trafficking organizations. Drug trafficking organizations do manage to create monopolies24 but they can also operate in highly competitive environments, and there is variation even when they operate more like networks (Dorn, Oette and White 1998; Malm and Bichler 2011). As the number of organizations competing for control of a given illegal market increase, we                                                                                                                 23 The use of the term cartel is highly controversial in the literature on drug trafficking. The term emerged mainly in journalistic and policy circles and has been criticized because it portrays a level of organization and coordination that does not often exist in criminal organizations and networks. It is also criticized because the term cartel implies collusion to set prices, something that criminal organizations do not achieve often. Yet, sometimes organizations do have abilities to set prices and coordinate a broad range of actions, even if they operate in a decentralized manner. I mainly use the terms Drug Trafficking Organizations (DTO), and criminal organizations to refer to organizations devoted to drug trafficking. 24 When organizations manage to establish monopolies they can in fact behave (or try to behave) as cartels setting prices. For example in the 1980s the Cali and Medellín DTOs met to discuss the price of cocaine and in 2004 several Mexican organizations met in order to try to set prices. Furthermore as Molzahn, Rios and Shirk (2012) show, the term cartel has a broader connotation in economics that includes formal and informal arrangements that affect production, competition, and prices, and trafficking organizations do often engage in these types of agreements.   27   can expect violence to be more frequent. The evidence compiled for this project shows that when one organization dominates the market in a city, violence is likely to decline, as occurred in Medellín between 2003 and 2007, or in Ciudad Juárez in the 1990s. The question is, if trafficking organizations operate with structures that vary in their degree of cohesion and organization (Bruinsma and Bermasco 2004) how can we determine what an organization is and when it has created a monopoly? Furthermore, considering that enforcers are usually interested in depicting extremely powerful organizations, while traffickers may sometimes prefer to deny their association with an organization, how can we accurately measure the power of a criminal group? To address this complex conceptual issue I separate competition from the structure of the organization (its cohesion or consolidation) and assess whether one or more groups are claiming control over the territory of a city, regardless of their structure. This conceptual decision is also based on the empirical observation of cases. One may argue that it is the internal structure of criminal organizations, rather than competition, what determines their use of violence. It can be argued that small and less centralized organizations have less capacity to engage in violence than large, centralized ones (Camacho and Lopez 2001, Eilstrup Sangiovanni and Jones 2008). Yet, the case studies analyzed in subsequent chapters show that similar organizational types can be linked to different forms of violence; smaller organizations can be linked with less visible violence in Cali and Medellín in the 2000s, yet paradoxically small splinter organizations have also been linked to more frequent and visible violence in Mexico (Guerrero 2011). This is because, as research on terrorism and suicide missions shows, high impact violence can be carried by a diverse range of organizational types including small organizations or   28   splinter cells within extremist groups (Bueno de Mesquita 2008, Gambetta 2005, Ganor 2008). Additionally, the centralization of an organization may not necessarily determine whether there is a hierarchy that internally monitors the use of violence, because organizations may operate like networks and compartmentalize information. and yet maintain clear hierarchies (Shapiro 2005). The assessment of the structure of the drug market faces another complication, which is that drug trafficking is by nature a transnational phenomenon, and even if criminal organizations are not always extremely powerful transnational groups, drug trafficking flows transcend local, national, and regional borders. Yet, manifestations of violence are highly localized, and in fact the same criminal organization can behave differently in different locations as illustrated by the changing behavior of national scale organizations across the Mexican territory. For example the Sinaloa DTO in Mexico was significantly more violent in Ciudad Juárez than in Tijuana. Similarly, Pablo Escobar and the Medellín DTO had extensive power inside and beyond Colombia, yet this power was contested in Medellín where the criminal market also included guerrillas, militias, and youth gangs. 1.4.2. The structure of the state security apparatus The state should be at the center of understanding violence by non-state actors because violence challenges the state’s monopoly of force; law enforcement is crucial to maintain the monopoly of force, thus it should be crucial in analyzing violence too. Yet, in comparative politics there has been a relative neglect in analyzing how the practices and structure of law enforcement impact the state’s monopoly of force (Tanner 2000,   29   Taylor 2010). Since the 1990s there has been an increase in knowledge about law enforcement but scattered among studies of democratization, citizen security, and police reform (Fruhling 2003; Fruhling, Tulchin and Golding 2003; Bayley and Dammert 2005). These studies usually focus on the impact that policing frameworks have on human rights and citizen security in nascent democracies, or on the endemic problems that police forces experience in countries with weak institutions, but as Davis (2006) notes, they rarely analyze the interaction between law enforcement and political dynamics, and broader state structures (Marenin 1985). While corruption in police forces is often associated with violence and spikes in criminal behavior, most studies lack an analytical link that explains the mechanisms connecting law enforcement and violence. The concept of the state security apparatus unpacks the components of the state’s monopoly of force, describes the mechanisms that connect it to the behavior of criminal organizations, and illustrates how under certain circumstances, corruption may deter criminal violent behavior. The state security apparatus refers to the set of elected authorities (executives) and law enforcement agencies25 that determine the definition and implementation of security and antinarcotics policies. The security apparatus can be cohesive or fragmented depending on three elements: intergovernmental relationships (vertical relations among levels of government); interagency relationships (horizontal relations across enforcement agencies); and time horizons of elected and enforcement officials. Fragmentation in the security apparatus increases when there are conflicts between levels of government (i.e., the President and the Mayor), conflicts across enforcement agencies (i.e., the military and                                                                                                                 25 Although the concept of law enforcement can broadly encompass all agencies that promote adherence to the law, it is more often referred to the agencies that engage in patrolling, surveillance, investigation and apprehension (Hess 2009).   30   the police), and when time horizons of officials are short due to processes such as rotation and purges in police forces, or higher levels of electoral competition. In principle, changes in the time horizons of public officials affect the predictability of power relations rather than the fragmentation of power itself. However uncertainty, or the possibility that power can be lost, implies that somebody can challenge power, and thus, that power is not concentrated in a single actor. This creates a direct link with the minimalist definition of democracy: democracy is characterized by the uncertainty of outcomes, but outcomes are uncertain because no single political force can determine them unilaterally (Przeworski 1991, p.12). Thus, when power within the security apparatus becomes more uncertain it is because it is becoming more contested and fragmented. The three dimensions of the security apparatus are interrelated but do not necessarily co-vary: a purge within a given police corporation affects time horizons by increasing uncertainty about who is a cop and about how long someone can remain within the purged agency, but it may not necessarily increase conflicts between the heads of different enforcement agencies or different levels of government. Thus, a change in any of the dimensions that constitute the security apparatus is likely to increase the visibility of violence even if the other dimensions remain stable; the more changes along the three dimensions of the security apparatus, the larger the effect on the visibility of violence. This was the case, for example, when the state security apparatus became extremely fragmented and visible violence spiked in Ciudad Juárez in 2008, as there were conflicts between enforcement agencies, between levels of government, and between political forces.   31   The structure of the state security apparatus thus depends on institutional arrangements that distribute government responsibilities such as the type of regime (unitary or federal), on bureaucratic dynamics that determine rotation and responsibilities of enforcement agencies, and on political and electoral dynamics and processes of democratization. In federal systems, like the Mexican, fragmentation of the security apparatus can increase, as local authorities such as Governors have more power and this increases the possibility of conflicts among different levels of government. Federal systems usually have a greater number of police forces than unitary systems; in Mexico there are more than 1600 police forces (Benitez Manaut 2006), thus making it more difficult to coordinate enforcement actions. Yet, some of the fragmenting effects of federal systems can be mitigated when political forces are hegemonic as in the case of the Institutional Revolutionary Party (PRI) in Mexico that successfully managed to control subnational political units and thus regulate relations with trafficking organizations (Astorga 2004, Astorga and Shirk 2010, Florez 2005, Serrano 2007). Likewise, the cohesive potential of a unitary system can be destabilized by decentralization processes that augment territorial power, as occurred in Colombia after 1986 when the popular election of Mayors was introduced (Faletti 2005). Bureaucratic dynamics affect fragmentation when responsibilities in antinarcotics policies are assigned to new agencies, or when massive rotation in enforcement agencies takes place. In Colombia in the 1980s, when the government started to confront drug violence in Medellín, it faced a set of corrupt and unprepared enforcement agencies, and the army became a crucial actor in urban antinarcotics operations. Yet the army was not   32   completely prepared to deal with these operations, and its deployment also created conflicts with police forces and discontent within the military itself (Palacios and Serrano 2010, p. 120, Ramirez 2011). Thus employing a new actor in enforcement operations led to a fragmentation of the security apparatus. Finally, political and electoral dynamics and democratization processes are also likely to produce fragmentation in the security apparatus by increasing the number of political forces that influence policymaking. In a context of open electoral competition there might be an increasing number of political actors prone to establish relations with criminal actors, but also an increasing number of actors committed to prosecute criminals or at least willing to expose the evidence of corruption for electoral purposes (Grzymala- Busse 2008, p. 658; Murillo and Martinez-Gallardo 2007, p. 124). Thus, democratization can empower regional actors and provide renewed opportunities for criminals to access the state as occurred in Colombia26 (Gutierrez 2007, Romero 2004) but also reduce the ability of the state to provide credible protection to criminals, as occurred in Mexico as the PRI hegemony stumbled when the party lost Presidential elections in 2000. It is important to note that higher levels of political competition are not automatically linked to higher fragmentation in the security apparatus; in some cases high competition can exist in cohesive security apparatuses if political actors have incentives to cooperate beyond party lines or if powerful political elites dominate even in the face of competition. Thus, in order to characterize the security apparatus in each city I combine formal indicators of electoral competition, political alignments in different levels of                                                                                                                 26 Colombia only experienced a brief dictatorship in 1957. Yet, for 16 years (1958-1974) political competition was limited to two political parties in the framework of a power sharing agreement known as Frente Nacional.   33   government, number of agencies working on enforcement responsibilities, but also informal indicators of collaboration or confrontation between different authorities. 1.4.3. State and market interaction There are of course, interactions between the state security apparatus and the structure of the drug market. Actions of the state security apparatus can shape the criminal market, for example when the capture or death of leaders fragments and creates “vacancies” within criminal organizations (Friman 2004), or when organizations adapt to enforcement pressures (Dorn, Burnji and South 1992, p.35). Heightened law enforcement can professionalize and force cooperation among criminals who would otherwise work individually, thus pushing criminal organizations to increase power by expanding their territorial control (Andreas 2011). Law enforcement can also force organizations to reduce their size in order to divert attention away from them. Thus enforcement actions could determine violence indirectly by determining the structure of the illegal market. Yet, not all enforcement actions have an immediate impact on the illegal market because they depend on the structure of the security apparatus. For example in 1989 Mexican enforcement authorities captured trafficker Miguel Angel Felix Gallardo who at the time was considered the main leader among Mexican traffickers. The capture did not generate an immediate violent response, precisely because, in the words of one enforcement official, Gallardo had strict control of his organization and the state was cohesive.27 By contrast, the capture of another prominent trafficker (Alfredo Beltrán Leyva) twenty                                                                                                                 27 Author’s interview with former drug enforcement official from the General Attorney’s Office. Mexico City, March 9 2011.   34   years later in April 2008, unleashed violent responses from DTOs, because the state was far more fragmented. The criminal market can also shape the security apparatus; the state can change as it responds to criminals’ behavior and more violence can eventually force changes in the security apparatus. Criminality can be seen as competition for state-making because criminal organizations may compete with the state in creating power (Felbab-Brown 2011), challenge the national state’s monopoly of force creating fragmented sovereignties (Davis 2010), or because, as in Tilly’s approach (1985), states behave like organized crime while relying on self interested coercion and extraction when building up state power. But criminality can also drive state-making because the state security apparatus grows, expands, and becomes more sophisticated in its effort to confront criminality. In Colombia in the 1980s when drug trafficking and the violence related to it became a threat to the state, the security apparatus was unprepared for antinarcotics operations and plagued by problems of lack of coordination and corruption especially in the police (Camacho 1993, Pardo 1996, Casas Dupuy 2005). The need to confront criminal organizations pushed and motivated some of the most crucial changes and expansions in the Colombian security apparatus over the past three decades, such as the initial experiments with specialized elite units with army-police unified commands, the creation of a specialized judicial police in the Attorney’s Office, or the implementation of sophisticated intelligence capacities in the police. The complex interactions between the structure of the criminal market and the structure of the state security apparatus illustrate why rather than establishing a causal relationship between these two variables, that then can explain violence, it is necessary to   35   understand violence as the relational product of the interaction between these two variables at any particular time. The interactions can produce four main configurations of violence: Low frequency-low visibility (LF-LV); high frequency- low visibility (HF-LV); low frequency-high visibility (LF-HV) and high frequency-high visibility (HF-HV) that I briefly characterize below, and will be fully described throughout the chapters of this dissertation. 1.4.4. Patterns of drug violence LOW VISIBILITY – LOW FREQUENCY VIOLENCE (LF-LV) This configuration occurs when the state is cohesive and the market monopolistic. The monopoly in the market reduces conflicts for control of territory and the cohesion in the state motivates criminals to maintain violence hidden either because they fear losing state protection if protection is predictable (as in Ciudad Juárez between 1984 and 1994) or because they fear state prosecution (as in Medellín between 2003 and 2007). Violence can emerge to discipline members, to eliminate potential competitors, or to regulate “public threats” such as low-level criminality, drug users, or social “undesirables” but the evidence of this violence is hidden. In this scenario, if youth gangs exist, trafficking organizations are more likely to discipline them than to outsource armed coercion to them. HIGH FREQUENCY-LOW VISIBILITY This configuration occurs when the drug market is competitive and thus generates violence aimed at gaining market control, and the state security apparatus is cohesive, thus reducing the incentive to use visible violence. This configuration of violence is the   36   most common in drug markets, but tends to be overlooked by the media and enforcement agencies as illustrated in the cases of Cali (1994-2010) and Culiacán (1984-2008) where the state was cohesive and protected criminals, and Medellín (2008-2010) where the state was cohesive but less protective of criminals. Given the incentives of criminals to reduce the visibility of violence, outsourcing is not likely to occur in this scenario. LOW FREQUENCY –HIGH VISIBILITY This configuration is the result of a monopolistic market and an increase in the degree of fragmentation within a state that used to be cohesive and protected criminals. This fragmentation emerges when new enforcement agencies are deployed in a territory or when rotation and purges reduce the time horizons of enforcement agents and criminals use visible violence to retaliate the state and force it to refrain from enforcing the law. This configuration of violence is not common and tends to be very short lived because it constitutes a highly unstable equilibrium and can easily evolve into HF-HV if the state or the market further fragment. Tijuana in 1996-1997 illustrates such configuration. HIGH FREQUENCY-HIGH VISIBILITY (HF-HV) The polar opposite of the LF-LV configuration is the extremely high frequency and high visibility (HF-HV) created by a competitive market and a fragmented state along all three dimensions (conflicts among levels of government, conflicts between enforcement agencies, and short time horizons of public officials). In this configuration, violence can take on very extreme dimensions if criminals outsource violence to youth gangs as occurred in Ciudad Juárez between 2008 and 2010 and in Medellín between 1984 and 1993.   37   Figure 1.3. Trajectories of violence in each city 1984-2010 1984 2010 Cali LF-LV HF-LV Medellín HF-HV HF-LV LF-LV HF-LV Ciudad LF-LV HF-LV LF- HF-HV Juárez HV Culiacán HF-LV HF-HV Tijuana LF-LV HF-LV LF- HF-LV HF-HV HV LF: Low Frequency, HF: High Frequency, LV: Low Visibility, HV: High Visibility 1.5. Research Design and Methods This dissertation is based on subnational comparative analysis, which is crucial to analyze the unevenness of outcomes that occur within the same country (Snyder 2001). A particular configuration of violence in each city in a specific time period constitutes a case, therefore this project constitutes a medium N design (18 city-violence periods) that covers a wide range of variation in both the dependent variables (the frequency and visibility of violence) and independent variables (the state security apparatus, the market structure, and the type of armed coercion employed by criminals). The cross-national nature of the research provides greater scope and generalizability to the findings, thus addressing a potential limitation of subnational analyses, which are usually focused on just one country (Lankina 2012, Moncada and Snyder 2012). The limited development of theories on drug violence made this project a theory building exercise where cases represented values of the dependent variable (Lieberman 2007). Colombia and Mexico were chosen because of their long-term histories with drug trafficking and because despite their difference across important dimensions, most notably regarding the structure of the state (unitary vs federal) and the existence of armed conflict, they have experienced similar moments of violence. Furthermore, there is   38   increasing interest in policy and academic circles in analyzing if current violence in Mexico is similar to Colombia in the 1980s and if policy lessons for Mexico can be drawn from Colombia. The five cities selected for this study, Cali, Ciudad Juárez, Culiacán, Medellín and Tijuana, have been home to major drug trafficking organizations (DTOs) throughout the entire study period (1984-2010) and highlight the variation in violence that can occur over time and across the same territory. By focusing on cities located in the same country and with similar size, yet contrasting patterns of violence, I ensured unit homogeneity to explore systematically the causes of variation in drug violence. At the same time, I assessed variation across other potential explanatory variables, such as international influence while comparing for example two cities that border the United States and thus have a greater likelihood of international influence, with other non-border cities. To assess the impact of the independent variables and disentangle common causes along these five cities, I used systematic paired comparisons in three stages. First, I paired cities in the same country that experienced contrasting patterns of violence, Cali and Medellín in Colombia, Ciudad Juárez and Tijuana, and Culiacán and Ciudad Juárez, in Mexico. This juxtaposition held constant national similarities and thus constituted a “most similar systems design” (Przeworski and Teune 1970; Ragin 1987) that highlighted factors that explain contrasting trajectories across cities. In the second stage I paired similar configurations of violence in different countries: Cali and Culiacán on one hand, Medellín and Ciudad Juárez on the other, in order to illuminate cross-national similarities that explain these outcomes. Finally, I paired different configurations of violence across time within each city in an even stronger, longitudinal, most similar research design. The   39   systematic paired comparisons allowed me to control for the particularities that exist among these five diverse cities: for example by comparing Medellín over time I was able control for socio-economic and cultural factors that remained relatively constant; then I tested the validity of the argument by applying it to explain why Ciudad Juárez and Tijuana, despite being two border cities with similar historic and socio-economic issues, have experienced different trajectories of violence. The multiple, systematic comparisons allowed me to reduce the rich combination of factors that produce different configurations of violence in each city at a specific point in time, to the simplest combination of variables present in each case. Systematic comparisons also allowed me to take into account the dependence of time periods within the same city (what happens in Medellin at time t, is not independent of what happens in t+1). To operationalize and code dependent, independent, and alternative variables I collected a wide range of primary and secondary data including violence statistics and electoral statistics, bibliographic material, government documents, newspapers reports, some judicial records, and a large number of interviews. I collected the data during preliminary research I conducted in Colombia (summer 2008) and Mexico (summer 2009), and then over extended fieldwork conducted between September 2010 and November 2011 that included on site research in each of the five cities. During these research trips I conducted over 175 semi-structured interviews with police officers, intelligence officials, politicians, journalists, human rights workers, representatives of international organizations, and NGOs, using a snowball sampling technique. To characterize violent dynamics, I created a dataset that includes over 6,000 violent events coded from local newspapers in selected years for each city. Each event was coded along   40   60 variables that describe aspects such as the location of violent attacks, the number of victims, and the methods used to carry out the attack. To operationalize the dependent variable (configurations of drug violence) I triangulated many sources of information. To characterize the frequency I identified trends in homicide statistics in each city and I triangulated these trends with local and national newspapers and secondary sources (such as journalist’s accounts, biographies and local research). I characterized spikes in violence as moments when homicide rates in a given city more than tripled the national average. I used interviews to identify periods of high and low visibility, and then cross-validated this information with human rights reports and with the dataset on violence I constructed. Even though the dataset does not provide an entire time series amenable for statistical analysis, it provides enough data points to characterize and contrast high and low visibility within and across cities, through information about targets, methods, and location of violent attacks. Human rights reports were crucial to code cases of low visibility, which do not appear in headlines or officials statistics, such as disappearances. One advantage of the long-term character of the analysis is that over time evidence of less visible methods of violence that prevailed at a specific point, but where not reported in the news of the moment, can become available. To operationalize the state security apparatus I analyzed standard indicators of electoral competition (volatility, effective number of parties, vote gaps, coincidence of party affiliations) for executive elections at the national, state (department), and municipal level, as well as party affiliations of elected officials to identify moments of higher competition that shorten time horizons of elected officials, as well as levels of   41   collaboration among government levels. I used my interviews and secondary materials to identify informal dynamics of conflict-collaboration. I used government documents, press reports, and interviews, to characterize the levels of conflict or collaboration between enforcement agents. Finally, to characterize the structure of the criminal market and the reliance on insourcing or outsourcing I triangulated interviews, and primary and secondary materials. The main variables were not coded as strictly dichotomous, to capture complex variation. For example in each city the security apparatus was not only characterized as cohesive or fragmented, but as more or less cohesive. Thus if in a given city at a given time the security apparatus was characterized by conflicts among agencies, conflicts among levels of government, and reductions in time horizons due to an increase in electoral competition, the city was characterized as having a highly fragmented security apparatus, and this score decreased if one of these components was not present. A similar method was used to code the structure of the market, and the frequency and visibility of violence. Insourcing and outsourcing were coded as a dichotomous variable. Table 1.1 summarizes each variable, its definition, and the range of scores that it could obtain.   42   Table 1.1. Definition and components of each variable Definition Scores Frequency of Rate at which violent events occur High to Low violence Dependent Variable Visibility of Extent to which criminals publicly High to Low violence display the evidence of their attacks or claim responsibility of violence Market structure Number of organizations competing Monopolistic for market control to competitive Criminal Armed How criminal organizations carry out Insourced or Independent Variables Coercion violence: from within the organization Outsourced or by contracting out youth gangs State Security Set of elected authorities and Cohesive to Apparatus enforcement agencies that determine fragmented security and antinarcotics policies. Determined by 1) intergovernmental relations 2) interagency relations 3) time horizons of public officials All the coded variables and information collected along with a through literature review and analysis of secondary sources, constituted the basis of detailed analytical narratives and descriptions of sequences of events. Analytic narratives and process tracing were the core tools used to disentangle causal paths and mechanisms. The most important types of events I traced were: - Changes in violent techniques; this implied not only identifying whether an attack that exposed the evidence of violence emerged for the first time, but whether it became more routine; - Key law enforcement operations; the motivations for state agents to act against traffickers; how operations were conducted and what authorities participated in   43   them (for example whether they implied mobilization of new enforcement actors such a the military); the reactions and changes caused in criminal organizations. - Changes in the time horizons of public officials, tracked both by looking at major shifts in electoral competition, and massive rotations, or purges in enforcement. - To assess the alternative explanations within the narratives I carefully looked at how national policy changes, changes in international antinarcotics policies, and major market transforming events (such as greater US border control after 9/11) affected each city. 1.5.1.Conducting research on a sensitive topic Drug trafficking and violence are highly sensitive topics which pose two main challenges for research: first and foremost the safety and security of the researcher and her informants, and second the reliability of the information collected. Here I highlight some of the most important measures I undertook in order to mitigate these problems. To limit risks for my informants I have taken all precautions necessary to guarantee anonymity except in cases of high-level officials (unless they explicitly asked me to make their opinions anonymous). In the text I identify each interview only with its location and date, and in order to provide some guidance for the reader, with the broadest describer that can characterize the respondent without risking his/her identify (i.e police officer, journalist). At the personal level I took several precautions to ensure my safety, and this forced me to change plans constantly and be flexible with my research schedule. In the most violent places (and even in the apparently calm places) I gathered extensive   44   information about the situation on the ground prior to my visits, reading newspapers, talking to experts, and especially with journalists, scholars, and practitioners that had visited my research areas shortly before my trips. Having established key personal contacts beforehand was crucial to have a better sense of the security on the ground. Contrary to my initial plan, I did not spend long periods in each city. Instead, I conducted several short trips, because I realized that this reduced the possibilities of being singled out for asking too many questions. In Mexico this was particularly important because I was traveling between cities where rival organizations operate. Within each city I stayed with local families I knew beforehand and this became a great source of protection. Even though I used snowball sampling technique for the interviews, I did not accept every interview that was offered to me; on one occasion I could have had the opportunity to interview key criminal actors but I refused to do it because I did not trust the person who offered me the opportunity; it turns out that the person who made the offer was killed a week after I met him. I only conducted interviews recommended by people I trusted, and while of course this generated a bias in networks of people I interviewed, in the end, there was no case in which more than 5 interviews originated from the same contact and only in one case I had a large group of interviewees that could be somehow traced to the same network, with 20 people involved. I was always clear about the objectives of my research and never misrepresented my work, but in many cases I tried not to start my interviews saying I was conducting research on drug trafficking and violence, I rather used the word public security. All in all, my best protections were following common sense, and to a great extent, my gut feelings.   45   Addressing the issue of data reliability is highly complicated. As Andreas and Greenhill (2010) point out while discussing the politics of the production of statistics, the acquisition of good data about transnational illegal activities is politicized, particularly difficult to come by, and is plagued by morally charged debates. The truth on these topics, as William Finnegan pointed out in an article about Guadalajara,28 in conflict and crime ridden societies it is difficult to know who is behind a murder, a candidacy, the uncover of a corruption scandal, an enforcement operation; the truth is usually complex, fluid and difficult. By nature, the most important information on an illegal organization is hidden; official data on drug trafficking and drug violence is scarce, methodologically weak and biased by government’s efforts in showing progress; judicial reports are difficult to access, and when they can be accessed, they do not necessarily ask the questions that the researcher wants to address, and in turn, the traffickers response also tends to be biased; interviews are biased by recall issues, opinions about the government and the traffickers, and perceptions. In this context the main tool I have used to address data reliability is to triangulate as many sources as I can for each piece of information in my analytical narratives; in other words I cross-validate all the characterizations I present in this dissertation. Naturally, there is still room for ambiguity, biased information, and information gaps, but I have worked to the best of my abilities to use reliable information that can contribute in the expansion of a research field that is everyday more important in academia, and that is crucial for political, economic and social stability.                                                                                                                 28 Finnegan, W. 2012. “The Kingpins: the Fight for Guadalajara". The New Yorker, July 2.   46   1.5.2. Road Map In the chapters that follow I develop, test, and illustrate the relational political economy framework for explaining drug violence that I have sketched in this theory chapter. Chapter 2 discusses the concepts of visibility and frequency, as well as the complexities related to the definition and characterization of drug violence. The chapter analyzes how visibility is distinct from related concepts like brutality, symbolic violence, random or indiscriminate violence, and from dynamics of media coverage. The chapter also discusses the operationalization of drug violence, and the implications of introducing the concept of visibility for understanding violence. Chapter 3 presents an overview of the evolution of drug trafficking and the state responses to it in Colombia and Mexico. Building upon the paradox that may emerge from my argument, namely, that a cohesive state that is more efficient in enforcing the law could also be a state that protects criminals, the chapter argues that there may be a tradeoff between a state’s capacity to enforce the law, and the state’s autonomy from criminal influences, although the tradeoff is not necessarily a zero sum game. The chapter sets the stage for the subnational comparisons, showing that national level processes are insufficient to understand contrasting patterns of violence across cities. Chapter 4 explores the especially sharp variation in organized armed violence in Medellín. The comparison within one city over time provides a controlled test of the argument because it holds crucial variables constant, such as the specific economic importance of Medellín, its history as a hub for early industrialization, and the particular characteristics of its urban landscape. The chapter shows how the transformations in the   47   organization of the criminal world and in the state affect levels of violence in the city, and explain one of the most striking transformations that any city has experienced, from being the most violent city in the world in the 1990s, to a sharp decrease in homicide rates in the 2000s, when one criminal actor dominated the city. Chapter 5 further tests the argument by comparing Cali and Culiacán, two cities that despite being located in different countries have experienced similar forms of violence, characterized by their low visibility. These two cases illustrate how the persistence of strong political elites symbiotic with criminals, have created incentives for the latter to hide violence, even when facing disputes with other criminal actors. Cali and Culiacán also illustrate cases when criminals prefer to insource the use of violence. The relative stability of conditions in the two cities allows me to test how changes in the independent variables affect violence. I show that instances of visible violence have emerged when a component of the security apparatus becomes more fragmented, for example when rotations or purges occur in enforcement agencies. Chapter 6 focuses on Ciudad Juárez and Tijuana highlighting how cities within the same country, similarly shaped by their status as cities that border the United States, by the powerful influences of economic globalization, and by diverse migrant populations, can nevertheless experience diverging patterns of violence. The cross-city comparison allows me to challenge explanations that try to account for the most recent wave of violence in Mexico by focusing on the effects of proximity to the border. This chapter also illustrates how in the late 2000s when the drug war exploded in Mexico, insourcing and a more cohesive security apparatus in Tijuana led to a different dynamic of violence than the one experienced in Ciudad Juárez where outsourcing and   48   fragmentation became prevalent. Even though both cities experienced sharp increases in the frequency and visibility of violence in 2008, such increases were more limited and short-lived in Tijuana than in Ciudad Juárez. Finally, the conclusion summarizes the arguments and explores their potential to explain violent criminal behavior in other cities and countries. I identify avenues for future research, as well as implications for broader theoretical discussions. Finally, I discuss the policy implications of the argument for issues of violence reduction and drug control.   49   CHAPTER 2. SILENT TRAFFICKERS OR BRUTAL CRIMINALS: DEFINING AND ASSESSING THE FREQUENCY AND VISIBILITY OF DRUG VIOLENCE On November 6th 1986 “Los Extraditables”, an organization founded by members of the Medellin Drug Trafficking Organization (DTO), made its official appearance before the Colombian public opinion and government, releasing a document where it called for the end of extradition “in the name of national sovereignty, family rights, and human rights.”29 Over the following five years, the organization was responsible for a deadly period in Colombian history that claimed the lives of hundreds of civilians, judges, politicians and cops. The emergence and evolution of “Los Extraditables” was marked not only by its brutality, but also by its intriguing willingness to claim responsibility for its attacks, usually through communiqués where it expressed the criminals’ rationale and motivation for using violence. “Los Extraditables” interest in being publicized stands in sharp contrast to the efforts to hide violence deployed in the late 1990s and 2000s by the Arellano Felix organization in Tijuana-Mexico; perhaps the most illustrative example of these efforts was the use of acid to destroy corpses of people killed by traffickers. The remains of the bodies would then be dumped into sewage systems or buried in the outskirts of the city by employees of the criminal organization. Such techniques to hide the evidence of attacks only became public in 2009 when Mexican authorities captured criminal Santiago                                                                                                                 29 El Tiempo. 1991. “El Fin de Los Extraditables.” El Tiempo. July 4.   50   Mesa Lopez, known as “el pozolero30 del Teo.” Mesa recognized that over 10 years he and other collaborators had disintegrated more than 300 bodies as they worked for the Arellano Felix, and especially their enforcers, el Efra, and then for el Teo.31 The methods employed by the pozoleros reflected a completely different story to that of Los Extraditables. Rather than recognizing their responsibility on deaths, traffickers eliminated the evidence of their attacks. In fact, since the mid nineties a few human rights organizations were denouncing the tragedy of disappeared people thought to be victims of drug traffickers in the state of Baja California, and some of these victims could have been disappeared by “pozoleros.”32 The contrasting cases of “Los Extraditables” and “El Pozolero” highlight two crucial puzzles that have not been systematically analyzed in existing studies of drug violence: Why would drug trafficking organizations (DTOs) use visible violence, like that employed by “The Extraditables,” that can attract enforcement and media attention which can be detrimental for their business? And, if, as stated in chapter 1, violence is sometimes unavoidable to regulate illegal transactions, how can traffickers control and hide their use of violence, as occurred in Tijuana? To answer these questions, an unidimensional analysis of violence focused only on the frequency of violence would be insufficient to capture the more intricate variation illustrated by these contrasting examples, as well as the complex –and sometimes contradictory- incentives that drive traffickers’ violent behavior. As we will see in chapter 6, one of the most powerful aspects of violence illustrated by the case of “El Pozolero” is that the same traffickers’                                                                                                                 30 Pozolero can be translated as the stewmaker. Pozole is a traditional Mexican soup. 31 PGR. Declaración ministerial MESA III AP/PGR/BC/TIJ/217/09-M-III. 32 Author’s interview with the relative of a victim of disappearance in the mid 1990s in Tijuana.   51   violent strategies can change over time: Mesa’s testimony revealed that for almost a decade his job was to eliminate the evidence of violence; later in 2008 he was specifically instructed to, instead of burying the remains of the bodies, place them in the middle of a well-transited street, along with a note. Traffickers thus decided to change their strategy from hiding violence to making it visible. In this chapter I develop the multidimensional conceptualization of violence outlined in Chapter 1, which includes two dimensions: frequency and visibility. I argue that such conceptualization captures a wide range of variation in drug violence that is often overlooked in existing analyses. Thus, I get beyond the limitation of the still nascent literature on drug violence that tends to focus only on explaining extreme cases of high violence, and that usually is based on body counts, but does not explore systematically the methods used to carry out violence. At the same time, this conceptualization presents, in Sartori’s (1970) terms, precision to capture subtle variation in drug violence, and at the same time maintains a level of abstraction and a common taxonomy among the multiple ways in which drug violence can be performed and carried out. To develop the multidimensional approach to violence, the first section of the chapter analyzes the theoretical bases for considering frequency and visibility as two crucial dimensions of violence and explains how visibility is distinct from related concepts like brutality, symbolic violence, random or indiscriminate violence, and from dynamics of media coverage. The second section discusses key issues in defining drug violence and separating it from other forms of violence, and outlines causal mechanisms that explain the use and reproduction of drug violence. The third section explains the   52   operationalization of concepts. The chapter concludes by presenting the implications of introducing the concept of visibility for understanding violence. 2.1 Frequency and Visibility: Understanding the content of drug violence Literature on violence by non-state actors and during civil wars has provided us with the crucial insight that variation in levels, types and practices of violence needs to be explained and not simply assumed away as the result of pure irrationality (Kalyvas 1999, 2006; Tilly 2003). From this viewpoint, the apparently irrational behavior of criminals using excessive violence that attracts enforcement attention can be seen as a strategic decision. Of course, violence is not always rational, and not all violence perpetrators are strategic thinkers, but as we will see throughout this dissertation, many behaviors that may appear gratuitous are the result of the complex interactions between states and traffickers, as explained in Chapter 1. When criminals fear the action of a cohesive state, or receive protection from it, they may refrain from using visible violence and may actively hide it. Yet, when criminals do not receive predictable protection or do not fear enforcement actions, they may gain incentives to make violence more visible and thus take advantage of the capacity of violence to show power and induce fear. In this sense, the visibility of violence depends on the relationship of criminals with the state but it does not mean that visible violence is only used against the state: once the incentive to hide violence disappears, criminals may decide, depending on their interests, to use visible violence against their rivals, the state, or in some cases, against civilians. In civil wars, violence can be deployed strategically to manipulate the level of control and collaboration armed actors can elicit from the population, because violence   53   can signal power, strength, infuse fear, and serve as a powerful coercive tool. But not all forms of physical violence generate the same reaction, and thus insurgent actors can use symbolic violence and manipulate violent images in order to mobilize the population’s support because “different types of actions, against different types of targets, carried out under different circumstances, at different times will not only elicit different impressions on those who witness them, they will also influence the level of exposure these attacks can expect to receive in the first place” (Mc Cormick and Giordano 2007, p. 312). Literature that defines terrorism as a signaling game that emerges when rebels are weak and do not have the power to extract concessions or impose their will on their rival (Kydd and Walter 2006; Hoffman and Mc Cormick 2004) also highlights the signaling potential of violence, in this case terror, to show power, resolve, or trustworthiness. As we will see below, this signaling power can be very useful for criminal actors to scare away enemies, retaliate to government action, or attempt to modify the behavior of governments (as in terrorist actions). Yet, the survival of trafficking actors, unlike that of armed actors in civil wars and that of organizations that deploy terrorist tactics, may benefit from, but does not depend on, civilian support. Most importantly, unlike rebels and terrorist groups, criminals always need some level of cooperation within states in order to survive, and thus they depend on the direct or indirect support of state actors. Thus, an act of violence that can signal power not only increases the risk of detention as in terrorist actions (Hoffman and McCormick 2004) but also risks the friendly support or protection of certain state actors. Thus a characterization of violence in criminal markets should account for the effects that violence can have on the relationship of criminals with states, that are not only confrontational, but also collaborative. It also requires explaining   54   the range of violent strategies criminals can use, something that is not usually considered in the literature on terrorism that tends to study terrorism in isolation and not as part of a wider repertoire of violence. In criminal markets, different forms of violence lead the state to react differently to trafficker’s behavior. In Colombia in the 1990s, the government reacted and persecuted more forcefully Pablo Escobar and the Medellin DTO than the Cali DTO. As a former Minister of Defense acknowledged, the government’s focus on Escobar was a response to his extreme violent techniques. As the Minister put it, “ the Cali [DTO] became stronger while Escobar was attacked, and captured the business, because in the last months Escobar had become more an extortionist than a narco. A group placing bombs generates more social rejection.33” Similarly, in the context of the Mexican government’s declaration of war against trafficking organizations that initiated the deployment of federal and army troops to cities since 2006, certain forms of violence precipitated government’s decisions about where to mobilize federal troops. For example, a high level official I interviewed described how an incident that occurred in the state of Veracruz on September 20 2011, when masked gunmen stopped traffic on a busy highway and dumped 35 bodies in broad daylight, precipitated the federal government’s decision to deploy troops in the area. The event was not the first incident of violence in the region, but it was particularly visible: “[before that event] Veracruz did not report kidnappings and kept the statistics hidden. People didn’t know [about violence], but then, to make                                                                                                                 33 Author’s interview with former Defense Minister, Bogotá January 13 2011.   55   people aware, traffickers dumped 35 corpses. Those are clear images that say “I’m the boss.” Of course, we ran and we had to react to that. That’s how we move.”34 Different forms of violence also create different perceptions about criminals among civil society. People in Mexico tend to see the violence that occurred before 2006 as violence that respected the honor codes, and that was limited to “them” [the criminals], in sharp contrast to the generalized violence that ensued afterwards. Thus, as explained in Chapter 1, traffickers debate between using violence that is functional for them to solve disputes and show power, and avoiding violence that can be detrimental for the business while attracting attention, especially from law enforcement (Kenney 2007).35 Recent literature on illicit markets provides crucial clues to characterize variation in drug violence while exploring why it can change over time (Andreas and Wallman 2009, Garzon 2008, Snyder and Duran-Martinez 2009a) and reach extreme levels (Astorga and Shirk 2010, Lessing 2012a, Osorio 2012, Rios and Shirk 2010, Rios 2012). Yet this literature does not consider that different forms of violence can meet different strategic needs (Williams 2010) and that it is crucial to analyze these forms of violence systematically. Extant literature on illicit markets does not help characterize instances of violence like the ones occurring during narcoterrorism in Medellin, which stand in sharp contrast to instances where criminals deliberately attempt to hide their use of violence. Existing literature on drug trafficking and violence does not provide a useful concept that                                                                                                                 34 Author’s interview with official from the Federal Secretariat of Public Security, Mexico City, September 27, 2011. 35 Attracting media and enforcement attention may not necessarily be intended consequences of the use of visible violence by criminals. Their main intention may be for example to get the attention of rivals, but while using visible violence, they are more likely to call attention from law enforcement.   56   can assess when do criminals privilege showing power, or when they privilege avoiding attention. By contrast to other existing classifications of violence, the concepts of Frequency and Visibility capture the fundamental paradoxes in understanding drug related violence. Frequency refers to the rate at which violent events occur and, as explained in Chapter 1, it can increase as competition in the illegal market increases. When the market is competitive, disputes for territory or control among rival organizations (or within the same organization if there is a leadership dispute) generate more violence, as they are unlikely to be solved with a single homicide.36 In a competitive market, violence can beget violence (Kalyvas 2006, p.59; Scheffler 1999, p. 178; Osorio 2012; Poole 1995, p. 42) because violent acts may nurture revenge or retaliation impulses that may not stop until one of the competitors is eliminated or significantly diminished. The more competitive the illegal market is, the more likely that violence can be very frequent, as no organization is likely to prevail, as was the case in Cali and Medellin in the late 2000s when at least two organizations were fighting to control illegal markets in each city (the gangs associated with criminals Sebastian and Valenciano in Medellin and the organizations Los Rastrojos and Los Machos in Cali). By contrast, if a criminal organization holds a monopoly, violence can still be used to solve disputes, enforce contracts, discipline members of the criminal                                                                                                                 36 Rival organizations can also compete for other assets such as state protection. Under some circumstances, two organizations within the same territory can coexist (Rios 2012, Williams 2002, Idler 2012) but this is usually not ad-hoc but a result of coexistence pacts, or arrangements in which one organization charges another one for the right to operate in a given territory. In Mexico, such an arrangement can take the form of “derecho de piso” or the right to pass through a territory in exchange for a monetary payment. In Colombia, criminal actors and armed actors on coexist in a variety of arrangements (Idler 2012) with implications for the level of violence. There are indications that coexistence arrangements leading to peaceful interactions may have reduced violence in Cali in the mid 2000s, in Tijuana after 2010, and in Ciudad Juarez in 2011. Yet, the evidence also suggests that these criminal pacts tend to be fragile.   57   organization, punish transactions gone wrong, punish treason, or scare away enemies. Yet, under a criminal monopoly violence tends to be more sporadic because it is unlikely that it can generate retaliation, and thus more violence. The power of the organization deters revenge and may persuade other actors not to react. The objectives of violence, such as punishment, can be achieved with a single violent act. This can be the case even when a criminal organization holds the primacy or an almost monopoly of the market but is challenged by internal and external competitors. In Chapter 5 we will see how, for example, the Juarez DTO faced challenges in the 1980s and 1990s, but its power allowed it to control spirals of violence when market disputes emerged. As the level of control declines, the organization also loses its power to persuade other actors not to react or imitate its own violent actions. Thus, the frequency dimension can help assess situations in which violence declines, and thus, scenarios that seem to be counterintuitive to the idea that violence is inherent to illegal markets. Visibility refers to whether criminals expose the evidence and/or claim responsibility for their attacks. Visible violence provides a key advantage, which is the ability to show power or to infuse fear, but has the disadvantage of generating state attention. As we noted in Chapter 1, criminals may refrain from using visible violence if the state is cohesive, because a cohesive state may be more likely to predictably attack, or alternatively, protect criminals. Thus criminals may refrain from using violence than can force an effective or otherwise friendly state to act. By contrast, criminals may decide to use visible violence when the state is fragmented because they do not fear the action of the state, or they do not receive state protection that they can loose if they use visible violence that can force the state to act. The concept of visibility thus captures both the   58   situations that are usually more analyzed but not very common (high visibility) but also the most common forms that tend to be overlooked and less of a concern for the state, the public, and the media (hidden violence). Thus the concept can explain why and how traffickers decide to be either visible brutal traffickers, or “narcotraffickers in silence,”37 hiding their use of violence. Visible violence can be used to instill fear and show power and this is crucial in at least three situations: when criminals are fighting for leadership within an organization, when two or more organizations are fighting for turf within the same territory, and when criminals react to state action or want to persuade state officials not to enforce the law against them. In this sense one may argue that the structure of competition in the criminal market, and not just the structure of the state security apparatus, may also determine the visibility of violence as criminals benefit from intimidation and from creating a violent reputation when they need to eliminate rivals. Yet, as we will see in the case studies, if competition in the illegal market creates incentives to use visible violence but criminals fear the retaliation of the state, or fear losing the protection of the state, then they would prefer to keep their disputes hidden. This explains why criminal disputes in Medellin after 2008, as we will see in Chapter 4, have increased the frequency of violence, but are not as visible as the violence of the Extraditables and the Medellin DTO in the 1980s: criminals have learned that visible violence makes them more vulnerable to state action. Evidence derived from confessions made by paramilitary leaders after their demobilization in Colombia illustrates the tradeoff that exists between using visible                                                                                                                 37 I borrow this expression from an intelligence official from the Attorney General’s Office in Colombia (Fiscalía General de la Republica) I interviewed in Bogotá in 2010.   59   violence to signal power and hiding violence to avoid attention. It also illustrates the way in which relationships with the state determine the calculus to decide on which type of violence should be used.38 The confession of paramilitary commander Jorge Laverde before a judge captures the advantages and disadvantages of exposing violence: Laverde: Sir, In Villa del Rosario, where the sugar mill used to be, near the river, we built an oven where approximately 40 or 50 people could be cremated, once they were killed […] Judge: Aside from that oven, those ovens used to dispose of bodies of the remains of people that the paramilitary assassinated, did you use any other method to dispose of the bodies? Laverde: Yes sir. They were thrown in the river. Judge: Which river? Laverde: The Zulia River. Judge: Who made the decision to get rid of the people? Why get rid of the people? Why not simply kill them? Laverde: That began a long time ago in Urabá. The commanders in Urabá ordered that the people disappear. Salvatore Mancuso said “there’s a reason they’re killed. Let the community see who was killed and that we are cleaning up the area” But commander Castaño disagreed. He said, if dead bodies started pilling up, it would make the authorities look bad. And it could be detrimental to us as well.”                                                                                                                 38 Paramilitary groups in Colombia lie in the thin line between political and criminal violence as they had a political agenda and a consolidated control on international drug trafficking since the late 1990s. Their actions could thus be both motivated politically and by the need to advance the trafficking business. Their inclusion here does not assume a criminal or a political motivation in their actions but rather shows that the concept of visibility captures fundamental changes in the strategic use of violence, especially when state complicity is a possibility.   60   Referring to a similar effort to hide bodies in mass graves, paramilitary commander Ever Veloza described how the use of hidden or low visibility violence can be essential when criminals need to maintain the collaboration of state authorities: Judge: When did you stop leaving bodies in plain sight and start digging mass graves? Veloza: The commanders were responsible for that. More and more people were being killed in Urabá every day. The homicide rate was going through the roof. Judge: Under which commanders? Veloza: Police or army commanders. The homicide rate kept rising. They started being pressured by their superiors or the media or various government organizations or NGOs. Because violence was happening right under their noses. So they asked us to please get rid of the dead bodies, to bury them to keep the statistics from rising. That’s when the mass grave system began to be implemented and used. They authorized killings but only with the understanding that the bodies would disappear.39 The strategic interest in hiding violence was also evident in the declaration of Santiago Garcia Mesa, el pozolero, before Mexican authorities. He claimed that the main objective of the brutal technique to place corpses in acid was to disappear bodies. He was not involved in killing the victims and he received the corpses blindfolded –so he would not recognize them- after the actual killing had taken place. His only function in the organization was to destroy the corpses in acid and bury them; thus he had no link to the                                                                                                                 39 Confession hearings of commanders Jorge Laverde “El Iguano” and Ever Veloza “HH”. Footage on documentary Impunity, by Juan Jose Lozano and Hollman Morris, 2011. Emphasis on the text is mine.   61   motivations that caused the initial killing and his function was only instrumental: to hide the evidence of assassination. In sum, frequency and visibility capture instrumental changes in the use of violence, and the combination of these two dimensions could characterize a wide range of manifestations of drug violence. Figure 2.1 (also presented in chapter 1) characterizes variation in violence in the five cities analyzed in this dissertation considering these two dimensions, and each cell illustrates a specific type of violence. It is important to note that each dimension is a continuous rather than a discrete category and cells represent ideal types. Each type combines both visible and non-visible manifestations of violence, because we know that in general, forms of violence mingle and intermix and that there are always ambiguities in the classification of violence (Moodie 2010, p.60). In fact, the database on drug violence constructed for this dissertation, as will be explained in the third section, illustrates that visible acts of violence may not necessarily cause the majority of lethal victims in a city; given their nature, a few visible attacks can be enough to capture media and enforcement attention. In any given period there are fewer “visible” attacks than “low visibility attacks.” Yet, there are periods in which visible acts become much more recurrent, even if they still represent a minority of lethal attacks. Figure 2.1. Types of drug related violence and location of cases Low Frequency (LF) High Frequency (HF) Cali 84-88 Cali 89-10, Medellín 07-10 Low visibility (LV) Cd Juárez 84-94, Tijuana Cd Juárez 94-07, Tijuana 88- 84-87/10 95/98-08 Medellín 03-07 Culiacán 84-08 Tijuana 96-97 Cd Juárez 08-10 High visibility Cd Juárez 05-06 Medellín 84-93 (HV)   62   It is crucial to emphasize that frequency and visibility are not necessarily independent dimensions. Extremely frequent violence can be visible in and of itself. For example, when a city reaches a very high homicide rate it can gain media attention and sound the alarms of governments and civil organizations. But as will be explained throughout the case studies, especially those of Cali and Culiacan in Chapter 5, there are instances of very frequent but hidden violence that receive less attention than visible instances of violence. Visibility can also reduce frequency, because visible acts can sometimes act as deterrents for violent retaliation and thus, can reduce frequency, particularly when the organization perpetrating violence has monopolistic market control. But visibility can also create more frequent violence while inducing retaliation, especially when the perpetrators are operating in a competitive market. Most importantly, frequency and visibility are interconnected because criminals do not decide on the forms of violence they will use thinking about frequency and visibility as separate dimensions, but simultaneously. In other words, if the market is competitive and market disputes emerge criminals decide to deploy violence, but they simultaneously decide, depending on the conditions of the state security apparatus, whether that violence could be perpetrated visibly, or should rather be kept hidden. Yet, the disaggregation of the two dimensions and of the factors that affect the decision on each dimension provides a useful way to capture patterns of variation that could not be noticed by using a concept that conflates visibility and frequency.   63   2.1.1. Alternative ways of classifying violence In criminological studies, a well-known classification of drug related violence differentiates between psychopharmacological, economic-compulsive, and systemic types of drug violence (Goldstein 1985). Such classification thus differentiates between violence derived from direct consumption of drugs, violence derived from the need to find the economic resources to buy drugs, and violence derived from interactions within systems of drug distribution. This useful classification, however, does not capture the variety of forms that violence takes within systems of drug distribution (the primary focus of this dissertation), and that thus is more connected to societal, political, and criminal organizational factors, and more common than violence associated to individual behavior linked to the consumption of drugs (Goldstein 1997). Another classification of violence that has been particularly useful in the study of civil wars and that can potentially capture the paradox of hiding or exposing violence is the distinction between selective and indiscriminate violence (Kalyvas 2006). This classification, however, does not capture the fundamental puzzles presented here because selective and indiscriminate violence can be both visible and hidden. Acts of indiscriminate violence tend to be very visible, as was the case of car bombs during narcoterrorism, yet they can also be hidden, as when victims of massacres are buried in mass graves. Similarly, acts of selective violence can be both visible and hidden: violence in Mexico since 2006 was not less selective than it was during the relatively peaceful 1980s, the difference rather was that attacks became more public in nature even when criminal enemies were selectively targeted. Although one of the crucial transformations of drug violence in Mexico was that civilians started to be increasingly victimized after   64   2006, this was not only a result of the targeting of civilian populations by criminals or of indiscriminate attacks but rather of situations like misidentification both by authorities and criminals, “astray bullets”, and targeting of civilians when criminals engaged in extortion, among other circumstances.40 In the study of drug violence, one can also differentiate violence according to the purposes it can serve, such as solving contractual disputes, territorial, disciplinary, or succession issues (Reuter 2009). The problem with this classification is that a differentiation of the purposes of violence might be misleading as similar purposes can be met with very different forms of violence. As Kalyvas has noted (2006, p.23-25) a single violent event can mix multiple motivations (that range from ideological to pure revenge or sadistic impulses), and have multiple uses (from intimidation to provocation), and this is also the case in drug related violence. We can hypothesize the incentives that lead to a change in the observed performance of violence, but it would be more difficult to differentiate forms of violence trying to disentangle the purposes behind it. In drug trafficking markets, violence may serve instrumental objectives such as regulating transactions, building reputation and intimidating, or imposing social control. These objectives can overlap and serve a wider range of motivations (from solving succession issues to eliminating competitors), but the key issue here is that each of these objectives can be served with different performances of violence. For example, violence can emerge as a tool to regulate transactions in a contractual environment that cannot be [officially] policed and enforced by the state.                                                                                                                 40 Kalyvas notes that the effectiveness of selective violence hinges on the perception that a process of selection is taking place (2006, p.192). Along these lines one could characterize violence in Mexico since 2006 as indiscriminate as the population perceives that the “honor code” of criminals has been broken and they kill randomly and non-selectively. The problem is that such perception is created more by the visibility of acts than by a technology of non-discrimination.   65   Transactional violence thus targets rival traffickers or members of the organization retrospectively, when transactions go wrong, when members of organizations try to cheat or steal part of the merchandise, or when they provide information for enemies or for the state. If one organization holds market control this violence can be sporadic, used to discipline members, but if competition within and outside the organization increases, transactional violence can become more frequent as the perpetrators may have less capacity to control retaliation. Violence to regulate transactions can also be both visible and hidden. For example, Mexican organizations in the 1980s could punish traitors disappearing them or their remains, but punishments became a lot more visible in the late 2000s, as criminals employed a whole system of meanings in mutilated bodies that publicly exposed violence in order to signal the reasons for a transactional killing. For example a corpse with a mutilated finger was used to refer to a “dedo”, somebody who provided information to enforcement authorities. Criminals can also use violence on civilians as a way to impose social order, especially when they establish arrangements of protection with the state. The control of vulnerable sectors of the population seen as “public hazards” such as petty criminals, drug users, prostitutes, and beggars may increase DTOs’ legitimacy before certain sectors of the community and the state. As Arlacchi noted when describing the use of such violence by Mafiosi in Italy, this violence can actually be interpreted as collaboration to repress disturbances to the established order (1986, p.28). For the most part “social control” is not very frequent or visible, both because the pool of “undesirables” is limited and they do not have the power to retaliate, and because if criminals receive protection from the state they prefer to hide this violence, as occurred in Cali in the 1980s. Yet, in   66   some cases, social control can take on very visible forms. For example in Culiacan in 2009 out of 35 messages found on dead bodies, 30 were directed to common delinquents and car robbers; the notes, which were also accompanied by toy cars, threatened other car robbers with the same treatment if they did not stop stealing.41 This type of visible violence sometimes appeared to affect other “social maladies” such as rapists or even unfaithful husbands. The crucial point is that social control became visible in that juncture, because as we will see in Chapter 6, the state security apparatus had become more fragmented. As the examples of transactional violence and social control show, the instrumental objectives of drug violence can be served with both visible and less visible forms of violence, and the same violent act can serve multiple objectives, therefore the frequency- visibility classification is both more precise and comprehensive than one based on motivations-objectives. 2.1.2. Symbolic violence, media representations and brutality 2.1.2.1. Symbolic violence The concept of visibility is connected to ideas about the semantic and symbolic uses of violence (Blok 2002; Uribe 1990; Uribe 2004), which disentangle the use of violence as a communicating tool (Gambetta 2009)42 and highlight that not all forms of violence have the same communicating power. Uribe (1990), for example, analyzed and                                                                                                                 41 Author’s dataset on drug related violence. 42 Concepts like symbolic (Bourdieu and Wacquant 2004) and performative violence (Juris 2005) could also appear useful here but these concepts are usually applied to characterize non-lethal forms of violence such as protests, gendered violence, and not to parse out different manifestations of lethal violence.   67   described the symbolism and rituals entailed in massacres and mutilations in Colombia during the period of civil war known as “La Violencia” (1948-1964). She described how parts of the body were separated and sometimes relocated (i.e. the legs would be placed where the head used to be) to desecrate the victims. In some cases, the bodies were disposed so that those finding them would be scared away. Perpetrators would also put notes in massacre scenes so that people knew who the perpetrators were. Along these lines, many recent analyses of violence in Mexico have focused on understanding why certain methods, especially the use of beheadings, became so prevalent after 2006, and emphasize the combination of cultural and technological factors (such as the spread of communication technologies), and the varied meanings and symbols associated with different killing methods. The concept of visibility differs from concepts of symbolic violence that understand the performance of violence as a product of unique cultural manifestations rather than as an instrumental tool, and thus reduce the possibility of comparing manifestations of violence across cases (Kalyvas 2006, p.24). For example, Uribe (1990) explains the recurrence of massacres and mutilations among peasants that were very similar to each other as a result of a latent aggressiveness, a tendency to retaliate and defend honor codes, and superstitious ideas about the adversary ingrained in primary peasant loyalties and culture (Medina and Uribe 1995). Uribe was careful to emphasize that these practices were not simply pathological, and occurred under particular political circumstances, with instrumental objectives such as revenge and the elimination of members of opposite political parties. Yet, she did not explore why such cultural “propulsion” to kill in barbaric ways was activated only under certain conditions.   68   In some extreme cases, ideas about the symbolic use of violence as a cultural manifestation can lead to a simplification of the causes of violence and a tendency to stigmatize certain cultures. For example, the Mexican’s familiarity with death, as Lomnitz (2005) points out, has led to the characterization of Mexican lower classes as potentially barbaric. It has also been used to explain the brutality of violent methods as the product of cultural influences, such as Aztec indigenous practices, or in recent years, the proliferation of the cult to the Santa Muerte, “The Holy Death”, among popular classes and criminals (Bunker and Sullivan 2011). Even if such cultural influences inform brutal practices of violence such as beheadings, they cannot explain why such practices are not employed all the time, or why the same criminals may sometimes recur to these methods, while others they refrain from using them. The concept of visibility, as I advance it, focuses more on analyzing when and why these visible forms appear, rather than on disentangling the diverse meanings, rituals, and cultural practices attached to different forms of killing, although such analysis is equally important and interesting. 2.1.2.2. Media representations of violence The concept of visibility is also connected to research on media representations of violence and their impact on perceptions of insecurity, which emphasizes how different performances of violence can impact perceptions and ideas about security for those who are not the direct victims of violence (Slone 2000). This research also explores how media prerogatives affect the coverage of violent events. The cases of this dissertation do illustrate that media prerogatives may manipulate and create the visibility, or not visibility, of violence. In Medellín, for example, in the past decade local media has   69   promoted an active effort to reduce the prominence of violence in the news. Similarly, in Tijuana, Mexico, authorities along with business sectors have actively tried to shape perceptions of violence and the “image” of the city by manipulating the media coverage of violence. Indeed, a businessmen in Tijuana explained to me how a leading objective of businesses’ active engagement in security policies was to change perceptions about the city as a dangerous place: “We started to monitor tv stations, we looked at each newspaper to see how many beheadings and deceased appeared in the headlines, and then we identified who was the first announcer in each of them, and we sat with the newspaper owners and told them, “if you continue publishing violence your advertisers will stop buying publicity”.43” Media coverage of violence is not neutral or objective, and media outlets can be coerced or corrupted into reporting the acts of violence of certain criminal groups and not those of their rivals, or forced to reduce reports on violence when it is in the governments or business sectors’ interest to reduce perceptions of insecurity. As these examples suggest, media coverage of violence may account for the divergence between realities of violence and the public perceptions of it. Yet, the concept of visibility differs from the focus on how media can transform or manipulate representations of violence, because it focuses instead on the objective and instrumental changes of violence. Research on media usually downplays the importance of objective changes in the forms of perpetrating criminal violence, and usually focuses on the instances where media portray excessive violence, rather than on the instances where media refrain from reporting violence.                                                                                                                 43 Author’s interview in Tijuana, October 18, 2011.   70   Furthermore, media coverage of violence can shift as the violent techniques of criminals change, because certain forms of violence are more likely to be covered by media (a public display of a beheaded body is more likely to be covered than a person being shot in the outskirts of the city). For example, a local freelance journalist in Tijuana explained to me how her priorities in covering violence shifted as the violent techniques of criminals changed in the city. In 2008-2009 the city experienced an upsurge of “high- impact” violence (shootouts, beheadings) but in 2010 violence seemed to decline. According to her, violence still occurred in the city, especially in poor neighborhoods, yet it was not the kind of high impact violence experienced in 2008-2009. Consequently, she could not hope to compete successfully against colleagues in other parts of the country by reporting on this “low impact” violence when they could offer news agencies far more lurid stories on beheaded bodies or “terrorist” acts. Her decision not to report on hidden violence was thus a response to media interests, but media did not create the changes in violent techniques that led her to make that decision. Of course decisions on what can be covered can be affected by different reporting standards. For example in 2010, as the number of journalists killed and threatened in Mexico kept growing (30 journalists had been killed between 2006 and 2010)44 heated debates emerged and divided Mexican media outlets: some considered that brutality should not be reproduced in the media, while others considered that journalists did not create violence and could not be held responsible for reporting reality and informing the public. Many journalists I interviewed in Mexico agreed on stating that journalists cannot ignore the reality and their responsibility is to analyze, rather than simply reproduce,                                                                                                                 44 According to the Committee to Protect Journalists, 64 journalists were assassinated in Mexico between 2006 and 2012, and violence related to drugs and organized crime has made Mexico one of the most dangerous countries for journalism.   71   grisly scenes. The key point is that even if media play a role in making certain forms of violence more visible than others, or even if they are manipulated to report in favor or against a certain criminal organization, the government, or business groups, their choices of reporting also respond to objective changes in the performance of violence.45 2.1.2.3. Brutality The concept of visibility is also connected to, but differs from, ideas about brutality in the perpetration of violence. Research on traditional mafia organizations illustrates the difference between visibility and brutality. Mafiosi used highly brutal methods to impose sanctions, retaliate to offenses, gain prestige, and out power rivals (Arlacchi 1986, p.14; Blok 2000, p.28). Yet, the “honor dimension of murder” led Mafiosi to make that brutality visible only to the direct victims of violence, the members of the group, or the rivals. According to Arlacchi, the change from traditional to entrepreneurial mafia was marked precisely by an increase in the level and scope of violence, which was no longer “in check” by traditional honor codes, and thus became increasingly more visible for the public outside the highly localized niches of traditional mafia. The difference between brutality and visibility is also exemplified by the behavior of members of the Cali DTO, usually considered to be less violent than their Medellin counterparts in the 1980s in Colombia. They used brutality, but mostly in private spaces and usually as an instrument to hide their responsibility in violence. As we will see in                                                                                                                 45 It is also plausible to think that if criminals have the power to manipulate journalists and they allow reports on their use of violence, it is because they indeed have an interest in exposing their use of violence. Of course not all manipulation is direct: given the notable increase of violence against journalists in Mexico many of them have resorted to self-censorship.   72   Chapter 5, during the period of the Cali DTO’s operation, tortured bodies were thrown on the Cauca river and their fingerprints were destroyed, thus making it difficult for authorities to find and identify the victims, and consequently the perpetrators of violence. According to criminal investigators analyzing a similar case of bodies found in the Medellin River in 2011, “criminal organizations throw the corpses with the aim of disappearing them, and thus, eliminating the possibility of identifying the perpetrators.”46 Thus brutality could actually be instrumental in destructing the evidence of responsibility, and could be deliberately used to this effect. Increases in disappearances signal the same dynamic, as it is always more difficult for authorities to find the victims and the perpetrators of forced disappearance. The idea of brutality, thus, does not necessarily capture tactical or strategic changes in the use of violence. Visible violence can be, and usually is, brutal, but low visibility violence can be very brutal too. In sum, visible violence implies that brutality transcends the direct victims and is communicated to the general public. Here it is crucial to highlight that of course there can be different publics for violence, and violence is always visible for the direct victims of violence, but the concept of visibility advanced here captures those instances when visibility transcends direct victims. When brutality is displayed at the public level, then it constitutes a highly visible form of violence, as is the case with beheaded bodies that started to be displayed publicly on bridges or in the middle of streets in Mexico since 2005, but with tremendous recurrence since 2006. We could further disaggregate violence by considering who are the intended audiences for a particular form of violence (i.e. the state, rivals, citizens), but as in the case of motivations, we may not observe                                                                                                                 46 El Espectador. 2011. “El Rio la fosa más grande de Medellín.” El Espectador, February 8.   73   clearly who the audience is for a particular form of violence, and thus, focusing simply on the performance can still capture an observable change in violent behavior. 2.2.The perils of defining and determining the limits of drug violence In chapter 1 we disentangled the rationale of the argument of this dissertation by stating that violence emerges in drug markets to solve market disputes, succession and territorial issues, to protect turf, and this is in essence a minimalist – and clear- definition of drug violence. Yet, one of the crucial problems in analyzing drug violence is characterizing it and separating it from other forms of violence. This problem is especially severe in contexts like the Colombian one, where non-state armed actors with political or ideological motivations have opposed to, or allied with, trafficking organizations, and where illegal activities like the drug trade have become a crucial source of funding for these politically motivated actors. Paramilitary groups emerged closely tied to drug traffickers, but until the 1990s both actors had separated organizational structures, thus making it easier to separate drug violence from the political violence of paramilitaries.47 This situation changed after the demise of the Medellin and Cali DTOs in the mid 1990s, when paramilitaries became protagonists of drug trafficking in the country, making it even more difficult to determine whether a particular action was aimed at facilitating drug trafficking activities, at eliminating political enemies, or at securing territorial control. Regardless of the intentions that are                                                                                                                 47 It is widely known that in Colombia the FARC guerrilla has also made drug trafficking its main income source, yet with some notable exceptions, the FARC’s control has been mainly limited to cultivation and selling of coca paste, while their involvement in commercialization and distribution is limited (Tickner, Garcia and Arreaza 2011, p. 421).   74   difficult to disentangle, the reality is that violence had both political consequences and consequences for the organization of the illegal enterprise (Duncan 2005, p. 30). In Mexico there are similar problems for separating drug related violence from other types of violence. Even though the political component of criminal organizations is substantially weaker in Mexico than in Colombia48 there is debate on whether the actions of criminal organizations against the state should be defined as political violence. For some authors, criminals behave like criminal insurgencies that challenge the existing power structure (Bunker 2011), but the problem is still that for the most part, criminal actors in Mexico may not seek to replace or conquer state power (Lessing 2012a, Osorio 2013). Following Laitin and Fearon (2003)49 the actions of criminals could be considered forms of political violence as long as they weaken the state’s monopoly of force and territorial control and seek changes in state policy. Another difficulty in separating drug violence is that it may affect civilians, and thus a simple characterization of drug violence as affecting only trafficking organizations that fight for the control of turf, overlooks the profound complexity of violence. In fact, one of the most contentious issues since Mexican President Felipe Calderón declared war against trafficking organizations in 2006 was precisely the government’s claim that most violence in the country involved criminals and was related to criminal disputes. The government started to partially change this rhetoric after famous Mexican poet Javier Sicilia, whose son was killed in April 2011, organized a massive movement of victims                                                                                                                 48 Trafficking organizations in Mexico do not have explicit political aims, with the probable exception of La Familia Michoacana and other smaller groups that publicized ideological platforms (See Finnegan, W. 2010. “Silver or Lead.” The New Yorker, May 31; also Grayson, G (2009)) 49 Laitin and Fearon define a civil conflict as one that involves “fighting between agents of (or claimants to) a state, and organized, nonstate groups who sought either to take control of a government, to take power in a region, or to use violence to change government policies” (2003, 76).     75   known as the Movement of Peace with Justice and Dignity. The massive gatherings promoted by Sicilia unveiled the way in which drug violence reproduced, affecting many civilians that died as a result of abuses of force by state officials, of extortion, of “astray bullets” (being in the wrong place at the wrong time), and of a variety of causes that transcended, but were not detached from, trafficking disputes. Criminals can kill any civilian suspected ob being a member of a rival gang or suspected informant, and state authorities may abuse their power against any civilian suspected of being criminal, or worse, in collaborating with criminals, they can also target civilians that are suspected to oppose the criminals’ interest. Defining drug violence also becomes complicated because many organizations dedicated to drug trafficking may engage in other criminal activities such as extortion or human trafficking, and both of these activities can generate additional victims of violence that are not necessarily associated with a trafficking dispute. Could this be classified as drug related violence? If we consider that trafficking disputes generate dynamics that fuel other types of violence, then it does make sense to classify this type of violence as drug related. In Ciudad Juárez for example, extortion did not exist before drug disputes exploded in the city, military and federal forces moved in,50 and gangs became a crucial asset for criminal organizations; then extortion became a source of violence when civilians refused to pay extortion fees.51                                                                                                                 50 Observers in fact claim that federal forces where complicit, or carried out extortion themselves (Author’s interviews in Ciudad Juarez in July 2011 and March 2012) 51 For example on May 1 2010 a report from El Diario of Juarez read “The employee identified as Moises Hernandez Martinez, 17 years old, was assassinated by several men that shot him with two fire guns. According to the witnesses, the owner of the business had received several threats when he refused to pay “the protection quota”   76   Drug violence is also difficult to disentangle from the violence of other “common” criminals, because as explained in Chapter 1, criminal organizations may outsource the use of violence to actors outside their organizational structure, usually youth gangs that can provide cheap armed muscle. When outsourcing occurs, violence is likely to spike, as was the case in Ciudad Juarez in 2008 and in Medellin throughout most of the 1980s and 1990s. The mechanism explaining the spike is the symbolic power that the money and arms of criminals can provide to youth gangs. Since traffickers may not have the ability or the willingness to control all members of gangs, they may end up using their new skills and arms not only when contracted by traffickers, but also in the perpetration of common crimes like robbery or assault, or in the resolution of personal disputes (Gutierrez and Jaramillo 2005, p. 197). Outsourcing can also make drug violence more difficult to identify because it professionalizes gang members (engaging them in high level criminality) without providing enough training. Inadequate training can lead paid killers to make mistakes while carrying out actions, or to be deliberately less careful, thus for example opening fire in public spaces in order to target one person, increasing the possibilities of collateral victims of violence. Gangs may also take advantage of the symbolism of power derived from their association to organized crime and use violence to claim their own power, because gangs privilege symbolic claims to identity, territory, turf, and friend-enemy relations (Venkatesh 1997). In this sense, once criminal organizations outsource violence to youth gangs, gangs may internalize the use of violence for their own purposes. The crucial point to highlight here is that outsourcing to gangs can fuel forms of gang violence that are not necessarily related to drug disputes, and in that sense transcend a minimalist definition of drug violence.   77   Besides gang violence, drug violence, especially when it becomes highly visible and highly frequent, can be difficult to separate from the violence of militias or vigilante groups that emerge in response to increasing violence from criminal actors. For example in Medellin, death squads and militia groups emerged in the 1980s when drug violence was proliferating, as a way to “protect” citizens from common criminality. As we will see in chapter 4, eventually these groups became crucial protagonists of violence in the city (Salazar and Jaramillo 1992). Drug violence also is difficult to separate from the violence generated by non- trafficking actors that emulate violence of trafficking actors either as a way to free ride on the benefits of a violent reputation, or as a way to misguide authorities. For example, a common criminal may decide to behead the victim of a robbery with the hope that the authority will automatically think that it is a drug related homicide and stop investigating the real motivation of the crime. In Mexico this is illustrated with Los Zetas, a criminal organization that has been presented by the state, media, and academics as the most brutal drug trafficking organization, and as the precursor of the most visible killing methods such as beheadings in Mexico.52 The ever-increasing violent reputation of Los Zetas spawned waves of imitation, and the emergence of copycat groups that were not related to, but benefited from, the Zeta’s reputation while conducting other crimes such as extortion.53 Los Zetas themselves have learned from other groups: their first live                                                                                                                 52 The Stratfor Report “Polarization and Sustained Violence in Mexico’s Cartel War” January 24 2012 describes for example how Los Zetas prefer brutality while the Sinaloa cartel prefers bribing. Available at: http://www.stratfor.com/analysis/polarization-and-sustained-violence-mexicos-cartel- war?utm_source=freelist- f&utm_medium=email&utm_campaign=20120125&utm_term=engage&utm_content=link4&elq=632fd75 cce59438cb88a7ebe4b49cee4 [Accessed 1 February 2012 ]. 53 Bailey, J.2011. “What do the Zetas and Mc Donalds have in common?” Insight Crime December 5. Available at: http://insightcrime.org/insight-latest-news/item/1933-what-do-the-zetas-and-mcdonalds-have- in-common [Accessed 1 February 2011 ]   78   broadcasting of a torture in 2005, widely reproduced throughout the internet, closely mimicked a similar video posted by Hamas months before.54 Since the Zetas military antecedents55 and their brutal methods are undeniable, authorities and media can be quick to blame the Zetas for brutal attacks, even before they conduct a proper investigation, and this can benefit those copying their techniques. The difficulties of defining drug violence presented so far can be summarized as the result of: criminal objectives being mixed with other objectives (political, personal); drug traffickers targeting civilians; drug violence generating retaliation by non trafficking actors; drug trafficking generating imitation by non-trafficking actors, and traffickers empowering youth gangs. Given these dynamics, a minimalist definition of drug violence does not seem adequate to characterize the dynamics of drug violence. A more extensive definition of drug violence that includes the victims and events associated with the issues described in this section, however, could cast doubt on the validity of the argument that different combinations in the frequency and visibility of violence are the result of the interaction between the state security apparatus, the structure of competition in the illegal market, and the type of armed coercion criminals employ. Yet, as will be evident throughout the case studies, an extension of a minimalist definition of drug violence is compatible with my main argument. The problems derived from retaliation by non-trafficking groups, imitation by copy-cats, or violence derived from other criminal activities, emerge when the illegal                                                                                                                 54 Author’s interview with official from the Secretariat of Public Security in Mexico City, 2011. 55 Los Zetas was created in 1996 as the enforcement branch of the Gulf DTO and was formed by a group of deserted members of a special force in the Mexican military known as GAFES. Los Zetas remained dormant for a few years and since 2003 they have been growing in size and power, becoming one of Mexico’s most powerful criminal organizations, engaged not only in drug trafficking but also in a wide range of criminal activities such as extortion and human trafficking.   79   market is competitive and under dispute, but they are unlikely to emerge if the market is monopolistic. For example, imitation may only proliferate when the market is competitive and violence more frequent, because criminal organizations may have less power to discipline members or because they may benefit from the reputation reproduced by imitators. Violence by imitators may thus not be drug-related in the strict sense, but it is fueled by drug trafficking disputes. Likewise, the imitation of visible forms of violence by non-criminal actors may not emerge if there is not a change in the state structure that subsequently transforms the incentives of criminals, leading them to expose violence. As I mentioned earlier, not all violence, and especially not all visible violence, is rational and calculated, but it may only emerge on a sustained basis when the interaction between states and criminals has changed. Finally, the problems associated with gang violence being difficult to separate from drug violence may only emerge when traffickers outsource their use of violence to youth gangs. Otherwise traffickers may even try to actively reduce the independent violence generated by gangs, as occurred in Medellín between 2003 and 2007. The next section will discuss how to operationalize a broader definition of drug violence along frequency and visibility without stretching the concept. 2.3. Operationalizing frequency and visibility of drug violence If we decided to analyze drug violence from a minimalist viewpoint, reducing it to instances when criminals kill each other or state officials in their quest for market control, we would run into the problems of clearly determining and counting those instances.   80   These problems were painfully evident after the upsurge of violence in Mexico in 2006, which generated many efforts to count the victims of violence associated to organized crime, but also many unsolved controversies.56 The first efforts to count drug violence were initiated by the newspaper Reforma that published the numbers of “executions by organized crime.” Then, in 2011, the government released a database on “presumed homicides related to organized delinquency” that presented data on drug related homicides from to 2006 up to 2010. In 2012, the government officially abandoned the effort to keep producing the numbers, acknowledging that it was very difficult to determine a drug related murder. Both the government database and the newspapers account ran into the same limitations: justice and criminal investigation are weak57 and without having finalized judicial investigations and sentences to determine the perpetrator and motivation of a given homicide, it was impossible to fully determine if a given murder was associated to organized crime or not. As a result, the classification was based on aspects such as the method employed in the homicide (i.e .the corpse had high caliber/long arms shots or/and signs of torture or/and elements characteristic of organized crime such as blankets, duct                                                                                                                 56 For a discussion of some of the controversies counting drug related killings in Mexico see Ley, S. 2012. “El desafío de contra a nuestros muertos” Letras Libres, September 12, available at http://www.letraslibres.com/blogs/polifonia/el-desafio-de-contar-nuestros-muertos, [Accessed September 12 2012]; Mendoza, E, and Mossa. R. 2012. “El Presidente de las 83 mil ejecuciones” Zeta, November 26. Karlin, M. 2012. “Fueled by War on Drugs, Mexican Death Toll Could Exceed 120,000 As Calderon Ends Six-Year Reign” Truthout November 28, Available at http://truth-out.org/news/item/13001-calderon-reign- ends-with-six-year-mexican-death-toll-near-120000 [Accessed 28 November 2012] Molzahn, C; Rios, V., and Shirk, D. 2012. Drug violence in Mexico: Data and analysis through 2011. Transborder Institute, University of San Diego. Steller, T. 2012. “Years of killing hard to add up in Mexico.” Arizona Daily Star, November 27. 57 The impunity rate for homicides in Mexico was 80% in 2009 (Seguridad, Justicia y Paz, Consejo Ciudadano para la seguridad pública y justicia penal) and for Colombia it was calculated in 97% in 2008 (Barreto and Rivera 2009).   81   tape, or notes or/and the victim was presumably linked to organized crime).58 One of the biggest problems of this method was that the victims counted in this way were assumed to be members of criminal organizations, along the lines of a minimalist definition of drug violence. Not surprisingly, the government numbers generated anger among relatives of victims who found it unfair to stigmatize victims as involved in trafficking, simply because they had been killed in a certain way; human rights activists also criticized this counting effort for underestimating the violence perpetrated by state officials against civilians in an effort to generate results for the government’s “war against drug traffickers.” The difficulties for counting drug related deaths in Mexico have also been present in Colombia. Colombia has almost a six-decades history with violence and armed conflict, which has generated many efforts to count the victims of violence. Government, NGOs, and academics, have created sophisticated databases to count violent events related to the armed conflict (Restrepo, Spagat and Vargas 2006), but given the difficulties of separating drug violence and political violence, many murders that can be seen as serving both drug trafficking and political objectives are just classified as politically related. In the 1980s, an academic commission that presented one of the first comprehensive analyses of violence in Colombia, argued for the need to recognize the multiplicity of violent manifestations, to separate different forms of violence –economic, cultural, common-, and to recognize that political or organized violence were not necessarily the most common forms of violence (Comisión 1987). Later studies questioned the possibility of clearly separating common and political crimes, and                                                                                                                 58 Poire, A. 2010 “Los homicidios y la violencia del crimen organizado.” Nexos, February 2010. Full coding details about the government’s dataset on drug related violence are available at http://www.presidencia.gob.mx/?DNA=119   82   different forms of violence (Arocha, Cubides, and Jimeno 1998; Guzman 1993). Over the years, official data produced by the National Police, the Forensics Institute, the Presidency of the Republic, among others, has improved and some sources such as the National Police disaggregate aspects of deaths that could potentially help identify violence perpetrated by traffickers. All this data, however, generates the same questions that the database on organized crime related homicides generates in Mexico: short of investigation, can we assume that a homicide carried out in a certain way is drug related or not? Keeping in mind these problems for counting deaths, there is no easy solution for identifying drug related murders even if we employed a minimalist definition of drug violence, and for identifying the boundaries between drug motivated violence perpetrated by political actors, or political violence perpetrated by criminal actors, violence motivated by the maintenance of the profits of illegal activities, or even violence perpetrated by common criminals attempting to profit from the reputation of larger organizations. In this dissertation I address this problem by defining drug related violence as acts of lethal violence that emerge in the functioning of drug markets and can affect civilians, state officials, and criminals alike.59 Thus, I use drug trafficking as a lens to analyze the dynamics of violence in a given city, but I do not assume that every single act of violence emerges directly from a trafficking dispute and involves only criminals. I instead analyze how the power interaction between criminals and the state can inhibit or fuel dynamics of violence. Of course this choice runs the risk                                                                                                                 59 Drug trafficking can generate other non-lethal forms of violence such as displacement. This dissertation only focuses on lethal forms.   83   of overestimating drug related violence, but it provides the basis to distill the impact and incidence of drug trafficking. Following this definition, I operationalize the frequency of violence using homicide rates from official government statistics: the National Statistics Institute in Mexico (INEGI); the Forensics Institute (Medicina Legal) and the National Police in Colombia. Not all homicides in a given city can be attributed to criminal organizations, yet since the aim of this analysis is to understand how the power relationships between criminal organizations and the state can fuel or inhibit the general dynamics of violence in a city, the homicide rate is still the best indicator of the frequency of drug related violence. Of course raw homicide rates cannot be in itself an indicator of drug violence in any given city. Thus the trends in homicide statistics are analyzed in tandem with all other primary and secondary materials in order to determine whether up and down trends in homicide can be safely attributed to drug trafficking actors. As can be seen in Figures 2.2 and 2.3, all the cities analyzed in this dissertation have experienced puzzling variation in the frequency of violence both over time and also when compared with their country counterparts. In order to accurately characterize periods of high and low frequency I compare a city’s homicide rate with the national average, with that of its country counterparts, and with the evolution of the city’s rate over time. This is important because national and local dynamics can affect the absolute number of homicides, for example, between 1984 and 2008 on average Culiacán had a homicide rate (29.5) that was significantly higher than Mexico’s average homicide rate for the period (15.19) but that was significantly lower than Cali’s homicide rate for the same period (80.0). Thus compared to Cali, the frequency of violence in Culiacán was   84   not very high, but if one takes into account the existence of the armed conflict in Colombia and the fact that even within Latin America Colombia exhibits extremely high homicide rates (Soares and Naritomi 2007, Pinheiro et al 2012), 29 is indeed a very high homicide rate by international standards, but also compared to the country average. Figure 2.2. Homicide rates 1985-2010, Mexican cities Source: Author’s elaboration, based on data from INEGI   85   Figure 2.3 Homicide rates 1984-2010, Colombian cities Source: Author’s elaboration, based on data from Medicina Legal and Policia Nacional To operationalize the visibility of violence, I constructed a database on drug related homicides. I collected information from the main newspapers in each of the five cities, most of which, due to the years selected, are only available in hard copy at local libraries.60 Given resource constraints and the level of detail captured in the dataset, it does not cover the entire time period analyzed in the dissertation (1984-2010) but I selected years when violence had low and high frequency, to characterize the possible range of variation in the visibility of violence across different frequencies. The dataset comprises 6,497 violent events disaggregated along 60 variables that present information for at least three years in each city. The basic criteria for selecting events for the dataset was to include events that were directly attributed to drug trafficking, or that given the methods used, the victim, or the background provided, could                                                                                                                 60 I thank Mundo Ramirez, Rocio Duran, and Ernesto Cañas for their valuable assistance in constructing this dataset. Without their patience for going through piles of dusty newspapers and in a few cases, digital records, this database would have been impossible to construct.   86   be presumably linked to drug trafficking. Of course, as in the dataset created by the government in Mexico, the problem with this selection is that is can overestimate many events as drug related. I decided to err on the side of overestimation by excluding only events that could unmistakably be attributed to other causes such as domestic violence or common crime. However, the advantage of this choice in the framework of my definition of drug violence is that I am not assuming that all the victims of drug violence have to be criminals or law enforcers. It is crucial to note that news accounts do not provide an accurate source for the number and motivation of killings, but rather substantial information to characterize the way in which violence is carried out. Reporting biases, incomplete information, and over-reporting, are obvious limitations for using press sources as a coding source of violent acts. For example, in Mexico in certain areas, local journalists were forced to report the attacks of one organization thus contributing to bolster the reputation of power of this organization to the detriment of others’,61 or alternatively they may be persuaded not to report at all. The full coding criteria are presented in Appendix 1. The dataset assesses four dimensions of violent acts to characterize visibility: whether criminals claim responsibility for their attacks; the type of victims of violence; the number of victims per violent attack, and whether the methods used to perform violence expose the evidence of the attacks. The type of victims of violence captures visibility because when criminal actors choose to target public officials, enforcement officials, and public personalities like journalists, they know that they are exposing themselves to public attention. One does not have to assume that criminals target public                                                                                                                 61 Author’s interview with expert on security. Mexico City, August 2008.   87   officials to seek the attention of the state or the media, as in fact, public officials may be targeted simply for retaliation of their actions against criminals. Yet, it is clear that killing a rival does not expose the use of violence in the same way that killing a politician does.62 Since visibility captures an objective change in the performance of violence it is not necessary to assume the motivation or who the audience of a violent acts is. For example, the targeting of public officials can certainly signal a deliberate attempt to challenge the state, but state officials can also be targeted in retaliation when they protect criminals and somehow fail to deliver protection. Similarly, criminals can kill fellow criminals in visible ways to send a message only to their rivals, but they can also use visible violence against fellow criminals as an act of defiance to the state. The key point is that to capture the strategic change that visible violence signals, i.e. criminals do not fear the attention of the state, we do not necessarily need to determine the intended audience or the objective of the violent act. Thus the characterization of the type of victim captures a technique of visibility rather than an objective of it. The number of victims captures visibility because the larger the number of victims of a violent event, the more likely it is going to be noticed by the public, thus if criminals lose the incentives to hide violence, they may also lose the incentives and the capacity to avoid collective attacks. It is crucial to note, however, that many visible violent attacks may include only one victim, as when a single body is mutilated and exposed publicly. Collective attacks may be seen as a reflection of identification problems that can emerge when criminal markets are competitive, because if there are rivals invading territory it becomes more difficult for criminals to differentiate friend                                                                                                                 62 There might be cases of murder of public officials that do not capture much attention, as when there are isolated murders of street level cops. Yet when cops are systematically targeted, violence is visible, as occurred in Medellin in the 1980s or in Ciudad Juarez since 2008.   88   from enemy (Kalyvas 2006, p. 89). As we will see in Chapter 6, some massacres after 2008 in Ciudad Juarez reflected this problem: perpetrators looking for a rival but not knowing exactly who the rival was, attacked groups in parties, or even all the patients in a rehabilitation center. In this sense, one may argue that if collective attacks are a reflection of visibility, and they in turn reflect identification problems, visibility is driven not only by changes in the state security apparatus but also by changes in the organization of the criminal market, as identification problems are more likely to emerge when the market is competitive. The key, once again, is that criminals facing identification problems may still choose to avoid collective attacks that may expose their violence too much when they fear the reaction of the state or fear losing the benefits of its protection. Thus, collective attacks capture visibility because fearing loss of state protection or retaliation, criminals may still try to avoid committing collective acts of violence. Finally, the methods used to carry out violence capture visibility because they reflect whether the evidence of violence is exposed publicly or not. The list of methods I used was not established a-priori but constructed as information was collected. Table 2.1 presents the possible values of each indicator of visibility and whether they indicate low or high visibility.   89   Table 2.1. Indicators of high visibility and low visibility violence Dimension of violence Indicators H. Visibility/L. Visibility Criminals claim responsibility forCommuniqués, banners, notes on High Visibility attacks corpses Type of victims of violence Public officials; enforcement High Visibility officials; journalists; human rights workers Members of the underworld, Low Visibility “undesirables” (prostitutes, drug addicts, street criminals) Civilians, members of DTOs Number of victims per attack Three or more High Visibility One or two Low Visibility Methods63 Car bombs and explosions; High Visibility mutilation or incineration exposed in public spaces; corpses with notes Combats in the street; forced Medium Visibility disappearance when the victim is taken in public (known as levanton in Mexico and paseo in Colombia); drive by shootings (sicariato) Simple use of firearms; simple Low Visibility use of blunt objects and knives; strangulation Author’s elaboration based on dataset on drug related violence Table 2.2 presents some of the general findings of the database, summarizing the values for each indicator in each of city-years available. As visibility increases, so do the number of methods used to carry out violence, the number of victims per violent attack, and the number of victims that could be identified as public figures (politicians, cops, journalists). The table also includes information on the location of attacks (whether they occur in public spaces and whether they occur in the urban areas or in the outskirts of the city). Cases of high visibility, as indicated by the methods, type and number of victims, are also cases that tend to have a higher percentage of attacks taking place in urban areas                                                                                                                 63 The classification of methods by level of visibility was constructed through interviews and conversations with journalists and law enforcers about what events are more likely to elicit attention from the media and the state.   90   and not the outskirts of the city, and more attacks in public spaces.64 The location of attacks is useful because events that occur in the center of cities and public spaces make the use of violence more public than when they occur in the outskirts of the city, or in abandoned lots. Table 2.2. Dimensions of visibility in Colombian and Mexican cities Number of Number of Number public targets Attacks with attacks with City Year of [politicians, direct claim of % public % urban more than methods police, responsibility65 three victims journalists] Tijuana 1984 5 0 2 0 73.53 81 Cali 1984 7 2 2 0 78.05 92.6 Culiacan 1984 7 4 3 0 81.94 86 Culiacan 1996 7 1 2 0 90.52 60 Culiacan 1986 8 4 12 0 94.29 65.7 Culiacan 1992 8 3 4 0 87.21 64.7 Tijuana 1992 8 3 1 0 81.6 93 Cd. Juarez 1984 8 0 1 0 77.95 90.8 Cali 2009 9 3 4 0 58.33 100 Culiacan 2002 9 1 3 0 87.88 60.6 Medellin 1984 10 2 16 0 85.21 96.5 Medellin 2009 10 11 2 0 79.1 98.6 Culiacan 2009 11 11 8 13 88.6 74.1 Tijuana 2002 11 1 5 0 85.03 89 Tijuana 2010 12 20 9 9 78.77 93 Cali 1989 13 10 8 1 84.9 88.4 Culiacan 2010 13 19 22 14 88.7 65.7 Cd. Juarez 2010 15 101 61 4 80.3 96.3 Medellin 1989 17 26 27 3 85.05 97.6                                                                                                                 64 Yet, as can be seen in the table, the majority of homicides take place in public spaces and thus the category does not present enough variation to characterize visibility, especially because public spaces, as defined by the ownership, are not necessarily places that are constantly exposed to the public, as in the case of rivers or roads. A further disaggregation of location in the database has been used to complement this characterization. Thus for example, the cities that have more homicides in rivers are also cities where low visibility and high frequency have prevailed (Cali and Culiacán). The location of attacks has to be considered in tandem with the other variables, because an attack at a private space like a house can be very visible if it implies a shootout or an explosion. 65 It is crucial to note that many acts that involve notes and messages do not necessarily entail a direct claim of responsibility, that is, an organization signing. Here I include only the acts where there is direct claim.   91   To characterize scenarios of violence in each city, I triangulate information from several sources to cross check the validity of all characterizations and to complement the information for years in which the dataset is not available. I use official data on homicides to characterize frequency; the dataset on violence to characterize visibility; secondary sources such as academic analyses of violence and security in each city; my interviews; and human rights reports. The comparison between homicide rates and the results of the dataset allowed me to characterize different combinations of frequency and visibility. For example, Culiacán in 1992 and 1996 had a very high frequency of violence with homicide rates of 35 and 29 per 100,000 inhabitants, yet the indicators of visibility were very low, there were few methods for carrying out violence, few attacks with one or more victims, few public targets, and a very high percentage of attacks took place in the outskirts of the city. Thus these years are characterized as high frequency-low visibility. By contrast, Medellín in 1989 and Ciudad Juárez in 2010 present some of the highest homicide rates for all the city-years (258 and 250 respectively) and also some of the highest values in all the indicators of visibility, thus leading me to characterize those city-years as high frequency-high visibility. Human rights reports and interviews with human rights activists were crucial to identify instances of low visibility violence. For example, as described in chapter 6, interviews with human rights workers were crucial for me to identify the common, but not officially reported, practice of disappearing people that became widespread in Ciudad Juarez and Tijuana in the mid 1990s. Naturally, human rights reports also carry biases, but they are particularly useful to characterize non-visible violence, which is difficult to   92   identify precisely because criminals are hiding their actions. Careful triangulation of sources is the most important tool I used to consider all the possible biases found in any single source. One interesting finding of my dataset, as shown in Figure 2.3, is that the most common method for killing, the simple use of fire guns, is not the most visible one, and the most visible methods, such as mutilations with notes, represent only a small portion of the total number of victims, even in scenarios of high visibility. Yet, studying visibility, even if it does not generate most of the victims of violence, is crucial because a few visible acts can affect public responses and perceptions more directly than many non- visible acts of violence. For example, the Colombian government enforced extradition treaties for traffickers after the assassination of Minister of Justice Rodrigo Lara Bonilla in 1984. Likewise Mexican President Felipe Calderón officially deployed troops in his home state of Michoacán, which was the location for one of the first visible manifestations of violence up to that moment. In September 2006, the DTO La Familia Michoacana threw the severed heads of five bodies on the dance floor of a club in Michoacán, along with a note that read: “La Familia doesn’t kill for money, it doesn’t kill women, it doesn’t kill innocent people—only those who deserve to die. Everyone should know: this is divine justice.”66 In other words, visible acts of violence do not need to cause the majority of violence to be relevant, to reflect a change in the incentives criminals have, and to motivate state reactions. Of course, this also means that most victims of drug violence can be routinely ignored, and a relatively small number of attacks can create an image of                                                                                                                 66 Finnegan, W. Op. Cit.   93   extreme brutality. The common stigma associated to victims of drug violence by media and public (they died because they deserved it, or because they were engaged in something bad) perpetuates the invisibility of many victims of violence, the creation of what Chomsky has labeled as “worthy” and “unworthy” victims (Chomsky and Hermann, 1988). In any case, the key point to highlight is that the worthiness of victims is not only the result of media manipulation, but also of the way in which violence is carried out. Figure 2.3. Percentage of violent acts by method Author’s dataset on drug related violence 2.4. What does visibility tell us about criminal behavior? A multidimensional approach to drug violence like the one I advance captures complex patterns of violent behavior by criminals that transcend the usual scholarly focus   94   on the most extreme cases of violence. As discussed throughout this chapter, these extreme cases are as puzzling as those in which criminals seem to be peaceful and hide their use of violence. Introducing the concept of visibility has several implications for the analysis of the violent behavior of criminals, as well as methodological and policy implications. First, visibility shows that apparently low levels of violence sometimes hide situations where criminals have not abandoned the use of violence altogether, but rather recur to less visible forms of violence. For example, as we will see in Chapter 4, between 2003 and 2007 Medellin experienced a historic reduction in violence, yet some violence still existed but usually criminals hid it (for example in mass graves) so that authorities could not notice it. Introducing the concept of visibility also explains why some contexts where violence is very frequent may not generate strong state and civil society reactions, because violence is less visible and thus may not generate a sense of insecurity and fear in the population, or a sense that there is a threat to national security among government officials. Although the reactions to violence are not the focus of this dissertation, we will see that cities like Cali and Culiacán, described in Chapter 5, have not experienced the same outcry against violence as cities like Ciudad Juárez and Medellín, partially because their violence has been less visible. Second, visibility and a multidimensional conceptualization of violence reveal that low levels of violence should not be equated with the absence of criminality because peaceful outcomes, in fact, often result from the dominance of strong criminal powers. One of the most powerful illustrations of this is the above mentioned reduction of violence in Medellín, which occurred precisely when for the first time one criminal actor   95   gained monopolistic control of the criminal market. Thus, a reduction of violence does not equal the elimination of criminal organizations, or the elimination of the most powerful ones. Situations of relative peace thus need to be monitored constantly because monopolies and criminal arrangements that bring about peace are prone to instability, and thus to the recurrence of violence. Third, visibility implies that an evaluation of drug violence requires a lot more than just analyzing homicide rates and trends, calls for a diversification of sources of information, and requires attention to the ways in which violence is performed. In the most extreme cases of low visibility, if criminals eliminate successfully all evidence of violence, records may not even exist. In this case, records of disappearances and human rights reports are essential to document the extreme cases of low visibility, and may not emerge until after the cases occur. In less extreme cases, the characterization of violence becomes more time-consuming as it requires collecting and analyzing information like the one describe above (victims, methods, locations). Fourth, the disaggregation of visibility has implications for policies aimed at reducing or targeting violence. Any policy effort has to start from diagnoses of violence broad enough to cover varied manifestations of violence. If enforcement resources have to be relocated to address the most violent consequences of drug trafficking (Kleiman 2011), then the assessment of violence needs to be broad enough to capture instances of violence that may be ignored by authorities because they are not very visible. Such targeting also needs to recognize that refocusing enforcement on the most violent organizations can sometimes leave untouched powerful organizations that either hide   96   their use of violence, or that recur to less violence because they hold a criminal monopoly. Finally, and most importantly for the chapters that follow, the disaggregation of violent behavior reveals the varied and complex interactions that exist between criminals and state actors. It shows that even if criminal actors always need some collaboration and protection within states, such protection can occur in different ways and at different levels, with diverse implications for the behavior of criminals. Thus, only by unpacking the state security apparatus can we understand those varied interactions. The chapters that follow will illustrate how the frequency and visibility of violence describe the wide range of variation in drug related violence and capture in a parsimonious way the two fundamental paradoxes of drug violence, namely the ability to control violence in a context that requires violent intermediation, and the use of violence when it can attract unnecessary government attention. The cases will show that the distribution of power within the state security apparatus and within the criminal market, and the interaction between the two, determines when and how frequency and visibility vary. Before turning in the details of the five cities, the next chapter will present an overview of drug trafficking in Colombia and Mexico framed as a discussion about the potential tradeoff between a state’s coercive capacity and a state’s autonomy from criminal influences. The chapter will emphasize the complex interactions between states and criminals at the national level, but will also emphasize why a subnational perspective is necessary to understand violence.   97   CHAPTER 3. THE STATE CAPACITY TO ENFORCE THE LAW AND STATE AUTONOMY FROM CRIMINAL INFLUENCE IN CONTEXTS OF INSTITUTIONAL TRANSFORMATION Since the 1990s Latin American countries have experienced two crucial, and to a large extent, paradoxical changes. On the one hand, the region has democratized, establishing competitive elections and civil liberties. On the other hand, there has been an upsurge in violence, with complex roots and dynamics, and related in several countries to criminal groups and illicit markets. With only nine percent of the world’s population, Latin America concentrates 30% of the homicides in the world, and ranks far above regions with civil conflicts like Africa and the Middle East. While there exists wide variation within the region both in the quality of democracies67 and in the dynamics of violence, the coincidence of these two processes, puzzles scholars and generates many crucial questions. Perhaps one of the most pressing ones is whether states with stronger – although imperfect- democratic institutions have become less capable to enforce the law. In Colombia68 and Mexico this question acquires significant relevance because over the past three decades both countries have witnessed crucial transformations in their political regimes and also sharp variation in the levels of drug violence, with higher levels of                                                                                                                 67 Discussions about the stability and quality of democracies in Latin America, are far from over. Many problems have plagued new democracies where elections are stable, such as concentration of power in the executive. These problems have led many scholars to advance more fine-grained definitions of political regimes that capture the problems and paradoxes of democratic regimes, for example O’Donnell’s concept of delegative democracy (1994) or Levitsky and Way’s competitive authoritarianism (2002). 68 To be precise, the paradoxical coexistence of violence and democracy in Colombia predates the democratizing wave of the 1990s. For years, Colombia has been characterized as having both one of the oldest democracies in the region and the longest standing armed conflicts in the world (Gutierrez and Sanchez 2006) although as Bejarano and Pizarro (2005) point out, the nature of such coexistence changed radically with the democratizing reforms of the mid 1980s.   98   violence tending to coincide, in a very rough observation, with crucial periods of institutional change. This chapter presents an overview of drug trafficking and the responses to it by state institutions in Colombia and Mexico that illustrates why national level processes are important but insufficient to understand the contrasting patterns of violence seen across cities. Thus it sets the stage for the city-case studies and the subnational comparisons developed in the following chapters.. The starting point for this chapter is my argument that the visibility of violence can decrease when the state is more cohesive. This argument entails a potential paradox: a cohesive state that is more efficient in enforcing the law could also be a state where protection to criminals is more pervasive and predictable. Thus, this chapter tells the story of drug trafficking, state, and violence in Colombia and Mexico through two main arguments. First I argue that there may be a trade-off between the state’s ability to enforce the law and produce order and stability, and the state’s autonomy from criminal influences. However, such tradeoff is not a zero sum game: Lower levels of violence may result from less autonomy due to the existence of stable state protection to criminals, but not necessarily from capable and professional enforcement institutions. Alternatively, higher levels of violence, at least in the short-run, may indicate a breakdown of corruption networks and thus more autonomy, but not necessarily less corruption. And less cohesion in enforcing the law may occur in tandem with an increase in the professionalization and capacity of enforcement agencies. Second, and in relation with this, the evolution of Colombia and Mexico illustrates that increases in criminal violence, and the state confrontation with criminal organizations, may contribute to the growth and sophistication of the state security apparatus over time.   99   The chapter first presents a brief discussion about state capacity to enforce the law in contexts of institutional change. The second section describes the evolution of Colombia showing how drug trafficking consolidated within a fragmented state, then evolved in complex interactions with the armed conflict. Over time the fight against drugs and conflict contributed to increase the state coercive capacities and shaped criminal behavior, although drug trafficking persists and state capacity at the local level is uneven. The third section describes how drug trafficking in Mexico emerged attached to the centralized system of the one party rule of the Revolutionary Institutional Party (PRI), and then over time changed as the democratizing reforms and institutional transformations revamped the relation between criminals and state, reaching an extreme situation of disorder in which state capacity to enforce the law remains elusive. 3.1. State capacity and autonomy in relation to criminal actors In chapter 1, I argued that a cohesive state that is more efficient in enforcing the law, paradoxically, could also be a state where protection to criminal organizations is more pervasive and predictable. Processes of institutional change that redistribute power among state elites could affect state cohesion, as is the case with democratization. In consequence, democratizing states can be less cohesive than authoritarian ones, and therefore less likely to enforce the law efficiently or to protect criminals in a stable fashion. Authoritarian states can be more cohesive and thus more capable to establish the   100   monopoly of force, but they can also be less autonomous from criminal influence and corruption.69 Recent literature analyzing the paradox of increasing violence in an era of democratization in Latin America highlights the possibility of this tradeoff between state capacity to enforce the law and state autonomy from criminal influences. North, Wallis, and Weingast (2009) present this paradox arguing that states discourage the use of violence by providing elites with opportunities to access rents that are higher under conditions of peace than under conditions of violence. Thus, arrangements that can reduce violence may not necessarily bring elites (or criminals, for the purposes of this analysis) under the direct control of governments. As we will see, Mexico under the rule of the PRI illustrates this: the state had the authority to regulate the behavior of criminals but was unlikely to confront them. In this chapter I follow this line of reasoning, showing how during democratization states are more likely to confront criminals but do not necessarily have the capacity to control them. One crucial implication of this assertion is that absence of violence does not necessarily equate absence of criminality, or a triumph of the state in controlling criminals, but rather a regulated behavior of criminal actors. The second crucial implication is that violence may not necessarily reflect state absence or state collapse, but rather transformations in the terms of interaction between criminals and states. Of course, in many rural, isolated communities, violence and control by criminal actors may reflect a lack of state presence, or state collapse. But in many other areas, like                                                                                                                 69 An authoritarian state can potentially eliminate criminals, as was the case with the 1973 coup in Chile and the 1959 Cuban revolution. Yet, more research is required to determine how these authoritarian governments in fact dealt with drug traffickers.   101   the cities analyzed in this dissertation, the problems with violence cannot be characterized simply as an absence of the state. The idea of a tradeoff, however, has some limitations. First, it is possible that during democratization states can be indeed gaining enforcement capacities. Daniel Brinks (2012) asserts that there are two dimensions of law enforcement, horizontal (regulating relations between citizens) and vertical (regulating relations between the state and citizens). He argues that democratizing states can become better in vertical law enforcement as leaders become more accountable to their citizens and less likely to repress them. Yet, they may face more limitations in horizontal law enforcement, precisely because of the reduced capacities for repression, thus explaining increases in violence. In other words, democratizing states do not necessarily become less capable to enforce the law altogether, but less capable to coordinate such enforcement. The second limitation of the idea of a tradeoff is that an authoritarian state that is more cohesive, thus more effective in enforcing the law, is not necessarily a state with modern, professionalized, enforcement agencies. Conversely, democratic states can progressively acquire more professional enforcement agencies, but because of fragmentation, enforcement is less effective. An authoritarian state, can, because of its ability to control power, conduct effective enforcement operations, if it needs to regulate transactions with criminals. Yet, this ability may derive more from its centralized power than from its enforcement techniques. Thus, violence may not only reflect an erosion of the state’s coercive capacity, but also a fragmentation of the ways in which the state deals with non-state criminal actors.   102   The third limitation of the idea of a tradeoff between the state capacity to enforce the law and the state’s autonomy from criminal influence is that democratic states are not necessarily free from corruption. In fact, as will be seen in the case of Colombia, democratizing states can indeed increase the opportunities for criminals to access and influence the state, although at the same time can increase the number of actors willing to prosecute criminals. Overall, it is difficult to assess whether a democratic state is more or less corrupt than an authoritarian one. Yet, we can assert that corruption networks are less entrenched and less predictable in democratic states than in authoritarian ones, somewhat along the lines of Olson’s (1993) stationary and roving bandits model. In the following pages I illustrate how the potential tradeoff has played out in the processes of institutional transformation and criminal evolution in Colombia and Mexico. Changes in political competition, institutional reforms, and anticorruption policies have made criminal-state interactions more unpredictable. Thus, they have generated more incentives for criminals to employ visible forms of violence. The challenge for these countries is therefore to become more cohesive, but also more accountable and autonomous from criminal influences. 3.2. The evolution of drug trafficking and institutional transformation in Colombia 3.2.1. Drug trafficking consolidates within a democratizing state (1980-1998) The history of drug trafficking in Colombia dates back to the 1960s (Saenz Rovner 2008) when marijuana production emerged. In the 1970s marijuana became a highly profitable, and in many occasions, violent market (Britto 2010, 2013). Marijuana production infused the economy with constant cash flows; the government even eased   103   controls on exchange houses that were known to trade dollars derived from the marijuana income, in an effort to benefit from these revenue streams. This process was colloquially known as the “sinister window”, as it allowed marijuana profits to easily enter the legal economy. In the late 1970s, policies of the government of President Julio César Turbay (1978-1982), in tandem with the increasing production of marijuana in Mexico, led to the decline of the marijuana boom (“bonanza marimbera”) in Colombia. As marijuana production lost prominence, Colombians’ centrality in cocaine trafficking skyrocketed. Colombians replaced the Bolivian, Cuban, Chilean, and Peruvian, cocaine smuggling networks that dominated a nascent cocaine trade that had emerged in the 1950s and 1960s (Gootenberg 2009).70 Interestingly, cocaine trafficking did not grow through the networks of the marijuana trade (Betancourt and Garcia 1994, Britto 2012, Saenz Rovner 2008). Rather, cocaine networks emerged linked to other smuggling traditions, such as cigarette smuggling. By the mid 1970s, two organizations rose to dominate the international cocaine trafficking market: the Rodriguez Orejuela brothers, based in Cali, and Pablo Escobar and the Ochoa brothers, based in the city of Medellin. These two organizations fed an archetype that became dominant among enforcement officials and many journalists: that of drug trafficking organizations depicted as “cartels” that were centralized, hierarchical, and omnipotent in controlling the illegal market. Over time, it has become clear that these organizations operated more like networks than like highly centralized and organized                                                                                                                 70 Gootenberg traces the origins of cocaine as a legal global commodity to mid 19th century Peru. Then he describes how the legal market slowly deteriorated up until the 1950s, and then an illegal commodity chain emerged from the late 1940s onwards. One of the key trends highlighted by Gootenberg, but still underresearched, is the way in which key regime and political transformations, such as the 1952 revolution in Bolivia, the 1959 Cuban Revolution, or the 1973 Chilean coup, transformed relations between states and drug trafficking networks, and shaped the specific routes, flows, and behavior of criminal actors.   104   groups (Kenney 2007), and thus, that the label of cartel is often times misleading. Yet, it is also clear that these pioneer organizations were very powerful and more centralized than the drug trafficking groups of the 21st century. When these early powerful cocaine networks consolidated in the 1980s, Colombia was not yet the main coca leave producer in the world. Initially, Colombian traffickers focused more on processing coca leaves produced in Peru and Bolivia, and thus, the potential cocaine production from coca cultivated in Colombia was low. In fact, as shown in Figure 3.1, coca cultivation and cocaine production tipped off only after early trafficking groups were dismantled in the 1990s. Thus, even though the inflow of resources derived from drug trafficking was huge in the 1980s and 1990s, the power achieved by traffickers in the early stages of cocaine trafficking was not simply a result of the size of the illegal market. Rather than the size of the market itself, the interaction between a new and shocking wave of resources that empowered different sectors of the Colombian society, and the conditions of a democratizing political system, determined the violent dynamics of trafficking groups in the first stage of cocaine trafficking in Colombia.   105   Figure 3.1. Coca cultivation and potential cocaine production in Colombia Source: Author’s elaboration. Based on UNODC data from “Colombia Monitoreo de Cultivos de Coca 2010,” June 2011. Coca cultivation between 1994 and 1998 is from the US Department of State, cited in the UNODC report. Early cocaine traffickers consolidated their power precisely while the political system was undergoing a democratizing transformation in Colombia. For almost two decades (1958-1974), the political system functioned under a power-sharing agreement known as the Frente Nacional (National Front). Designed as a conflict mitigation strategy, this agreement evenly distributed electoral posts and bureaucracies between the two main political parties (the Liberal and Conservative Party) that had been the center of a civil war since the 1950s. The scheme excluded any effective electoral participation by alternative political forces, and thus, even though the country was democratic in a minimalist sense (there were regular elections), there were serious restrictions for electoral participation. In 1974, the agreement came to an end, thus paving the way for a   106   democratic opening. In 1986, the democratic opening expanded with the initiation of an extensive decentralization process (Faletti 2005) that established the popular election of Mayors and effectively reshaped the territorial bases of power in the country, during the presidency of Belisario Betancourt (1982-1986). The institutional reforms of the 1980s fragmented the distribution of political power in Colombia, creating a set of contradictory opportunities for emerging traffickers: on the one hand, the fragmentation of power increased the opportunities for traffickers to access the political system, but on the other, it increased the incidence of politicians that were likely to oppose the growing power and influence of traffickers, thus making state protection very unpredictable. This contradiction is illustrated vividly by the political behavior of traffickers like Pablo Escobar and Carlos Lehder who formed their own political movements. As we will see in chapter 5, Pablo Escobar ran as a candidate for Congress and was elected as an alternate member for the Lower Chamber of Congress in 1983. His infamous attempt to become a politician reflected the doors that democratization was opening for traffickers to access the state. Yet, the fierce opposition by other recently formed political actors (such as the party Nuevo Liberalismo) to Escobar’s political strategy reflected that democratization could also undermine the interests of traffickers. 3.2.1.1. Criminals confront the state (1984-1993) Justice Minister Rodrigo Lara Bonilla, a member of Nuevo Liberalismo, became one of the leading voices opposing Escobar’s participation in Congress, as well as the central figure in an enforcement campaign against the Medellin DTO. His efforts led to   107   Escobar’s ousting from Congress in 1983, and to the discovery and destruction of a vast cocaine processing facility known as Tranquilandia, located in the southern Amazon region of Colombia, in 1984. In Tranquilandia, the Medellín DTO established housing, landing strips, and laboratories for processing coca leaf originated in Perú and Bolivia, and also initiated the expansion of coca cultivation in Colombia. At the time, the operation constituted the largest drug bust ever (Thoumi 1995, p. 213). In retaliation to these actions, Escobar assassinated Lara Bonilla on April 30, 1984. The government in turn enforced extradition treaties, and extradition became the target of traffickers’ violence and war against the state. In 1986, the Medellin DTO created “Los Extraditables” an armed branch aimed at achieving the prohibition of extradition. In the years that followed, the country experienced an extremely violent period known as narcoterrorism, that left thousands of civilians, journalists, more than 200 judges, thousands of police members, and several prominent politicians dead, and that was imprinted in the memories of Colombians with prominent terrorist attacks such as the bombing of a major national newspaper and the headquarters of the civilian intelligence agency DAS. Lara Bonilla’s assassination led the government to define drug traffickers, especially those of the Medellín DTO, as a national security threat for the Colombian state. Yet, the state security apparatus was weak and unprepared to face the threat. Police forces were not very professional and were plagued by problems of corruption and inefficiency. Enforcement actors lacked coordination. The state, even though potentially more autonomous from criminal influence, had little capacity to enforce the law effectively and reassert its coercive capacity against the emerging narcoterrorist threat.   108   In 1983, the General Procurator Carlos Mauro Hoyos had already expressed that the extreme fragmentation of the state security apparatus was a grave hindrance in the fight against drug trafficking. According to him [the] judicial police lacks coordination. The Law gives the Procurator the mission of overseeing and coordinating the judicial police, but the reality is that such police is made of autonomous contingents that answer to their own commands and that come from the Police, the Security Administration Department DAS, from the Procurator’s office itself […]. As long as there is no specialized policing body responding to one command and with unity in terms of hierarchy, direction and policy, and as long as it does not have specialization in the drug trafficking function […], we won’t be able to say that we have a policing mechanism that corresponds to the gravity and urgency of our problems.71 Thus, even though the Colombian police had been formally centralized since 1950, it lacked the capacities and coordination to combat the emerging threat derived from drug trafficking violence. The first governments that confronted the visible attack of the Medellin drug traffickers faced a dilemma: they could not use the police that was perceived to be corrupt and inefficient, and thus they attempted to use the military, but the military did not cherish the idea of being deployed in antinarcotic responsibilities, especially in urban areas.72 The reluctance of the military reflected their focus on combating guerrillas in the context of the armed conflict that affected the country since the 1950s. Trafficking organizations were relatively independent from the leftist guerrillas, and thus, the military did not see the combat to drug trafficking as a way to achieve the greater objective of eliminating guerrillas.                                                                                                                 71 Procurator’s intervention before the Senate, in Anales del Congreso, August 30 1983. 72 In the 1970s, the military participated in marijuana crop eradication and it did not necessarily welcome these early antinarcotics roles (Palacios and Serrano 2010, p.120) but as these operations where mostly rural, they were not as contentious as the first attempts to deploy the military in urban antinarcotics operations.   109   It is essential to note that early drug traffickers interacted with armed actors in a variety of ways: traffickers sometimes hired guerrillas to provide security for the expanding coca crops and may have provided funding for some guerrilla operations of groups like the M-19.73 Paradoxically, traffickers also allied with, and even created, some of the first paramilitary groups that opposed guerrillas. One of the first groups that mixed traffickers’ violence and antiguerrilla actions was the group Death to Kidnappers (Muerte a Secuestradores MAS) formed after the M-19 guerrilla kidnapped the daughter of Jorge Luis Ochoa, one of the main traffickers of the Medellin DTO. Paramilitaries, traffickers, landowners, and sectors of the military also created strategic alliances against guerrillas (Gutierrez and Baron 2006, Romero 2003) but the alliances were unstable and changed constantly (for example some sectors of the paramilitaries eventually allied with the Cali DTO against the Medellin DTO). Despite the complex connections that linked early trafficking groups to the armed conflict, until the 1980s the government still perceived conflict and trafficking as separate threats, and the military was reluctant to fully engage in the fight against drug trafficking. As we will see later, this situation eventually changed in the mid 1990s. In 1986, and as the threat of narcoterrorism kept growing, President Virgilio Barco (1986-1990) assigned a crucial role of policing to the military in antinarcotics operations. Barco did not trust police forces and wanted to centralize all antinarcotics actions in the military. The decision generated discontent in the military, conflicts with                                                                                                                 73 An event that remains particularly controversial regarding the collaboration between traffickers and guerrillas is the seizing of the Palace of Justice by 34 members of the M-19 on November 6, 1985. According to a Truth Commission appointed to clarify the events of that day, drug traffickers may have funded this operation partially, which favored their larger opposition to extradition policies (Comisión de Verdad, 2006). However, the analysis of the Commission does not provide conclusive evidence proving the extent and type of collaboration between the traffickers and the guerrilla.   110   police forces, and a grim record of human rights violations, especially in the urban operations that were carried in Medellin in the late 1980s, and that we will analyze in chapter 4. The security apparatus of the democratizing state was thus extremely fragmented and weak because of the conflicts, corruption, and lack of capacity of the enforcement agencies. The first election of local mayors in 1988 added an additional layer of complexity to the decision making process in security policies. On top of this, the government of President Virgilio Barco adopted a political arrangement known as the government-opposition scheme that further undermined the government’s capacity to carry out any successful security program. The government-opposition scheme was aimed at protecting the political opposition by clearly separating the public posts that belonged to the opposition from those that belonged to the government; this ended up making it more difficult to forge political alliances to approve any new policy proposal in a highly convoluted security period that required rapid, but often difficult, decisions.74 Paradoxically as the state was looking for ways to strengthen its security apparatus, it was also facing a huge challenge in its capacity to maintain the rule of law. The problems in using the military led President Barco’s to attempt an alternative enforcement scheme in 1989, creating “islands of efficiency” within the weak and delegitimized police, such as the “elite police force” (Cuerpo Elite). This approach also proved limited and fraught with problems and the Cuerpo Elite was plagued by scandals of corruption and abuse. By the end of Barco’s term, Colombia was undergoing one of the most violent crises in history, and the narcoterrorist war had taken an even more extreme dynamic with the assassination in 1989 of Luis Carlos Galán, Presidential                                                                                                                 74 Author’s interview with former Defense Minister. Bogotá, January 13, 2011.   111   candidate of the Nuevo Liberalismo, who actively opposed drug traffickers and Pablo Escobar. 3.2.1.2. Constitutional reform in the midst of crisis In 1990 César Gaviria, the appointed successor of Luis Carlos Galán, was elected president. Gaviria summoned a Constitutional Assembly and eventually sanctioned a new Constitution in 1991, thus completing a political process initiated during the Barco administration. The 1991 Constitution introduced crucial political changes for the country, and important transformations in the structure of law enforcement and justice, such as the creation of the Attorney General’s Office (Fiscalia General de la Nación) and of multiple mechanisms for the protection of individual rights. Crucially, the Constitution also banned the extradition of Colombian nationals. Even though banning extradition had been one of the key objectives of the narcoterrorist campaign, violence continued75 (Pardo 1996). The profound institutional changes in 1991 provided an opportunity for Gaviria’s government to experiment with specialized police units again, with the expectation that the checks and balances created by the new constitution would prevent abuses like the ones that occurred with the Cuerpo Elite (Pardo 1996, p.448). Thus, in 1993, the government created a joint military-police command known as the Search Bloc (Bloque de Busqueda) to dismantle the big trafficking organizations and specifically to hunt down Pablo Escobar. This unit received training and military resources from the United States (Bowden 2002) and also facilitated the coordination and channeling of US aid, which up                                                                                                                 75 It is important to note that Pablo Escobar handed himself to authorities in June 1991, one month before the proclamation of the Constitution. After his brief stint in prison, as well will see in Chapter 4, violence took on a new extreme level.   112   to that moment had been fragmented between several US and Colombian enforcement agencies.76 The Search Bloc eventually succeeded in bringing down Escobar in 1993. Over time increasing evidence has suggested that the Search Bloc did not operate alone, and received support and information from Escobar’s trafficking enemies. However, the model set up with the Search Bloc became a blueprint for subsequent enforcement operations. In fact, a new Search Bloc was set up to dismantle the Cali DTO. The operation of this new Bloc, however, reflected that the strategies of drug traffickers and their relations with local and national states varied significantly across regions, and in consequence, enforcement operations also varied depending on those conditions. Because of the pervasive corruption of the police, the Search Bloc in Cali depended more on the Army than on the Police. 3.2.1.3. The transformation of the National Police After Pablo Escobar’s death, the pressure to dismantle the Cali DTO increased, not because the Cali DTO was a security threat similar to the Medellin DTO, but because its corruptive power became too difficult to hide. During the Presidency of Ernesto Samper (1994-1998) the extensive links between the Cali DTO and large sectors of the police and the political class came to light when accusations emerged that Samper’s presidential campaign received funds from the Rodriguez Orejuela brothers, as we will see in chapter 5. The judicial process initiated with these accusations, known as Proceso                                                                                                                 76 Author’s interview with former Minister of Defense, Bogotá, January 13, 2011.   113   8000, linked 21 senators and representatives of Congress,77 hundreds of cops, private companies, and national level authorities such as the General Comptroller of the Nation and the General Procurator, with the Cali traffickers. President Samper himself was absolved, but the scandal seriously undermined his authority, especially in front of the United States government, that decertified Colombia and cancelled Samper’s visa. In the midst of the diplomatic crisis, the Director of the National Police at the time, General Rosso Jose Serrano, emerged as the most credible interlocutor for the US government that perceived him to be a champion against corruption. The US prioritized its relations with the police and the military (Ramirez 2010, p.92). Serrano took upon the task of “cleaning” up the police forces and dismantling the Cali DTO, and by the time he stepped out of his post, the international image and power of the National Police had been completely revamped. The reforms conducted under Serrano’s leadership, most notably an increase in the police investigative and technical capacities, appeared as a sort of counter reform to a reform effort initiated in 1993, aimed at strengthening the civilian nature of the police, professionalizing it, decentralizing it, and increasing community oversight. The scandals of trafficking related corruption, especially in the Cali police, and the need to effectively dismantle the Cali and Medellin DTOs, countermined the democratizing impetus of the previously proposed reform. The director of the National Police recentralized the institution in order to conduct the massive firing and rotation of about 7,000 cops between 1995 and 1998 (Casas Dupuy 2005). The institution also became more independent from civilian oversight, and police roles became closer to the military than to                                                                                                                 77 El Tiempo. 1996. “Detenido el senador Gustavo Espinosa.” El Tiempo, February 15.   114   a civilian police (Camacho 1994, Casas Dupuy 2005, Llorente 2005). Thus in the mid 1990s the police gained more enforcement capabilities, although it did not necessarily become more democratic. However, by 1995, the National Police could present itself as a highly effective institution, that eliminated the most dangerous man in Colombia (Pablo Escobar) and incarcerated the highest echelons of the Cali DTO78 thus dismantling the big trafficking organizations. In sum, during the period of consolidation and demise of the first big trafficking organizations between 1980 and 1998, the structure of political power in Colombia, the democratizing changes taking place, and the fragmentation and corruption of enforcement agencies, made it impossible for politicians to centralize the control over trafficking networks, as occurred in Mexico (Palacios and Serrano 2010) or to enforce the law efficiently. This in turn, created the environment that incentivized criminals to confront the state. In turn, the state’s response to the rising threat of traffickers radically transformed the state security apparatus, and police forces emerged more powerful and legitimate at the end of this period. Levels of coordination between police and military increased, and the technical capacities of the police grew significantly. The pressure to confront narcoterrorism contributed to extend the power of both the national police and the military over the course of a decade, thus highlighting the state-making potential of the war against drug trafficking. It is important to emphasize that the Colombian state in this period could be characterized as a fragmented state at the national level, yet there was significant variation throughout the territory. In Medellín, as we will see in Chapter 4, the                                                                                                                 78 Even though after their capture the Rodriguez Orejuela continued operating in jail until 2000, when they were extradited.   115   fragmentation of the security apparatus made it very difficult for the Medellín DTO to establish predictable links with state officials. While many cops and politicians collaborated with Pablo Escobar and the Medellín traffickers, others opposed him actively. By contrast in Cali, as we will see in Chapter 5, traffickers secured protection within a more cohesive state security apparatus at the local level, and this ability motivated the less visible violent strategies used by the Cali traffickers and their successors. This subnational variation illustrates two points: first, a fragmented state at the national level operated differently across regions and subnational authoritarianism coexisted within the democratizing state (Gibson 2005). Second, even though cocaine trafficking thrived under the state weakness that characterized a war-torn country like Colombia, with limited ability to reach all corners of national territory (Thoumi 1995), drug trafficking networks did not simply emerge where the state was absent. Especially in the urban areas where early DTOs located their operational bases, the state was present, even if fragmented. 3.2.2.The aftermath of the big “Cartels” 1998-2006 Three main changes marked the mid 1990s: the big trafficking organizations had been dismantled; guerrillas and self defense groups became more actively engaged in cocaine production and trafficking; and the security apparatus became more sophisticated, coordinated, and cohesive. At a first sight, anti trafficking operations seemed to be successful as the threat of narcoterrorist violence disappeared, and big trafficking organizations were fragmented into many small organizations, colloquially called “microcartelitos” (Bagley 2012; Gootenberg 2012; Lopez 2006). In practice,   116   however, the situation was far more complex: trafficking networks had been beheaded but not completely eliminated, and the successors of these networks engaged in a set of complicated interactions with guerrillas, and especially, paramilitaries. Rather than pure fragmentation, there was a process of re-accommodation that fueled an intensification of the armed conflict, associated with the expansion of paramilitary groups and their attempt at winning territorial control over guerrillas. In the mid 1990s, Colombia’s main guerrilla group, the Revolutionary Armed Forces of Colombia (FARC), increased significantly its involvement in the taxing of coca crops, thus leading both domestic and international voices to emphasize the narco- guerrilla thesis, that is, the idea that the guerrilla was no longer motivated by ideological grievances but by pure greed. As many excellent analysis of this topic have demonstrated, the FARC found in drug trafficking not only a source of funding, but also a mechanism of social control and for the creation of support bases among peasants and coca growers (Felbab Brown 2010; Gutiérrez 2004). The increasing involvement of the guerrilla in coca cultivation provided the motivation that the Army needed to fully engage in anti trafficking operations. Coordination between the Army and the Police increased, thus reducing a source of fragmentation in the security apparatus. A military colonel who was deployed in the 1990s as commander of an elite unit in charge of dismantling the Cali DTO summarized the changing relation between military and police as follows “[in the 1980s] the relation with the police was respectful but distant […] institutional distrust was a clear obstacle […] Nowadays the relationship is more fluid and it is easier to work together, perhaps   117   because, when you introduced guerrilla to the narco equation it became more difficult for the military to be bribed.79” In 1999 during the Presidency of Andres Pastrana (1994-1998), the United States government launched Plan Colombia, an antinarcotics and antiterrorism plan that consolidated the merging of antinarcotics and antiguerrilla operations and thus provided an additional opportunity to promote cooperation between military and police forces. The Plan Colombia increased the funding, technical training, and mobility of armed forces throughout the territory. The military component of the plan80 totaled US$ 3,378 million provided by Colombia and US$ 2,787 million provided by the United States (Mejia 2010). The plan focused on two goals, reducing cocaine production by 50% in six years, and restoring security in the country (Mejia 2010). Implicitly, the plan gave more preeminence to attacking guerrillas and greatly ignored the expansion of paramilitary groups, thus contributing indirectly, to their consolidation. Paramilitaries and the remnants of older drug trafficking groups benefited from the dual antinarcotics- antiguerrilla campaign. Many of the remnants of the old DTOs allied in diverse and complex ways with paramilitary groups. Some trafficking groups had, in fact, remained strong, as was the case with the Norte del Valle DTO, successors of the Cali DTO, who were far from being simply a “microcartelito.” As we will see in Chapter 5, the Norte del Valle DTO engaged many prominent traffickers and controlled a large part of the                                                                                                                 79 Author’s interview with former Army Colonel and former commander of the Search Bloc against the Cali DTO. Bogotá, 2010. 80 There was also a social development component of the plan, but it was small compared with the resources and political prominence given to the military component.   118   trafficking chain. In the words of an Army general “Between 1995 and 1998 the main problem started to be the guerrilla, and the Norte del Valle cartel benefited from that.”81 Authorities from the US and Colombia attribute to Plan Colombia the reduction of coca cultivation and the strengthening of military presence that occurred in the late 1990s and early 2000s. Yet, the implementation of the plan also coincided with some of the worst Human Rights violations and with the expansion of paramilitary forces. This was crucial because the re-accommodation of forces in the armed conflict allowed the paramilitaries to become the sole controller of many territories and consequently of criminal markets,82 with important consequences for the levels of violence in the country. 3.2.2.1. Alvaro Uribe’s presidency and the changing role of paramilitaries One essential aspect in the consolidation, expansion, and centralization of military power was the election of right wing politician Alvaro Uribe as President in 2002. Uribe got elected on an agenda of all-out war against guerrillas that reflected, among other things, the citizen’s disappointment with a failed peace process with the FARC that former President Pastrana carried out and that contributed to the expansion of FARC’s power in the southern part of the country. Uribe expanded the territorial reach and capacity of both the military and the police. Military investment increased under a far-reaching plan aimed at gaining                                                                                                                 81 Author’s interview with former Army Colonel and former commander of the Search Bloc against the Cali DTO. Bogotá, 2010. 82 It is crucial to note that even after the creation the AUC (United Self Defenses of Colombia) in 1997, an umbrella paramilitary group that joined previously regionalized and fragmented paramilitary groups throughout the country, the paramilitaries kept being a relatively decentralized organization, constantly subject to fragmenting pressures (Gutierrez and Baron 2006). In this sense, international drug trafficking was not in the hands of a completely centralized and monopolistic group. But it was not in the hands of completely uncoordinated “cartelitos” either.   119   territorial control known as the Democratic Security Policy (Politica de Seguridad Democrática). Military forces grew from 145,000 personnel in the late 1990s to 431,253 in January 2009 (Granada et al 2009). In 2003, the government also established Joint Command Forces (Comandos Conjuntos) that improved the operational coordination of different state forces (Granada et al 2009, p.81). Uribe’s plan was effective in reducing kidnappings, deaths of civilians and combatants, and in weakening the guerrillas (Granada et al 2009, p.51). Both political power and enforcement coordination concentrated around the President, and thus the state became more cohesive. However, paramilitary groups, now the protagonists in the international distribution, the most profitable part of the drug trafficking chain (Duncan 2005), consolidated their relations with wide sectors of the political establishment in temporary protection rackets (Snyder and Duran-Martinez 2009b). In 2006, the Supreme Court of Justice and the Attorney General’s office initiated the investigation of links between politicians and paramilitaries through a judicial process known as “Parapolitica”. The proceedings of this large-scale process that by 2012 had investigated 943 politicians including 139 legislators, 41 mayors, and 20 Governors,83 unveiled how paramilitary forces established symbiotic relations with large sectors of the political class. Probably for the first time in the history of drug trafficking in Colombia, the state achieved a level of centralization that facilitated the predictable protection to paramilitary groups, by then as much political actors as drug traffickers. The tradeoff between state autonomy from criminal influence and state coercive capacity emerged again: more security meant less                                                                                                                 83 Verdad Abierta 2012. “Cinco años de parapolítica.” Available at: http://www.verdadabierta.com/Especiales/cinco-anios-parapol/parapolitica-2012.swf Accesed [3 October 2012].   120   autonomy, even though police and military were significantly more professionalized than in the 1980s.   Paramilitary groups entered a peace negotiation with the government in 2002 that led to the demobilization of approximately 32,000 combatants (CNRR 2007). Human rights groups and other voices criticized the process, because it was not framed within a clear legal framework that guaranteed transparency about the terms of the negotiation, and justice and reparation for victims of crimes against humanity (Romero 2007, OHCHR 2006). Once the legal framework for the negotiation was created, there were also concerns about the possibility of impunity (OHCHR 2006, Uprimny and Saffon 2006). While an analysis of the many shortcomings and the advances associated with this process exceeds the purpose of this chapter and this dissertation,84 the crucial point is that the consolidation of paramilitary forces, their successful entrance into drug distribution, and the possibilities of protection provided both by the peace process and by the networks of political protection, created the appropriate context for drug violence to become both less visible and less frequent at the national level. It is crucial to note here that a subnational perspective provides us with a more nuanced story about the impact of paramilitary demobilization and reconfiguration on the levels of violence. Paramilitaries successfully merged with trafficking actors in Medellín, as described in Chapter 4, and thus contributed to the creation of a criminal monopoly and a reduction in both the frequency and the visibility of violence. Yet, in Cali trafficking did not merge successfully with paramilitary actors, and as a result, violence remained very frequent in Cali even after the paramilitary consolidation.                                                                                                                 84 One positive advance of the process, although limited by the above-mentioned problems, is the conformation of a Truth Commission that produced reports about armed actors in Colombia. Some of these reports were especially useful for clarifying the links between drug trafficking and paramilitarism.   121   As we will see in the next section, a combination of new trafficking actors, non- demobilized paramilitary forces, recidivate soldiers, and older criminals became the center of the sharply different dynamics of drug trafficking in the new millennium. 3.2.3. Drug trafficking in the 21st century (2006-2012) Towards the end of the first decade of the 21st century, the dynamics of drug trafficking had changed radically in Colombia. Statistics on both coca cultivation and cocaine production declined, while evidence suggesting the greater power of Mexican trafficking organizations increased. Although all events point to a declining power of Colombian traffickers, it seems more difficult to claim that drug trafficking is no longer important. In many ways, the institutional transformations described in previous sections led traffickers to adapt, becoming less open in their operations and more careful in avoiding media and enforcement attention. Many of them became very localized “war- lords” and criminal networks seem to have more localized roots that are unlikely to pose a “national security” threat. Yet criminals retain destabilizing power,85 links with sectors of security forces (OHCHR 2011), and links with political and electoral actors (Avila and Velasco 2012). The peace process with paramilitary groups dismantled the national network of paramilitary groups established in the late 1990s, the United Self Defenses of Colombia (AUC). Yet, remnants of paramilitary structures, especially middle ranks that did not demobilize and rearmed demobilized members, along with old and new criminals,                                                                                                                 85  Garzon, J.C. 2013. “La violencia que seremos” Razón Pública, 10 February. Available at http://www.razonpublica.com/index.php/conflicto-drogas-y-paz-temas-30/3556-la-violencia-que- seremos.html, Accessed [10 February 2013]   122   merged into new criminal structures. Some experts and academics defined these armed groups as neo-paramilitaries or new illegal armed groups, while the government designated them as criminal gangs (Bacrim) to deemphasize the idea of a connection with former paramilitary structures. Assessments about the nature and size of new groups varied widely depending on the source. By 2012 the reported geographic presence of these groups ranged from 4,800 members with presence in 167 municipalities, to 8,000 members with presence in 406 municipalities (ICG 2012). Thus, these groups, although relatively dispersed throughout the territory and less able to challenge national security, acquired significant geographic presence and became crucial actors both in the armed conflict and in illegal markets. The new criminal structures diversified their illegal portfolios and were more likely to engage in other illicit and semi licit activities, such as illegal gold mining, gambling, and extortion (ICG 2012, Indepaz 2012).86 Alliances between groups that appeared to be enemies before, such as guerrillas and neoparamilitary groups also became more common (CNRR 2007). However, these associations were not entirely new, and while alliances existed in some places, in others there was also dispute for territorial control (ICG 2012, Idler 2012, Indepaz 2012). It is crucial to note that diversification of illegal activities was not new; in Medellín, as we will see in chapter 4, extortion and local drug markets were widespread in the 1980s. The difference with the new scenario was that in the 1980s trafficking groups rarely attempted to control other illegal activities. These illegal activities became more consequential for criminal income, thus reflecting some decline in the power of drug trafficking networks.                                                                                                                 86 This economic diversification sometimes even extended to investment in legal activities such as oil exploitation or palm cultivation.   123   Yet a decline in trafficking power was far from meaning disappearance, and drug trafficking remained crucial as an income source for illegal actors, and as a security issue. Groups operated as security providers for owners of cocaine processing facilities, rather than as controllers of international distribution.87 But even in this context, criminal groups supplied cocaine for more powerful Mexican traffickers, and played crucial roles in cocaine trafficking to Europe. Even though reductions in cocaine production and coca cultivation in Colombia were offset by increases in Peru and Bolivia88, Colombia still played a major role in supplying coca and cocaine. Although statistics are highly controversial, according to the most recently available comparable estimates from the UNODC, in 2010 Colombia produced as much coca as Peru, and more than twice the amount produced in Bolivia.89 3.2.3.1. Criminal adaptation and silent traffickers After almost four decades of interaction between criminal groups and state actors in Colombia, traffickers have learned to lower their profiles and thus reduce the                                                                                                                 87 This is the case of a structure known as the ERPAC (Popular Revolutionary Anti-insurgent Army of Colombia) that rather than directly controlling production or international distribution, acted as a provider of security and overseer of cocaine processing facilities (ICG 2012). 88 This is an illustration of the “balloon effect” of enforcement: as enforcement pressure in Colombia has increased, coca production has moved back to other locations. 89 An important controversy about numbers emerged in 2012. While the UNODC reported a slight increase in hectares of coca cultivated in Colombia from 62,000 to 64,000 and a minor decrease in coca production of about 1% from 350 to 345 tons, the US reported a decrease in cocaine production in Colombia of 25% from 270 to 195 metric tons of cocaine, falling behind Bolivia and Peru (UNODC 2012; ONDCP 2012). Besides the wild divergences between the two sources, the US report was criticized because it was not clear how Colombia produced less cocaine that Bolivia even though it still had more hectares of coca cultivated. For an analysis of these numbers see: WOLA, 2012. “UN and US cocaine estimates contradict each other”. July 31. Available at: http://www.wola.org/commentary/un_and_us_estimates_for_cocaine_production_contradict_each_other Accessed [16 February2013] and Fox, E, 2012, “US report on Colombia Cocaine raises more questions than answers” Insight Crime , August 1, Available at http://www.insightcrime.org/news-analysis/us-report- on-colombia-cocaine-production-raises-more-questions-than-answers Accessed [16 February 2013]   124   possibilities of being captured by enforcement agents. An intelligence officer expressed to me “traffickers are no longer flashy [dan pantalla]. You don’t see many flashy traffickers [traquetico] but they extort people and threaten. It’s a very strong phenomenon […]. Trafficking is a time bomb, not like Mexico, but the BACRIM [criminal gangs] are uncontrollable, this hidden criminality has a very strong capacity to penetrate [state institutions].”90 The intelligence officer’s remarks about the criminals’ capacity to penetrate institutions summarize the consequences of the institutional transformations during the past four decades. On the one hand, it is clear that as a result of the efforts to address the challenges of a persistent armed conflict and drug trafficking, military and police forces are more sophisticated and have more territorial reach than in the 1980s. During the period of President Uribe these changes occurred in tandem with serious distortions in electoral competition, but the moderation of some of these distortions during Santos’ term (2006-2012) reflect that to some extent Colombia has been able to reduce the tradeoff between democracy and state coercive capacity. However, on the other hand, crucial challenges still diminish the state capacity to enforce the law especially at the local level. The growth and power of the security apparatus could be testified to by the enforcement officials’ confidence in their own operational capacity, which stands in sharp contrast to the fears and distrust in the state apparatus of the 1980s. In the words of Oscar Naranjo, the director of the National Police between 2007 and 2012: “Police forces in the world are threatened by four monsters: inefficiency; police brutality, lack of                                                                                                                 90 Interview with intelligence official of the General Attorney’s Office, September 13, 2010. Intelligence officials of the national police may be less likely to recognize the destabilizing power of criminal structures in the new era of drug trafficking.   125   solidarity, and corruption. We have combated all of them. Today the penetration of corruption is minimal. Today, drug traffickers do not feel at peace in Colombia, although it does not mean that there is not much left to do.”91 Naranjo’s remarks reflect the power and legitimacy that police forces have gained in the past decades, and their own self- confidence. The security apparatus in Colombia has grown more complex, with multiple specialized units, and the power of the police represents a node of cohesion in the security apparatus. Institutional jealousies exist, and as many of the cops and intelligence officials I interviewed put it, there is always “confrontation” about “who will appear in the picture” or in other words, about who will take the credit for a particular enforcement operation, capture, or seizure. However, it is clear that the police is the leading agency in antinarcotics operations, and it is the agency with the most comprehensive territorial reach and the most resources. As one police commander put it “Institutions work alone, and the one that has more power and reach is the police. With the CTI (intelligence police of the Attorney General’s office) we divide cases, but they are inefficient. The DAS (extinct civilian intelligence agency) is weakened […]. The Army is not that efficient, because they want to use their spirit d’ corpse to hide things. The Army has many scandals […]   But   because of hypocrisy and because of the power of the President, forces must be united.”92     The sophistication and territorial reach of law enforcement agencies has modified the incentives of criminals to use highly visible forms of violence. Yet, there are still huge                                                                                                                 91 Amat, Y. 2012. “'Debe permitirse el consumo controlado de marihuana': general Naranjo”. El Tiempo. April 21. Interview with General Oscar Naranjo. 92 Author’s interview with Police Commander, Bogotá, August 30 and September 5, 2011.   126   challenges for the state’s coercive and law enforcement capacity. At the local level state capacity is still fragmented in some cases, or symbiotic with criminal actors in others. In Cali violence does not recede but is not very visible because state actors protect criminals. In fact, in 2012, Cali appeared as the 7th most violent city in the world93 but this situation did not capture much media attention. In Medellín, a modern and progressive political coalition is in government, but local police forces still collaborate with the local gangs (combos). Even though these corruption networks are not necessarily linked to the highest rankings of police authority, they greatly affect the well-being of local communities, as expressed by a young inhabitant of a highly violent neighborhood in Medellín “There have been some corruption cases, but those that get discovered are the unlucky ones. But the cops [tombos] here are the ones that handle the black market in arms, the robberies, the drugs. The Police is highly corrupt.” 94 Other areas in the country face extremely dangerous security situations associated with drug trafficking. This is the case in areas of the Pacific coast of Colombia, in the departments of Choco95 and Valle del Cauca. For example, since 2006 the port city of Buenaventura was besieged by an increasingly brutal but silent violence, derived from the confrontation between several neo-paramilitary and criminal groups such as La                                                                                                                 93 “San Pedro Sula otra vez el primer lugar a nivel mundial; Acapulco la segunda”. Seguridad, Justicia y Paz Consejo Ciudadano para la Seguridad Pública y Justicia Penal A.C” Available at http://www.seguridadjusticiaypaz.org.mx/sala-de-prensa [Accessed 12 February 2013]. 94 Focus group conducted in Medellin with Arturo Alvarado, Alberto Concha-Eastman, and the author in the framework of the project “La violencia que afecta a los jóvenes en América Latina”. 95 El Tiempo. 2013. “Son 903 los desplazados por bandas en el litoral del San Juan, Chocó” El Tiempo, January 17.   127   Empresa, Los Urabeños, Los Machos, and Los Rastrojos.96 Thus, even if they do not appear as a national security threat, current criminal organizations do have a large potential to destabilize public security in many regions. These local enforcement challenges sometimes tend to be overlooked in the optimist atmosphere of declining violent trends in the country. In sum, the Colombian security apparatus has acquired abilities to react more quickly to criminals, especially those that concentrate too much power. Yet, enforcement capacities are still hindered by local security threats and by weak judicial systems. Colombia has good functioning higher courts (Brinks 2012) that have been critical in investigating large corruption scandals that connect wide sectors of the political class with illegal armed and trafficking actors. Yet, impunity rates for crimes are very high and many trafficking actors have managed to get light sentences,97 a situation that also occurs in the United States.98 Colombia has improved its enforcement capacities in the context of a democratic state but the situation is still precarious in several aspects. One of the key lessons of the framework proposed in this dissertation is that one cannot automatically infer the conditions of drug trafficking organizations from the level of violence in a given location. In other words, less violence does not necessarily mean weaker organizations or more capable state enforcement. In this sense, even though                                                                                                                 96 Molano, A. 2013. “Buenaventura: entre la pobreza y la violencia”. El Espectador, February 23. Interestingly, Victor Patiño, an old time trafficker from the Norte del Valle DTO, commands the dominant group Los Rastrojos, reflecting the complex links between old and new traffickers. 97 A case in point is Lorena Henao, a member of the Norte del Valle DTO, who was captured in 2004 and released shortly afterwards in 2011 after completing her sentence. El Tiempo. 2012. “Quien era Lorena Henao Montoya la viuda de la mafia?” El Tiempo, December 27. 98 In the United States the practice of reducing sentences in exchange of information has led many extradited criminals to get light sentences such as in the case of former members of the Cali DTO William Rodriguez, Phanor Arizabaleta or Victor Patiño Fomeque who obtained 4 years, 9 months, and 8 years in prison respectively. See Semana. 2012. “Fracaso la extradición de narcotraficantes?” Semana April 14.   128   Colombian enforcement agencies have increased their capacity to control territories and establish the monopoly of force, trafficking organizations are not necessarily powerless, because in many cases they have adapted to enforcement changes by reducing visible violence. Additionally, state capacity is not evenly distributed throughout the territory and does not imply the elimination of criminal actors. Even if weaker, more territorially fragmented, and more diversified than in previous periods, trafficking organizations still retain capacities for violence, and still derive significant rents from illegal activities. The persistence of high levels of violence and corruption suggests that there is still a long road to consolidate both the state capacity to enforce the law and its autonomy from criminal influences. 3.3. Mexico: From state sponsored protection to the perils of incomplete democratization 3.3.1.The origins of drug trafficking and the PRI rule Drug trafficking in Mexico, as Luis Astorga (2005) has documented, has an older history than in Colombia. Opium and marijuana production date back to the late 19th century, when these products were cultivated in small amounts to satisfy a demand of drugs for medicinal purposes, and a growing, and increasingly criticized use for social recreation. As worldwide efforts to control drugs started to intensify in 1909 with the Shanghai opium convention, the 1912 Hague Convention, and the US 1914 Harrison Act, the Mexican government also increased the pressure on the use and production of drugs. In the 1940s two important transformations occurred. First, in 1947, narcotic drugs started to be treated as a national security issue, and the responsibility for antinarcotics policies was centralized in the Attorney General’s Office (PGR). Second,   129   global demand for marijuana, and specially opium, increased during and after World War II, and traditional sources of supply lost preeminence, opening the space for Mexico to become an important opium and marijuana supplier. Illegal crops proliferated, and the state of Sinaloa, in the pacific coast of Mexico became a new node for trafficking organizations (Astorga 2005). Unlike Colombia, drug trafficking in Mexico emerged in the context of a system where political power was centralized after the decades-long rule of the Revolutionary Institutional Party (PRI). Drug trafficking consolidated clearly tied to the political power and ability of the PRI to determine terms of interaction between traffickers and the state (Astorga 2005). Drug trafficking was initially concentrated on opium and marijuana, although Mexicans had early incursions in cocaine trafficking in the 1960s (Gootenberg 2009, p. 273-275). Marijuana became particularly important as Mexican traffickers created a seedless variety (“sinsemilla”) that had stronger psychoactive potential. In 1977, and with the support of the United States government, the military carried out the “Operación Cóndor” aimed at eliminating marijuana production. Although the Operation Cóndor did not effectively eliminate trafficking groups, it caused a dispersion of trafficking networks outside the state of Sinaloa to other Mexican territories. As we will see in Chapter 5, this dispersion contributed to generate spikes of violence between trafficking groups in some Mexican states, and even some violent reactions against authorities, especially PGR officials, as traffickers attempted to resist the eradication operations (Astorga 2005, p. 101). Yet, for the most part the interaction between traffickers and state officials remained stable and not violent.   130   In the 1980s US authorities cracked down on cocaine transportation routes in the Caribbean, and Colombian traffickers started to use transportation routes inland through Mexico, thus increasing the incidence of Mexicans in cocaine trafficking (Andreas 2000, p.45-53). Initially, Colombians paid Mexican traffickers a fee for their help in smuggling the drugs, but over time Mexicans started to be paid in kind, or with a percentage of the profits, thus changing the nature of Mexican organizations. By the mid 1980s, drug trafficking had become a great concern for national and international audiences, and the cases analyzed in this dissertation focus on the transformations occurring after this juncture when trafficking emerged more clearly in the public scene. Some analysts argue that as a result of their greater involvement in cocaine trafficking, Mexican organizations became more powerful and increasingly willing to use violence. Yet, as we will see in detail in chapters 5 and 6, there was variation in the level of violence across cities where cocaine traffickers operated. Additionally, violence did not increase immediately after cocaine trafficking became more important. One case in point is trafficker Amado Carrillo and the Juárez DTO, who since the early 1990s mastered relations with Colombian traffickers. Carrillo was one of the traffickers who first requested from Colombian traffickers a direct share in the profits of trafficking. Yet, despite his prominent role in cocaine trafficking, Carrillo favored non-visible violence strategies. Allegedly, he even offered deals to the Mexican military: in exchange for the liberty to carry out his business he offered to control disorganized traffickers, to ban drug sells within Mexico, and to limit violence (Astorga 2005, p.169).   131   During the Sexenio99 of President Miguel de la Madrid (1982-1988) drug trafficking was declared a national security priority and as a result, enforcement agencies engaged more actively in the policing of drug trafficking. The Attorney General’s Office (PGR) remained as the central institution of the law enforcement and antinarcotic apparatus. Although multiple other agencies existed, the power of the PGR in controlling antinarcotics enforcement was uncontested. Within the PGR, police units such as the Federal Security Directorate (DFS) became the nodes of protection and interaction between the state and criminal actors (Lupsha 1991; Snyder and Duran-Martinez 2009a). The military was subordinated to the power of the PGR, and the PGR concentrated both policing and judicial responsibilities. 3.3.1.1. The Camarena case and its implications In November 1984 the Mexican Army, in collaboration with the DEA, discovered and dismantled the Buffalo Ranch, a large marijuana plantation in the north of Mexico, seizing thousands of tons of marijuana.100 A few months later, on February 7 1985, and in retaliation to this operation, Mexican trafficker Rafael Caro Quintero in association with trafficker Miguel Angel Felix Gallardo, kidnapped DEA agent Enrique Camarena and his Mexican pilot Alfredo Zavala, who were later found dead on March 6. Camarena’s death unleashed an aggressive response from the United States. Days after the kidnapping, the DEA declared that 30 of its agents worked in Mexico, that Guadalajara –Caro Quintero’s turf- was the epicenter of national and international drug trafficking, and that 75 major                                                                                                                 99 Sexenio refers to the six-year term of Mexican Presidents. 100 Paradoxically, military personnel of the garrison of the small town of Jimenez protected the plantation (Artz 2011).   132   drug traffickers controlled the drug business in the country, in many cases protected by state authorities. These assertions generated discomfort among Mexican authorities. PGR representatives denied that Mexico was a launch pad for international drug trafficking, and announced that only seven traffickers operated in the country, and that there was no police corruption (Astorga 2005, p.134). Tensions between US and Mexican authorities followed the Camarena case, with US authorities accusing Mexicans of corruption and Mexicans accusing US authorities of intervening in the country. The episode of visible violence of the Camarena case forced the Mexican state to respond to traffickers and to international pressure, and in 1985 Mexican authorities captured Rafael Caro Quintero and dismantled the DFS101 on the basis of the pervasive corruption that affected it. However, the initial reluctance of Mexican authorities to recognize the incidence of drug trafficking and corruption, and the ability to quickly conduct operations that responded to the pressure of the events, testified to the fact that the state had the ability to regulate criminal interactions, and that it protected criminals to the extent possible. Despite the scandal and the dismantling of the DFS, there were no major anticorruption reforms that transformed the structure of enforcement after Camarena’s death. The Camarena incident illustrates that the state protection to criminals that may emerge when a state is cohesive, as occurred during the PRI rule in Mexico, does not necessarily imply that enforcement actions cannot occur, or that criminals cannot retaliate to those actions. The key is rather that under conditions of state cohesion, the spillover                                                                                                                 101 In an illustration of the corruption of the DFS, Caro Quintero carried a DFS badge.   133   effects of both state actions and violent criminal retaliation can be easier to control, at least in the short run, than when the state is fragmented. In sum, up to the mid 1980s, Mexico illustrated the tradeoff between a state’s coercive and enforcement capacity and a state’s autonomy from criminal influence. The concentration of state power in the PRI and of enforcement responsibilities in the PGR allowed the state to maintain the monopoly of force, even with very corrupt and non- professional police forces. At the same time, the formal and informal links with trafficking organizations were very strong. Even though many Mexican analysts emphasize that it was the PRI who held the strong hand in dominating criminals, the key is that in order to control traffickers the state conceded to them benefits that empowered them, and thus the state was not an autonomous actor. Criminals were not necessarily weak, but the state’s capacity to coordinate its actions against them made it too costly for criminals to confront the state, as reflected in Caro Quintero’s quick capture after Camarena’s death and the lack of a strong retaliatory response from criminals after his capture. It is crucial to note that the centralization of political power effectively motivated traffickers to regulate the use of, but did not completely eliminate, violence. Subnational variation existed in the behavior of traffickers: as we will see in Chapters 5 and 6, in the 1980s while Culiacán experienced extreme levels of violence that were usually above the national average, Ciudad Juárez and Tijuana had lower levels of homicide, at least up to the mid 1990s. It is even more interesting to note that the higher levels of violence in Culiacán occurred precisely in a PRI dominated state, thus showing that state cohesion was consequential mainly for the visibility of violence. Along the lines of the framework   134   I propose in the dissertation, the PRI domination did not necessarily guarantee the absence of violence that could derive from intra and inter organizational disputes, but rather determined its low visibility. 3.3.2. Democratization and the initial transformations in the state security apparatus (1988-2000) President Carlos Salinas de Gortari was sworn in as President of Mexico in 1988, amidst accusations of fraud in the election. Drug trafficking had become more public and accusations of corruption and state protection proliferated, forcing the government to carry out more aggressive institutional and anticorruption reforms. In 1989 the Federal Judicial Police PJF captured trafficker Miguel Angel Felix Gallardo, who was at the time the most powerful trafficker in Mexico. Upon his incarceration Felix Gallardo divided the control of drug trafficking along regional lines that separated the traffickers from Sinaloa led by Joaquin Guzman and Ismael Zambada, Baja California led by the Arellano Felix organization, and Chihuahua led by Amado Carrillo. Although Felix Gallardo retained command over these groups for some time, by the early 1990s, confrontations between the different regional trafficking groups increased, thus changing the geography of trafficking in Mexico. During his Presidency, Salinas dealt with high level assassinations such as that of Cardinal Jesus Posadas Ocampo in 1993, presidential candidate Luis Donaldo Colosio in March 1994, and general PRI Secretary Jose Francisco Ruiz Massieu in September 1994. The cardinal’s assassination was allegedly an accident that left the cardinal between the cross fire of members of the Arellano Felix organization and their rival trafficker Joaquin Guzman. In the other two cases, the circumstances were never completely clarified, even   135   though the Ruiz Massieu case led to the imprisonment of the President’s brother Raul Salinas, and both cases seemed to weave political and drug trafficking interests in complex ways. These events and the general sense of corruption within the PGR led Salinas to a constant rotation among the heads of the PGR (Artz 2011) and each new Attorney arrived with a new team of collaborators. These rotations altered the power interactions within the PGR, and along with the dismantling of the old police forces such as the DFS, changed the distribution of enforcement power. However, the PGR remained as the central actor in law enforcement and the newly created police agencies eventually became plagued by corruption, as it is the case with the Federal Judicial Police (PJF). The PJF became the most powerful force within the PGR, and as its predecessor the DFS, also became the central node for criminal-state interactions. Throughout the 1980s the main opposition party to the PRI, the PAN (National Action Party) won mayoral offices in a few municipalities in the country, especially in the northern states of Baja California, Chihuahua and Sonora. In 1989, the slow opening of electoral competition took on a new dimension when the PAN won for the first time a governorship in the state of Baja California. The electoral success in Baja California, was followed by other successes at the state level in Chihuahua (1992), Guanajuato (1991), and Nuevo Leon (1997), and by the first opposition dominated legislature in 1997. These opposition successes created a more heterogeneous political landscape and curtailed the power of the federal government and the PRI. However, as we will see in chapter 6, even though these gubernatorial elections represented a watershed in the process of democratization in Mexico, in the short run they did not have a large impact on the   136   structures of intermediation between state officials and traffickers. Crucial processes of transformation were accumulating but the old power structure was still very ingrained. The last year of Salinas’ term was marked by the emergence of the infamous EZLN (National Liberation Zapatista Army), which led an armed uprising in the southern state of Chiapas that shook up the country precisely the same day in which the North American Free Trade Agreement NAFTA went into effect. Thus, when President Ernesto Zedillo (1994-2000) initiated his term in December 1994 the legitimacy of the regime was questioned, and the opposition PAN was gaining more electoral power. Zedillo gave more concessions to the opposition, and one of the most significant ones was to name for the first time a PAN official as Attorney General in 1994, thus breaking down a traditional pattern of PRI leadership in this agency. Antonio Lozano, the PAN Attorney General promoted more reforms in the structure of the PGR in an attempt to decentralize its power and to reduce corruption, although his tenure was very brief.102 One reform that illustrates the attempted transformations in the PGR was designed by Lozano but implemented by his successor Antonio Madrazo, and changed the geographic organization of the PGR in 1997. The reform replaced the centralized, functionally based division of power of the PGR, with a decentralized, geographically based division of power, that created three jurisdictions (subprocuradurias) joining non- contiguous states. The logic behind the reform was that in contiguous jurisdictions corruption could trickle up, whereas a dispersion of jurisdictions could curtail corruption.103 In fact, the new organization led to a disjuncture between geographies of                                                                                                                 102 This pattern resumed after PAN Attorney Antonio Lozano left in 1996, but was finally broken with the election of a PAN President in 2000. Another significant change in the PGR was that since 1995 the appointment of the Attorney General by the President is ratified by the Senate. 103 Author’s interview with former high-level official of the PGR. Mexico City, September 26 2011.   137   enforcement and geographies of criminality (Snyder and Duran-Martinez 2009a), multiplying the number of enforcement actors and making it more difficult for them to coordinate. The reform increased fragmentation although it was short-lived, precisely because it generated problems of coordination and sharp divergences in the way in which each subprocuraduria conducted judicial processes and evaluated its own security achievements.104 In 2002, the organization of the PGR changed again to one combining deconcentration and specialization, by strengthening the delegations (Delegaciones) of the PGR in each state, and at the same time creating specialized subprocuradurias for specific crimes such as organized crime, electoral crimes, and piracy. These changes of the late 1990s did not eliminate but rather fragmented the power of the PGR. At the national level, violence displayed a downward trend, although at the local level some of institutional changes affected the terms of interaction between criminals and state, as occurred in Tijuana, where some anticorruption reforms and enforcement operations generated visible violent reactions by the Arellano Felix trafficking group. Attorney Jorge Madrazo continued with the transformation pattern of the PGR dismantling the National Institute for Combating Drugs (INCD) and the Center for Drug Control Planning (CENDRO), which had been created in 1993. He created two new specialized units, the FEADS (Specialized attorney office for crimes against health) and the UEDO (Specialized Unit for organized crime) in April 1997. The decision came after General Jesus Gutierrez Rebollo, who was the head of the INCD, was investigated and sentenced to prison for collaborating with trafficker Amado Carrillo. This pattern of                                                                                                                 104 Ibid   138   constant creation and recreation of offices added instability to the PGR without solving its core problems, but the PGR was still the cornerstone of the security apparatus. Throughout the mid 1990s, the military increasingly engaged in public security tasks. This process, rather than a demand of the military establishment, was a strategy of governments to contain increased challenges of public security; as in Colombia in the 1980s, governments thought the military could countermine the weakness and corruption of police forces (Artz 2011). Attorney Lozano, for instance, named a military member as the director of the Judicial Police PJF and involved the military in antinarcotics operations. Yet, for the most part, military commands remained formally subordinated to the PGR. This “behind the scenes” militarization illustrates that despite the initial and important transformations in the PGR, the security apparatus was still mostly cohesive. In the late 1990s, a key change in the structure of enforcement started to pave the way for a reduction of the PGR’s role in public security: a renewed role of the Secretary of Governance (Secretaria de Gobernación SEGOB) in public security. In 1989 Zedillo’s Government created the civilian intelligence agency CISEN attached to the SEGOB, and in 1995 established a Law to create a System of National Public Security (Artz 2011). Zedillo also created the Federal Preventive Police (PFP) in 1999, which unified the former Federal Road Police (Policia Federal de Caminos), the fiscal police, and some personnel from CISEN, and included members of the military police. The creation of the PFP generated criticisms among sectors that feared excessive intervention of the federal government in municipalities and states, and feared duplication of the roles of the PGR and the newly created police.   139   3.3.2.1 Protection starts to break down During Zedillo’s administration, former President Salina’s brother, Raul Salinas was accused of laundering money for the Gulf DTO, and this scandal affected the configuration of drug trafficking groups in Mexico. The Gulf Drug Trafficking Organization led by Juan Garcia Abrego had become more powerful during the Salinas sexenio, and thus became completely independent from the Sinaloan traffickers (Astorga 2005). There were serious suspicions that the Gulf DTO enjoyed protection from the highest echelons of power. Partially the public pressure generated by these suspicions, led to Raul Salinas’ indictment on money laundering charges105 and to the detention of Garcia Abrego in 1996. Abregos’ successor, Osiel Cardenas, seemed resolved to taking over his rivals more violently. In 1997 Cardenas created a semi-paramilitary armed branch, the Zetas. The emergence of the Zetas reflected the concurrence of two processes: on one hand, the capture of Garcia Abrego, after being protected by high ranking officials, signaled that state protection was becoming less predictable, and created an incentive for Cardenas to create an armed apparatus that could help him in the quest for increasing territorial power. On the other hand, starting in 1997, military skills became available as an important number of well-trained soldiers deserted from the GAFES, special military units created in 1994 with the aim of maintaining national security in the midst of the EZLN uprising. Ex-kaibiles and former Guatemalan soldiers supplied additional military skills. The kaibiles, as the GAFES, were soldiers trained for special anti-guerrilla                                                                                                                 105 And also for being the intellectual author of the assassination of his brother-in-law Jose Francisco Ruiz Massieu.   140   operations, and with equal eagerness to move to the side of the most profitable employer.106 3.3.2.2. A system in the brink of transformation Despite all the institutional reforms and the transformations in the criminal market, in the end the Zedillo administration put more emphasis on economic liberalization, the consolidation of NAFTA, and on tackling the political consequences of the Zapatista uprising, than on making organized crime and drug trafficking a policy priority (Benitez Manaut 2010). Many of the new trafficking challenges were kept under the radar; both the US and Mexican governments shared an interest in the success of NAFTA and as a result, the public reactions to drug related corruption and violence were somehow contained. For example, the US government did not decertify Mexico in this period even though corruption scandals led to Colombia’s decertification over the same period. The corruption scandals and political assassinations forced the Zedillo government to restructure security practices (Bayley and Chabat 2002) but these changes were still contained. By the end of Zedillo’s term, many changes were brewing, and criminals seemed to recognize that protection was not as predictable as it was before, as this is probably one of the reasons for the creation of Los Zetas. Yet criminals still seemed reluctant to confront the state or to deploy full-scale violence. In fact, it was not until 2003 that the Zetas became visible for the media. With the election of a PAN                                                                                                                 106 It is estimated that between 2000 and 2004, 61000 members of the GAFES and GANFES deserted the military for their “lack of adaptation to the military environment”. El Universal. 2004. “Desertan 1382 militares de elite”. March 28. La Crónica de Hoy. 2005.“Los Zetas fueron capacitados por los kaibiles antes de desertar del Ejército mexicano”. La Crónica de Hoy. October 3.   141   candidate as President of Mexico in 2000, the process of democratic transition was completed, and in this context, politicians and traffickers’ incentives changed radically. 3.3.3.The end of the PRI hegemony During the 2000s, transformations in the security apparatus became more pronounced. The election of the PAN’s candidate Vicente Fox as President of Mexico signaled the democratic transition at the national level. He appointed an Army General as the head of the PGR thus changing the balance of power between civilian and military law enforcement, and giving more formal power to the military. During Fox’s term, the Secretary of Defense (SEDENA) and Secretary of Marine (SEMAR) engaged more directly in public security, and carried out joint operations with the PGR. The PGR was clearly weakened; the military had more de facto power in controlling operations, and deployed far more personnel in anti organized crime operations than the PGR itself (Artz 2011). This situation stood in sharp contrast to the militarized operations during the Zedillo administration where the PGR deployed military personnel temporarily but maintained operational control. Another important change within the PGR related to the replacement of the Federal Judicial Police with the Federal Investigative Agency (AFI). Fox created the Secretariat for Public Security (SSP), and moved the Federal Preventive Police (PFP) from the Secretariat of Governance (SEGOB) to the SSP, thus signaling that he wanted to give more independence to the PFP. The Secretariat for Public Security formally separated the actions of preventive policing, investigation, and prosecution. In addition the judiciary increased its authority and influence in the political system (Schatz, Concha and Magaloni 2007). These reforms broke down the   142   concentration of policing and investigative responsibilities in the PGR, generating conflicts between different enforcement agencies: federal police and other police forces on the ground resented that the “ministerial” or judicial police did not communicate on the results of investigations of people that had been previously captured by them. Before the creation of the SSP, the police did not have a voice107 and thus conflicts with the PGR were not common. Many of the law enforcement reforms that occurred during Fox’s term were related to the democratization process. Up to that point, unlike the case of Colombia, state enforcement institutions had changed not in the process of responding to the violence generated by criminals, but rather in response to domestic and international pressure to democratize, and to increasing accusations of corruption within different branches of the government. The threat of violence, however, started to increase in response to several changes within criminal organizations. 3.3.3.1.Drug violence increases In the first years of the Fox government, important changes occurred in the criminal market. The government publicly acknowledged the existence of the Zetas, as their power, violent capacity, and territorial reach were rapidly growing. In 2003 the PGR reported grenade attacks by los Zetas in the state of Tamaulipas, and increasingly, the Zetas were associated with particular methods of violence, such as beheadings, putting marks on corpses, and using weapons and gear used by the Army. On September 11 2004, another incident marked increased confrontation between criminal groups.                                                                                                                 107 Author’s interview with former high-ranking officer of the PGR. Mexico City, September 26, 2011.   143   Members of the Sinaloa DTO assassinated Rodolfo Carrillo, a member of the Juarez DTO, and his death deepened the territorial confrontations among trafficking groups that had been brewing throughout the 1990s. These confrontations generated increasingly frequent occurrences of violence as leaderships constantly re-accommodated and organizations competed for territories. The available data on homicides in general, and drug violence in particular, although imperfect, suggests that drug related homicides started to increase in 2005 (Osorio 2013, Rios 2012, Molzahn, Rios and Shirk 2012) although general homicide rates were still showing a downward trend (Escalante 2009). Highly visible forms of violence sporadically appeared, although they were for the most part scattered across the country. The Mexican government claimed that changes in criminal behavior resulted from transformations in drug trafficking after the 9/11 terrorist attacks in the United States. The increasing enforcement pressure along the border, officials argued, made it more difficult for criminals to move drugs into the United States, and thus motivated them to diversify their portfolios into internal drug distribution. Shrinking profits from international distribution, along with a heightened need to control the micro territories associated with internal distribution, made organizations more violent. Evidence from drug use surveys in Mexico does point to an increase in internal drug consumption after 9/11, which in turn can reflect the criminals’ diversification to internal markets. Yet a disaggregation of survey data suggests that an upward consumption trend preceded 9/11, as can be seen in Figure 3.2, although it increased significantly around 2005. Likewise, although drug seizure data108 suggests that after 9/11 seizures increased in the border.                                                                                                                 108 Of course, this data suffers from all the problems of statistics discussed in chapter 2, and thus it cannot be interpreted clearly. For example, depending on the interests of the generators of statistics, an increase in   144   Yet, seizures also spiked in 1999 and then, in the case of cocaine, declined steadily since 2007 (NDIC 2011), thus making it difficult to clearly connect market and violence trends. Thus, it is not clear that heightened border controls after 9/11 reduced profits, increased internal drug distribution, and in turn caused violent behavior.109 The changes in criminal behavior more likely resulted from heightened disputes as organizations became more geographically differentiated, and as the change in the state apparatus and political configuration of the country transformed the incentives of criminals to use violence. Figure 3.2. Number of new drug users in Mexico 1984-2008 Source: ENA 2008, author’s calculations. These numbers correspond to new drug users of marijuana, cocaine, and methamphetamines in a given year according to addiction surveys. They provide only a rough calculation of drug use. The worsening of the security situation towards the end of the Fox government, especially in border regions, led the President to initiate a military operation known as “Mexico Seguro” (Safe Mexico Plan) that dispatched military troops to five cities in the                                                                                                                 seizures is usually interpreted as a result of heightened pressures and not as a sign of growing markets. Yet decreased seizures are usually interpreted as a reduction in markets and not as a sign of lenient policing. 109 Although of course it many occasions heightened border controls affect criminal behavior, for example making it more sophisticated as criminals adapt to increased controls (Andreas 2001).   145   states of Baja California, Sinaloa, and Tamaulipas in June 2005, and was later extended to the states of Coahuila, Sonora, Chihuahua, and Veracruz. The operation was conceived as a joint operation to face and confront growing violence derived from “confrontations between organized delinquency gangs.”110 The operations formally involved all actors of the public security system: the Governance Secretariat, the Secretariat for Public Security, the PGR, the Army, the Marines, and even the Finance Secretariat, and also federal, state, and municipal forces. The plan formally dispatched soldiers to the regions, and thus represented an omen to the large-scale military display that Fox’s successor Felipe Calderón initiated in 2006. As we will see next, Fox’s plan did not have the same fragmenting effect on the security apparatus that Calderón’s plan did because it had a more limited scale, both in size, and in duration. 3.3.4. The road towards breakdown and transformation In 2006, the PAN candidate Felipe Calderón won the Presidency of Mexico with a very thin margin: He got 35.89% of votes compared to 35.31% of his main opponent from the Party of the Democratic Revolution (PRD) Andrés Manuel López Obrador. With serious doubts hanging over his electoral victory, Calderon was in dire need of a legitimacy boost, and a declaration of war against trafficking organizations became one of this crucial legitimizing tools (Osorio 2013). Calderon’s decision to deploy military troops throughout the country became the catalyst for the state fragmentation that had been occurring since Fox’s presidency.                                                                                                                 110 Presidencia de la República, Boletín de la puesta en marcha del operativo México Seguro. June 13 2005.   146   The first joint military operation started ten days after President Calderón took office, in his native state of Michoacán; military deployments continued in Tamaulipas in December 2006; Tijuana in January 2007 with more than 3,000 military elements; and in Chihuahua in March 2008 with 5,000 military troops and 2,700 federal police officers. An extreme and tragic spike in violence followed Calderón’s declaration of war: over Calderón’s term there were at least 65,000 deaths related to drug violence and about 120,000 total homicides. According to a tally of drug homicides kept by the newspaper Reforma, drug related killings increased from 2,280 in 2007 to 5,153 in 2008 (Shirk 2010), with a notable, although changing, pattern of geographic concentration. In 2008 for example, Baja California, Chihuahua, and Sinaloa comprised more than half of the entire drug related killings in the country (Shirk 2010). The upsurge in drug violence in Calderón’s term quickly caused a sharp increase in the overall homicide rate of the country, after almost two decades of constant decline: Homicide rates increased from 8.5 per 100,000 inhabitants in 2007 to 13.5 in 2008, and by 2011 they had reached 24.22. In 2010, the country experienced the highest level of drug related killings (15,273 according to government figures) as well as a growing dispersion of violence throughout the country (Duran-Martinez, Hazard, and Rios 2010). By the end of Calderón’s term in 2012, violence seemed to have leveled off and to have relapsed from a pattern of geographic dispersion (Molzahn, Rodriguez, and Shirk 2013) yet, the tally of violence during Calderon’s term had surpassed the lethality of many civil wars (Osorio 2013). Additionally, during the period, other violent crimes such as kidnappings, extortion, and violent robberies skyrocketed (Mexico Evalua 2012).   147   The initial spikes of violence during Calderon’s term occurred precisely in areas where the state deployed military troops. Paradoxically, the state’s inability to establish the monopoly of force became clearer in those locations where the state had displayed most publicly its coercive capacity through urban military operations. This paradox was again a reminder that violence is not synonymous with state absence, and it can actually emerge in those places where the state has attempted to perform its authority more vividly. Military deployments worsened the upward trend in drug homicides that seemed to be occurring since 2005. However it is crucial to note that, as we will see in chapters 5 and 6, even in those locations where the state deployed its military force, local political conditions and particular correlations of power between criminal forces, marked different paths for violence. For example, while Ciudad Juárez experienced an increase of over 700% in homicides between 2007 and 2008, the increase in Tijuana was 259% and in Culiacán 205%. Thus, even though the military deployment was pivotal for the increase of violence in the entire country, in different cities and states violence evolved depending on the level of cohesion of the local state apparatus before the deployments, and on the dynamics of confrontation between trafficking groups. Alongside the military deployments, other reforms contributed to completely reshape the security apparatus during Calderon’s term. The most important was a shift away from the centralization of antinarcotics responsibilities in the General Attorney’s office, towards a major emphasis on a National System of Public Security. The federal investigation agency (AFI) created under Fox, was dismantled and merged with the Federal Preventive Police, eventually becoming the Federal Police (PF). This shift   148   consolidated the fragmentation of power that the PGR experienced in former years. The growth of federal police forces, from just over 6,000 members to about 35,000 in six years (2006-2012), was key in this process. Since the beginning of the term, Calderón requested support from the United States through an Initiative known as Plan Mérida announced in 2007, whereby the US government would provide 1,400 million dollars for criminal justice reform, training, equipment, and technical and military resources to support the combat against drug trafficking in Mexico, Centroamerica, Haiti, and the Dominican Republic (Rodriguez 2010). From the beginning, the Initiative suffered problems such as lack of transparency in design and implementation, as well as an excessive focus on punitive measures (Rodriguez 2010). Plan Merida reinforced the militarized approach of the Calderon government, but unlike Plan Colombia, it seemed to be less consequential for the reconfiguration of state enforcement. If the processes leading to a fragmentation of state, and to confrontations between criminal groups had been evolving before Calderón’s term, then the crucial question is why violence exploded and spiked the way it did after 2006. And the answer is that the policing decisions of the government exacerbated and accelerated the process of state fragmentation. Unlike previous governments, the impact of Calderon’s mobilization of the military and strengthening of the federal police was immediate. At the same time, the confrontations between criminal groups that were proliferating since the late 1990s, had not, until 2008, became all out wars for territory. For example in Ciudad Juarez, the confrontation between the Juarez DTO and the Sinaloa DTO worsened during the early   149   2000s, but it was not until late 2007 that it implied an invasion of territory by forces of the Sinaloa DTO. The state enforcement actions further exacerbated the process of criminal confrontation, as the capture of criminal leaders generated succession disputes between and within groups. In 2007, the government officially recognized six criminal organizations operating in Mexican territory. With the splits within the Tijuana DTO in 2007 and the Sinaloa DTO in 2008, the number of organizations increased to eight in 2008; by 2010 at least 12 organizations resulting from splits within the original six groups operated alongside smaller local organizations (Guerrero 2011). By 2011 two major organizations seemed to be growing stronger –the Zetas and the Sinaloa DTO- while smaller groups still operated throughout the territory (Guerrero 2012). One aspect that deserves comment is the growth and action of the Zetas during Calderon’s term. Many government officials, academics, and observers, attributed to Los Zetas a substantial transformation in the methods of violence, and therefore, a great responsibility in the spikes of violence during the Calderón era. Without a doubt Los Zetas represented a different breed of criminal organization, both because of their military origins that put them closer to a paramilitary organization, and because of their willingness to expand into other illicit ventures such as migrant smuggling. It is also undeniable that the Zetas are responsible for some of the most horrid violent actions, such as the massacre of 72 Central American migrants in Tamaulipas in 2011 or the incineration of a Casino in Monterrey that led to the death of 56 people in 2011.111 Thus, Los Zetas undoubtedly played a crucial role in the upsurge of violence and in the                                                                                                                 111 In the first case, the migrants were killed for not paying an extortion fee (and probably also in a failed attempt of forced recruitment). In the second case, the casino was burned because the owners had not paid their protection quota.   150   transformation of drug trafficking in Mexico and also in Central America. However, simplifying the explosion of violence in Mexico to Los Zetas is excessive. For one thing, the cities that became more violent in 2007 and that are the focus of this dissertation – Ciudad Juárez, Culiacán and Tijuana- did not have permanent presence of the Zetas. Even though Los Zetas apparently created alliances with the rivals of the Sinaloa DTO in each of these cities, they were not the source of the criminal disputes in these areas, and with the exception of Culiacán, did not have a permanent presence of troops in these cities. 3.3.4.1. Has state capacity increased or decreased in Mexico? One crucial question that emerges when analyzing the explosion of violence in Mexico since 2006 is whether the state increased its capacity to enforce the law while directly confronting criminals. There is no doubt that the scale of violence was too big and the human consequences of the war on drugs too extreme to claim that the Mexican state had the capacity to establish its monopoly of force. But along the lines of the potential tradeoff posed in this chapter, the real question was if state enforcement actions were increasing state independence from criminals, and contributing to a professionalization and improvement in state enforcement capacities in the long run. In other words, the question was, is it possible to see violence in Mexico as the price to pay to regain the state control over criminal organizations that were becoming more powerful and violent? The rather simplistic answer to this question was constantly given by the Calderón administration to justify its decision: declaring war against traffickers was the only option   151   available to confront extremely powerful and already violent drug trafficking organizations, caught amidst disputes for turf and territory. Using the military was necessary due to extended corruption in local and state enforcement institutions. The state was in fact becoming stronger with more professionalized police forces, and a record number of drug related detentions at the federal level (188,244)112 and the capture of 21 out of 37 criminal leaders. In the words of a high-level public official of the Secretariat for Public Security “the state is improving and one has to evaluate tendencies. Homicides cannot be the only variable [to assess] a process of war occurring in tandem with institution building.” 113 Yet, the answer was a lot more complex. On the one hand, several changes in the state security apparatus during the Calderón administration could indeed contribute to a more efficient enforcement of the law in the long run. An attempt to clearly separate policing from legal prosecution by reducing the involvement of the Attorney General’s office in law enforcement was necessary to modernize public security provision. The attempts to increase the budget and the technical and investigative capacities of the police, and to professionalize its ranks, were necessary to build modern police forces. The need to respond to violence led civil society organizations to become more active in advocating for the respect of human rights, and for the first time a Law for the Protection of Victims was issued in 2013. Thus, in some ways the pressure created by violence sophisticated civilian discourses on public security and contributed to a more substantial professionalization of law enforcement and security.                                                                                                                 112 Between January 1 2007 and 30 June 2012, according to the Sixth Report of Activities of the Attorney General’s Office, August 2012. 113 Author’s interview with high-level official of the Secretariat of Public Security, Mexico City, September 27, 2011.   152   However, on the other hand, the evidence questioning the idea of a strengthened (and independent) state is overwhelming. For one thing a death toll of about 100,000 victims of homicide is difficult to be understood as a process of state consolidation and as a sign of state capacity to enforce the law. Even if violence decreased, as seemed to be the case by the end of the Calderón Sexenio, the process of professionalization and consolidation of the state security apparatus appeared to still be very limited. First, the increasing number of detainees during the Mexican “drug war” was not matched by an equal ability to effectively prosecute and punish criminals, and it is estimated that during the Calderón administration 80% of organized crime detainees went free.114 This may signal either incompetence of the judicial system or an indiscriminate detention policy that led to many unfair detentions. Second, federal police forces grew significantly, and efforts to professionalize and to eliminate corruption became more comprehensive, but mechanisms to train officers, evaluate police performance, and to effectively investigate and prosecute corruption were still weak. Corruption cases were generally followed by massive firings rather than by investigation and strategic planning. Third, even if the stable state protection networks that criminals enjoyed during the era of PRI domination had been disjointed, corruption was not eliminated; it rather became more dispersed. Finally, human rights violations and abuses of force by state forces proliferated during the Calderón administration. Forced disappearances skyrocketed, and it is estimated that there were more than 26,000115 victims of disappearance. According to a study conducted                                                                                                                 114 Stone, H. 2012 “Under Calderón 80% of Organized Crime Detainees Went Free” Insight Crime November 4. Available at http://www.insightcrime.org/news-briefs/calderon-80-organized-crime- detainees-free Accessed [28 January 2013] 115 According to data from the Registro Nacional de Personas Extraviadas y Desaparecidas. Available at http://sirenped.secrretariadoejecutivo.gob.mx/CritMinTPadronPersonalPublicoSFotoCNT?idExtDes=1, Accessed [28 February 2013].   153   by Human Rights Watch (2013) analyzing 249 cases of disappearance, state forces participated actively in at least 149 cases, and investigations were plagued by state inefficiency, anomalies, and impunity. In sum, the situation in Mexico illustrates the tradeoff between state coercive capacity and state autonomy, between order and democratic checks and balances. The authoritarian state under PRI rule had more capacity to regulate criminal behavior than the recently democratizing state that is building up institutions. Yet, as suggested at the beginning of the chapter, the tradeoff is not a zero-sum game. The problems of the justice system and of the increasingly powerful –but not necessarily professional- police forces suggest that democratic checks and balances have not been effectively introduced even though the one party rule does no longer exists. It is possible that in the long run some of the steps taken in recent years (such as separating the prevention, investigation, and prosecution of crimes) can strengthen the state ability to enforce the law, but in the short run, criminal power, corruption, incapacity, and state abuse are still rampant. 3.4. Conclusion There are many differences between Colombia and Mexico when it comes to the evolution of drug trafficking. As Palacios and Serrano (2010) point out, the lack of effective centralization of political power in Colombia, the existence of an armed conflict, the different composition of the illegal markets, and different patterns of interaction with the United States, constitute sharp differences between Colombia and Mexico. Yet, the two countries are similar in many other aspects, especially because they illustrate the paradox of institutional transformation and democratization with growing   154   violence, and thus the possibility that a state that is transforming its capacity to regulate interactions with citizens, can also be a state that is losing capacity to regulate interactions between citizens, as Daniel Brinks (2012) suggests. In the terms of the theory I propose, a cohesive state that is more able to enforce its coercive capacity is also a state that is more able to provide predictable protection to traffickers. In this sense, the challenge for countries that like Colombia and Mexico confront criminal actors and are still building professionalized, accountable, and democratic, state security forces, is to, as Bejarano and Pizarro (2005) put it, rebuild authority without authoritarianism. As illustrated in this chapter a tradeoff does not imply a zero sum game: processes of institutional transformation that unleash violence can also lead the state to improve its crime fighting capacities. As violence increases, the efforts to combat crime can over time strengthen the enforcement ability of the state. It is possible to think along these lines, that eventually Mexico can emerge out of its public security crisis with stronger police institutions, as occurred to some extent in Colombia. However, this does not mean that the state security apparatus has achieved its full potential, especially at the local level, as the recent trends of crime and violence in Colombia illustrate. Even if the state coercive capacity increases, the ability to institute the rule of law can still be limited by impunity, corruption, and abuses of power. As this chapter sketched, and as will be seen in the following chapters, drug violence is the result of complex interactions between states and criminals. The possibility of a tradeoff between state capacity to maintain the monopoly of force and state autonomy from criminal influence reminds us that while the absence of violence can reflect a better capacity of the state of control territories, it can also imply that the state is   155   not autonomous from the influence of criminal actors. Moreover, relatively low levels of violence do not necessarily imply an absence of criminal activities, frequent but not visible violence can still exist even when the state is cohesive, and levels of competition in the illegal market can affect the frequency of violence. These elements are particularly relevant in the case of Medellin, the subject of the next chapter.   156   CHAPTER 4. FROM A “PERFECT STORM” OF VIOLENCE TO UNSTABLE PACIFICATION: THE CASE OF MEDELLÍN Medellín, capital city of the northwestern department (state) of Antioquia, and second largest city of Colombia, has been besieged by multiple armed and criminal actors for the past three decades, and has witnessed stark transformations in the forms and perpetrators of violence: an initial period of very frequent and visible violence, followed by frequent but not visible violence, a subsequent phase of pacification, and then a fourth phase characterized by a gradual return of low visibility violence. Not surprisingly, Medellin has gained significant attention from observers and academics alike, especially after 2003, when the city experienced a historic decline in its homicide rate that coincided with broad urban transformations, which led many observes to label this process the Medellín “miracle.”116 What explains this sharp variation within one city? How can we explain one of the most striking transformations that any city has experienced worldwide, from being the most violent city in the world in the 1990s, to a sharp decrease in homicide rates in the 2000s? This chapter shows how transformations in the organization of the criminal world and in the state affected violence in the city. I explore how the constant violence in Medellin generated organized civil society and state responses that contributed to improvements in living conditions and urban governance. Still, these factors alone cannot explain the evolution of urban violence in Medellin because, in 2008, despite                                                                                                                 116 See for example Faiola, A., 2011. “Sustaining the Medellin Miracle: Colombia Struggles to Hold on to Gains from Globalization”, The Washington Post, 11 July; and Fukuyama F., and Colby, S., 2011. “Half a miracle.” Foreign Policy, May-June.   157   these significant transformations, violence spiked again. To explain this resurgence of violence, I focus on the increasing competition in the criminal world. The case of Medellin illustrates powerfully how within the same city the type of criminal armed coercion and the relations between gangs and criminal organizations can change radically, with important consequences for violence. The comparison of Medellin over time also provides the most controlled test of this dissertation’s argument as it holds crucial variables constant, such as the specific economic importance of Medellín, its history as a hub for early industrialization, and the particular characteristics of its urban landscape. Medellín was the center of an early industrialization and modernization process in Colombia, heavily promoted by a local, conservative elite (Franco 2005, Safford 1965, Salazar and Jaramillo 1992, Walton 1977). As a center for economic development since the 1950s, Medellin attracted large migratory flows117 motivated by economic opportunities, but also by the need of rural populations to escape the violence that affected Colombia since the 1950s.118 Economic incentives attracted large and young labor forces that in times of economic boom were ready to be employed, but in times of economic crisis suffered the consequences of the lack of infrastructure and services that the city was not prepared (or willing) to provide. These economic processes transformed profoundly the social structure of Medellin, and traditional conservative values clashed with new consumer habits and ideas, traditions derived from older small crime traditions, and the advent of drug trafficking in the city (Roldán 2003, Salazar and Jaramillo 1992).                                                                                                                 117 In Medellin by 2005 the population was more than double its size in 1973 (Lamb 2010) 118 This period is known as “La Violencia” (with capital letters) and refers to a period of interparty violence initiated in 1948 that constituted the seed of modern armed conflict in Colombia (Palacios and Safford 2002).   158   The history of Medellin’s illegal actors is also inextricably linked to its geostrategic importance as the city connects key drug producing areas, important transportation routes and infrastructure projects, and highly profitable market centers for both legal and illegal products (CNRR 2011).119 It would be impossible to fully understand violence in Medellin without considering the deep social, cultural, and economic transformations of the city, and its geostrategic importance. Furthermore it would be inappropriate to pretend that drug violence has been the only form of violence affecting Medellin because a multitude of actors, mixing criminal, political, and personal agendas created an extremely complex dynamic where actors, alliances, and motivations have been extremely fluid (Gutierrez and Jaramillo 2005, p. 197). Yet, social and economic factors that move slowly cannot explain sharp changes in the content and dynamics of violence (Escalante 2011, Kalyvas 2006) like those Medellin has experienced; and despite its complexity, it would be impossible to understand the history of violence in Medellín without considering the influence of drug trafficking. To analyze the evolution of violence through my political economy framework for explaining drug violence, the chapter is divided in four sections that analyze the main periods of violence in Medellin since the 1980s as the result of interactions between the state security apparatus and the illegal market structure. When the state security apparatus has been more cohesive, violence has tended to be less visible, as states are more credible in protecting, or alternatively, persecuting criminals. In turn, criminals lose incentives to use visible violence that can risk the protection they receive from the state, or spur the                                                                                                                 119 Author’s interviews with experts in Medellin, October 2010.   159   actions of an effective state. Alternatively, violence has been more visible when the state has been more fragmented. Violence has also been more frequent when there is competition in the illegal market, and pacification occurred when there was a monopoly of criminality. Outsourcing of criminal coercion and violence to youth gangs has also characterized the most violent periods in Medellín. The first section focuses in the period 1984-1993, a period known as the narcoterrorist war, when both the frequency of violence - the rate at which homicides occurred-, and the visibility of violence –criminals’ claim of responsibility for their attacks and their use of methods that publicly displayed the evidence of violence- increased exponentially, as the city was besieged by the Medellin DTO and Pablo Escobar. This period was the “perfect storm” when all the variables leading to higher drug violence combined to produce the worst possible outcome: an extremely fragmented state security apparatus; a competitive criminal market; and the outsourcing of violence to youth gangs. The second section analyzes the period immediately following Escobar’s death in 1993 until 2002. During that period drug trafficking underwent a re- accommodation that contributed to maintain highly frequent violence, although less visible than in Escobar’s period because the state was becoming more cohesive and thus criminals started to refrain from using visible violence that could force the attention of a more effective state. The third section analyzes the period of pacification between 2003 and 2007 and shows how a cohesive state and a criminal monopoly allowed a reduction of violence in the city. It also shows how the criminal leader of the time, Diego Fernando Murillo a.k.a Don Berna, strategically disciplined gangs in the city in order to maintain “peace”. The final section explores the return of violence that occurred in 2008, after   160   Murillo was extradited to the United States, and disputes for succession of his leadership increased competition in the criminal market, and led to an explosion in the frequency of violence. This section shows that as a tactical adaptation to a cohesive state that became more effective in capturing criminals, criminals have learned to use less visible violence. 4.1. The perfect storm of violence 4.1.1. The rise and fall of Pablo Escobar and the Medellin Drug Trafficking organization The origins of drug trafficking organizations in Medellin were tied to long-term smuggling traditions dating back to the nineteenth century. In the 1970s the city witnessed the emergence of the first successful drug traffickers that, coming from very poor upbringings, amassed large fortunes. This decade also witnessed a boom of drug carriers, known as mules (Riaño 2006, Salazar 2001). In 1976 Pablo Escobar emerged publicly as a criminal when he was arrested for the first time on drug trafficking charges; by the 1980s Escobar and the Medellin DTO made the city famous worldwide due to their extreme violence, their declaration of war against the state, and Escobar’s megalomaniac personality that transformed the world of drug trafficking (Bowden 2002). Understanding violence during the period of domination of the Medellin DTO requires analyzing both its extreme frequency, and its extreme visibility. The first, as I explain in this subsection, was a result of the violent disputes within a competitive market, and the second, as I explain later, a result of the fragmentation of the state security apparatus, which made protection unpredictable and thus created incentives for criminals to use visible violence that could signal their power, as not using violence would not guarantee that they were not prosecuted by law enforcers.   161   At a first glance it seems counterintuitive to classify Medellin’s criminal market as competitive when Escobar’s reputation was precisely that of having enormous power, which in fact, he had.120 Yet, the power of Escobar and the Medellin DTO was far from being a complete monopoly even at the height of his power, and criminal competition became more acute as profound divisions increasingly disjointed the Medellin DTO. Growing criminal competition explained the sustained increase in the frequency of violence since 1984: an average homicide rate of 232 homicides per 100,000 inhabitants between 1984 and 1993, far above an average homicide rate of 63 for the entire country. A historic peak rate of 381 homicides occurred in 1991 and constituted the most violent year in Medellin’s history. (Figure 4.1.) Figure 4.1. Homicide rates in Medellin 1984-2010 (per 100,000 inhabitants) Source: Author’s elaboration, data from Policia Nacional de Colombia, Secretaria de Gobierno de Medellin By the mid 1980s, Escobar consolidated a strong organization that controlled previously fragmented trafficking groups. At the same time, urban guerrilla militias, and                                                                                                                 120 Even though recent studies have questioned the idea of Escobar’s organization as highly centralized and organized (Kenney 2007) it is undeniable that his leadership and power were far more centralized than the leadership that exists nowadays in Colombian trafficking organizations.   162   self-defense militias, also struggled for the control of Medellin. In 1984, demobilization camps were established in Medellin as part of a peace processes between the national government and the M-19 guerrilla. These camps became a military training ground for young men who would later become urban guerrilla militias or gang members (Salazar 1990, Lamb 2010). Gangs started to proliferate due to the money provided by traffickers, and to the combat expertise provided in guerrilla camps, and as a result, problems of criminality multiplied: extortion, robberies, drug dealing, and killings. Criminality in turn provided the breeding ground and motivation for self-defense militias and vigilante groups, which became another player in violence in their quest for “protecting” the community and “reducing” criminality. Even though guerrillas and militias were not directly fighting for turf with traffickers, they challenged the Medellin DTO’s power, and complicated the patterns of alliances and confrontations among armed actors. Militias usually targeted lower level gangs at the service of traffickers while tolerating the more powerful broker organizations in charge of hiring sicarios (hitmen or contract killers). In some cases, militias stirred conflicts among gangs, so they would destroy each other (Medina Franco 2006). In other cases, militias recruited members from the very same groups they were targeting. Traffickers did not control these violent dynamics, and gangs supported by traffickers fought gangs supported by militias, and the confrontations snowballed in a war of all against all (Salazar and Jaramillo 1992, p. 92). Competition within the criminal market became even more severe at the beginning of the 1990s, sustaining the upward trend in homicide rates. By then, Escobar had instigated deep opposition within his organization and also with his biggest   163   competitors, the members of the Cali DTO. The Cali DTO repeatedly rejected Escobar’s plans to confront the state121 and on January 11, 1988 members of the Cali DTO activated a car bomb against one of Escobar’s properties and he in turn responded with serious terrorist attacks in Cali, thus generating an all-out war between the two groups. Furthermore by early 1990, some of Escobar’s partners started to openly criticize terrorist actions (explained below) and the monetary quotas Escobar extracted from them to wage war against the state.122 In June 1991 Escobar handed himself to the authorities under a policy known as the “Surrender to Justice” (sometimiento a la Justicia), whereby the state offered judicial benefits to traffickers that confessed all their crimes (Pardo 1996).123 During the year Escobar spent in jail he killed three of his closest allies -Gerardo Moncada, and the brothers Fernando and Mario Galeano- thus fracturing his organization irreparably. The killing apparently had to do with money not paid for a drug transaction, but also reflected Escobar’s fear that the armed and enforcement wing of the DTO, known as the Oficina de Envigado, was getting too strong while he was in prison (Serrano 2010, p. 85). The Oficina was a broker criminal organization created to hire assassins for the Medellin DTO, but later became the main contract enforcer for traffickers. The killings thus unleashed ire among some other members of the Medellin DTO, specially Diego Fernando Murillo, who was the Galeanos’ chief of security and then became one of the founding members of the vigilante-paramilitary group Los Pepes (Persecuted by Pablo                                                                                                                 121 The relationships between the Cali and Medellin DTOs were tense at least since 1984 when the Cali DTO leader Gilberto Rodriguez Orejuela was captured in Spain and allegedly Escobar was trying to take advantage of this situation to gain market share. 122 Communiqué of Los Extradibles to the Colombian people. April 1 1990. 123 Escobar escaped from prison in July 1992 after a scandal made public his luxurious living conditions in jail, forcing the Government to announce Escobar’s prison transfer, which motivated him to escape.   164   Escobar) formed in January 1993. Los Pepes were an eclectic group that provided the opportunity for Escobar’s enemies, both internal and external -the Cali DTO, paramilitary groups, and former allies betrayed- to work together, and with the police, towards the objective of bringing Escobar down. The emergence of Los Pepes thus reflected the internal fragmentation and the external competition the Medellin DTO, and specially Pablo Escobar, confronted. As Escobar faced a larger number of competitors, his ability to control criminality further decreased, even though his increasingly difficult situation forced him to try to regulate low-level non-trafficking enterprises that he had previously ignored. As Medina (2006, p. 63-64) describes it plazas (local drug markets) were second category offices that Escobar rarely used during his first period as a capo. But things changed for the boss after his escape from La Catedral [prison]. He was no longer the omnipresent and almighty figure that media and the DEA depicted [...] His resources started to be scarce and he had no way to collect the tax he imposed over other traffickers to use his export routes. Kidnapping thus became the salvation resource to maintain provisionally his war against the Colombian state, and guarantee as a minimum the resources to remain hidden. [...] Second category offices as the drug plazas gained unprecedented centrality. The underworld of the plazas constituted eyes and ears to detect potential victims for kidnapping, or to find out the destination of people who had been kidnapped and thus collect intermediation fees. […] Eventually the Cali DTO and Los Pepes, alongside with government and international enforcement, succeeded in their objective of eliminating Escobar in 1993 (Bowden 2002, Chaparro 2005, p. 212-230, Clawson and Lee 1998, p. 176). My argument that frequent violence resulted from the increasing competition of the criminal market in Medellin provides a different reading to the common depiction of Escobar as an almighty criminal. Even though his power and influence were enormous,   165   and there are probably only few criminals that fit the imagery of a “capo” as he did, he did not monopolize criminality in Medellin. Yet, two other aspects were equally crucial to understand the extreme scenario and spiral of violence: the outsourcing of violence to youth gangs, and the fragmentation of the state security apparatus, which prevented traffickers from controlling the state, and made it difficult for the state to confront them effectively. 4.1.2 The narcoterrorist war and the fragmentation of the state security apparatus In 1984 Pablo Escobar killed Justice Minister Rodrigo Lara Bonilla in retaliation for Lara’s open criticism and enforcement actions124 against drug trafficking and Escobar; then the state struck back by enforcing extradition treaties, which in turn led the Medellin DTO to declare a war aimed at deterring the use of extradition (Camacho and Lopez 2001) that came to be known as the narcoterrorist war. These events unleashed a deadly period in Colombian history that claimed the lives of hundreds of civilians, more than 200 judges, 700 police members, and several prominent politicians. Violence thus became extremely visible. The use of notes or communiqués to claim responsibility for violence, tactical innovations in the methods of killing, collective violence, attacks that took place in broad day light or in the middle of public spaces, and the targeting of public officials, characterized the visibility of violence in Medellín. The visibility of violence reflected the complex interactions taking place within a fragmented state that was unable to effectively confront criminals and enforce the law, but that also made it difficult for traffickers to obtain credible protection from the state.                                                                                                                 124 Such as the dismantling of the “Tranquilandia” drug processing facility described in Chapter 3.   166   The framework I propose focuses less on the events that initiated this particular wave of violence and more on the power relationships that framed these events. In this sense, it matters less who made the first move to create violence, but rather how the degree of power fragmentation within the state shaped the effects of that move. An enforcement action may have different consequences if it occurs within a fragmented state security apparatus, and the actions of criminal organization may unleash different reactions depending on who its competitors are and how is the security apparatus organized. In other words, in a counterfactual scenario, it is likely that the narcoterrorist war in Medellín would have been shorter and less intense had the government had more capacity to dismantle the Medellin DTO, or had Escobar found protection from a more cohesive state. As described in Chapter 1, state fragmentation can be extreme when there are conflicts among levels of government, between enforcement agencies, and when events such as democratization or anticorruption purges reduce the time horizons of elected and public officials. All these elements were present in Medellin: an opening of electoral competition in the mid 1980s started to shake political life in a city accustomed to domination by a conservative political elite (Jaramillo and Salazar 1992); in response to increasing violence, the central government deployed the military in urban anti-gang operations generating bitter conflicts between military and police; Medellín’s geographic location, contiguous to nine municipalities125 made efforts to control the city extremely dependent on the actions of the mayors of the other municipalities thus increasing the                                                                                                                 125 The region comprised by Medellin and other 9 municipalities is known as the Metropolitan Area of the Valle de Aburra. Administratively each municipality is independent but in practical terms life is extremely interconnected, to the extent that Medellín’s subway system runs from the municipality of Bello in the North to Itaguí in the South.   167   difficulties for coordination; and finally, early attempts to transform police forces drastically changed the rules of the game for traffickers. State fragmentation created more incentives for the Medellin traffickers to use visible violence as a tool to intimidate, punish, and reaffirm their power. The first indication of the traffickers’ willingness to use visible violence appeared in December 1981 with the creation of the group MAS (Muerte a Secuestradores, or Death to Kidnappers). The group was the first manifestation of an explosive combination between paramilitary groups and drug trafficking that besieged Colombia in later years. Pablo Escobar and 200 traffickers founded it as a reaction to the kidnapping by the guerrilla group M-19 of the sister of the Ochoa brothers, Escobar’s trafficking partners. The group announced its emergence raining leaflets over the crowd attending a soccer match, stating its objective to publicly execute kidnappers, to offer rewards for information, and to hang, shoot, and mark the executed (Lamb 2010, p. 55). Over the next few years the MAS became the perpetrator of multiple acts of violence that blurred the lines between criminal vendettas and acts of political violence (Salazar 1990), many copy cat groups emerged, and in the end both the original group and the copy cat cells nurtured the paramilitary phenomenon and violence. After the creation of MAS, traffickers incrementally introduced new methods for conducting violence as can be seen in the difference between methods used in 1984 and methods used in 1989 (see Table 4.1.). New methods were used to attack state targets, but also to eliminate competitors, and to discipline and punish members of the organization. Many of these methods publicly displayed the perpetration of violence, for example using drive-by shootings (Lamb 2010, p. 52) and explosions.   168   Table 4.1. Methods used in violent attacks, Medellin 1984 and 1989 1984 % 1989 % Banner (without victims) Car bombs and explosions 12 4.65 56 4.83 Combat with fire arm or explosives 1 0.39 33 2.84 Corpse with a note 2 0.17 Corpse wrapped in blanket 2 0.17 Fire 24 2.07 Head in cooler/frozen corpse 1 0.39 1 0.09 Levantón o paseo (forced 2 0.78 17 1.47 disappearance) Mutilation or incineration with note Mutilation or incineration without 4 1.55 2 0.17 note Sexual violence 2 0.17 Sicariato (Drive by shootings) 75 29.07 66 5.69 Simple use of fire arm 131 50.78 832 71.72 Simple use of knives 28 10.85 94 8.10 Strangulation 4 0.34 Torture 4 1.55 21 1.81 Total number of events 258 1160 Source: Author’s dataset on drug related violence, information compiled from El Colombiano newspaper (Medellín). As the presence, power, and violence of the Medellin traffickers grew, their relations with the state became more complicated. The political and electoral liberalization that Colombia experienced in the early 1980s, after a long power sharing agreement between two political parties,126 affected profoundly how illegal actors interacted with political forces and also the feasibility of coordinating enforcement actions. One of the cornerstones of electoral liberalization was the introduction of popular elections for Mayors in 1986, which increased the number of actors involved in local public security. Even though antinarcotics was not officially within a Mayor’s realm of                                                                                                                 126 The Frente Nacional, described in Chapter 3.   169   action, the violent consequences of it were, thus creating a multiplicity of actors that could eventually determine a city’s security policy and augmenting the possibilities of conflicts between levels of government. The challenge of coordination was greater given the extent of violence and the need to deal with the other mayors in the metropolitan area.127 The other cornerstone of political liberalization in the 1980s was an increase in electoral competition (Gutierrez 2007) which provided renewed opportunities for criminals to access the state, but also made it increasingly difficult for them to maintain stable corrupt deals, as the denunciation of corruption became a powerful tool of electoral competition. The effects of political liberalization were profoundly felt in Medellin because it had been traditionally the stronghold of a conservative political elite and started to experience notable increases in electoral competition in the late 1970s and early 1980s.128 A powerful and entrenched elite that started to see its power challenged, reacted strongly to Pablo Escobar’s attempt to mobilize the support of popular classes, and Escobar in turn reacted to the “bourgeoisie” that blocked his intentions. In the words of former minister of Defense Rafael Pardo, Escobar’s violence was not only a reaction to extradition but “the most violent expression of a social class trying to be recognized and finding a place in the Colombian society” (1996, p. 193). This class struggle was more                                                                                                                 127 Sergio Fajardo, mayor of Medellin between 2003-2007 and credited with a reduction of violence in the city expressed in an interview with the author that during his term there was not collaboration with any other major in the metropolitan area “I could not count with the nearby government, for example, think about councilman Upegui in Envigado who was murdered because of his links with drug trafficking, and yet was mourned in the city’s council” (Interview conducted in Bogotá on November 6, 2010). 128 The effective number of parties (a measure of political competition) for Senate elections in Medellin for the period 1970-1982 was 2.42 and it increased to 2.6 for the period 1982-1994. Similarly the same indicator for the Lower Chamber was 2.40 in 1970-82 and increased to 3.88 in 1982-94, and in local council the effective number of parties went from 2.54 in 70-82 to 3.30 in 1982-1994. (Author’s analysis of data provided by Congreso Visible, Universidad de los Andes).   170   profound given the political and economic gaps and inequalities that characterized Medellin. The main opponent of drug traffickers in general, and champion of morality in politics was the Nuevo Liberalismo, a new party formed in 1980 by Luis Carlos Galán, and its role reflected the impact of political liberalization on corrupt deals of protection between criminals and state. Nuevo Liberalismo had its main electoral stronghold in major cities like Bogotá and Medellín. In February 1982, Pablo Escobar appeared on the Congress election ballot as an alternate member for the Lower Chamber with Jairo Ortega, a liberal politician of Antioquia, who was seeking support from Nuevo Liberalismo. When Galán and Ivan Marulanda, coordinator of Nuevo Liberalismo in Antioquia, learned about Escobar’s intention, they requested Escobar’s elimination from the list. Despite Nuevo Liberalismo’s rejection, Escobar and Ortega ran with the political movement Renovación Liberal and were elected for the Lower Chamber of Congress. Renovación Liberal also elected three council members in the Medellin’s neighboring municipality of Envigado, who became loyal servers of Escobar.129 Nuevo Liberalismo did not stop its opposition, and managed to expel Escobar from Congress in 1983. Thus Pablo Escobar’s infamous attempt to become a politician reflected the doors that democratization was opening for traffickers to access the state, but the Nuevo Liberalismo’s reaction also reflected that democratization could seriously undermine the possibility for traffickers to get protection. Escobar reacted quickly and visibly to his expulsion from Congress. In 1984 he assassinated the Justice Minister Rodrigo Lara Bonilla, member of Nuevo Liberalismo                                                                                                                 129 Arrázola, M., 2011. “Todo empezó en Envigado” El Espectador February 12.   171   and one of the most vocal opponents of drug traffickers. As a reaction to Lara’s assassination, then President Belisario Betancur issued an extradition order for Escobar’s trafficker partner Carlos Lehder. On November 6 1986 “Los Extraditables”, an organization founded by Pablo Escobar and members of the Medellin DTO with the main objective of impeding the enforcement of extradition, made its official appearance before the Colombian public opinion and government, releasing a document where they called for the end of extradition “in the name of national sovereignty, family rights, and human rights.”130 Los Extraditables distinguished their actions with a willingness to claim responsibility for their attacks, usually through communiqués where they explained their crucial objective of preventing the enforcement of extradition through intimidation (later on complemented with the objective of opposing “human rights abuses” by police forces) but also expressed their lack of concern in avoiding media or enforcement attention.131 Los Extraditables occasionally claimed responsibility for their attacks using notes on corpses, especially in their attacks against their Cali DTO rivals. Corpses would appear on highways pierced with hand drills, burned, and shot with notes such as “Members of the Cali DTO, for attempting against people of Medellin" (Salazar 2001, p. 183). After the emergence of Los Extraditables, Medellin experienced the systematic assassination of local state officials such as Antioquia’s state governor Antonio Roldan Betancur (July 1989), Medellin’s police chief Valdemar Franklyn Quintero (August 1989), and local Senator Federico Estrada Velez (April 1990). In August 1989, Luis                                                                                                                 130 El Tiempo. 1991. “El Fin de Los Extraditables”. El Tiempo. July 4. 131 For example, in a threat note sent to a judge The Extraditables claimed “not to be afraid of press scandals and repressions.” These communiqués belong to the author’s personal archive.   172   Carlos Galán, who was running for President, was assassinated. This event marked both the government’s strongest reaction towards traffickers and the traffickers’ all-out war against the state. Violence escalated and the killing methods aimed at exposing the evidence of attacks, such as massacres and car bombs, proliferated. The bombs provided the basis for using the “narcoterrorism” label, and dramatically affected Medellin: between 1988 and 1993 there were 719 cases labeled as terrorism.132 The escalation of violence after Galan’s assassination motivated military urban operations in Medellín, which further fragmented the security apparatus. The Army had established its first Battalion with an epicenter in Medellin in 1984. Originally the Battalion had responsibility in anti insurgency but not in anti narcotics operations, and was extremely wary of participating in antinarcotics action. In the words of Colonel Augusto Bahamon, who was in charge of first establishing the Batallion Back then, the Army was not interested in confronting a war that they knew in advance they were going to lose with a high toll in lives and prestige, because the mission of the Army is very different to that of the Police, and this was very clear to all our superiors (Bahamon 1991, p. 14-15) As Medellín became the epicenter of the state’s military response to narcoterrorism, the Fourth Brigade received the “Plan Genesis” with instructions from the Superior Command to move the Brigade’s monitoring posts into the neighborhoods most affected by drug trafficking (Bahamon 1991, p. 27). By 1989, the Army was engaged in massive urban antinarcotics and antigang operations. The Army’s use of indiscriminate search techniques and illegal detentions133 undermined its relationship with the                                                                                                                 132 According to statistics from the Investigative Police DIJIN. In the same period Cali experienced only 270 terrorist acts. 133 After the assassination of Presidential candidate Luis Carlos Galan, President Cesar Gaviria issued a series of public order decrees. One of the measures adopted authorized the retention of suspects for seven days.   173   community, while its parallel role in dismantling corrupt local police agencies created constant confrontations with the police and made even more unpredictable the framework in which criminals interacted with the state. According to Colonel Bahamon In the fight against sicarios there was no way to copy a model from another country as has been done in antiguerrilla campaigns both urban and rural. There were no legal procedures that facilitated the troop’s work. The Executive tried to modify this situation creating Military Commands in Bello, La Estrella, and Envigado. After operating with and without them it can be said that they were just a path that the military was not used to travel. The government believed that creating legal attributions was enough to obtain the desired results. The reality was quite different because the Military Commands were institutions without means, which were born and died without leaving a trace (Bahamon 1992, p. 34, 37) Bahamon’s remarks capture the Army’s frustration as well as the deep conflicts that divided the institutions in charge of the anti-drug trafficking campaign, thus illustrating the impact that the mobilization of the military, and the addition of different enforcement agencies can have in the fragmentation of the state. As military and police operations against traffickers increased, and Escobar’s networks of protection within police agencies became unstable, he initiated a campaign to eliminate policemen in Medellín. Traffickers had enjoyed protection among police agencies known as Departamentos de Seguridad y Control (Security and Control Departments) in the municipalities surrounding Medellin. In fact, fired members of these agencies later made it to the ranks of criminal organizations.134 Evidence of corruption within these agencies led the national government to eliminate municipal police forces in the metropolitan area. As the agencies were removed and reorganized, and the police engaged more forcefully in operations against the DTO and the sicarios, state protection                                                                                                                 134 Author’s interview with Rafael Pardo Bogotá January 13 2011, and Sergio Fajardo Bogotá November 6 2010. Also El Tiempo. 1990. “Envigado: detenidos doce miembros del departamento de seguridad y control”. El Tiempo March 14.   174   became increasingly unstable for traffickers. The result was that despite enjoying protection from some sectors of law enforcement, Pablo Escobar and the Medellin traffickers also faced an increasing number of actors willing to attack them, thus adding to the unpredictability of the situation. The systematic assassinations of municipal police in Medellin can be seen as a manifestation of visible violence and of the impossibility for Escobar to find predictable protection in enforcement officials, even though he had ample corruption networks.135 Pablo Escobar set a price for every cop killed in Medellin, a strategy that led to the death of 737 policemen: 420 in 1990 and 317 between 1992 and 1993 (Pardo 1996, p. 344). Escobar’s offer of paying about five hundred dollars per cop killed unleashed a “witch hunting” like period in which young sicarios would kill cops in order to get the money; cops in turn refrained from wearing uniforms making targeting more difficult and increasing “mistaken” attacks. The systematic killing of cops also unleashed violent reactions: sectors of the Elite Police created in 1989 to combat traffickers and paramilitaries killed young men they suspected to be sicarios even without having any evidence. Dangerous alliances emerged in some sectors of the police that looked for the protection of nascent anti-gang militias and paramilitary groups (Medina Franco 2006). In sum, the high fragmentation in the security apparatus in Medellin created an environment in which both credible enforcement and credible protection were difficult, thus eliminating the incentives of criminals to hide violence. This explanation as such challenges a narrative that would explain the narcoterrorist period only as a result of Escobar’s megalomaniac personality. Although his personality undoubtedly played a                                                                                                                 135 At the beginning of 1990 Los Extraditables denounced that the police had massacred 13 people with criminal antecedents at a farmhouse in the outskirts of Medellin. This was another of Los Extraditables alleged motivations to declare a war against cops in Medellin.   175   crucial role in explaining the methods he used, it only reflects one part of the story. In fact, if the use of visible violence was only the result of Escobar’s preferences then it would be difficult to explain why in some occasions he was able to hide the use of violence, for example during his brief stint in prison. Before escaping from prison in 1992, Pablo Escobar strategically conducted hidden forms of violence. Furthermore, not only Escobar used visible violence but Los Pepes also recurred to visibility in their targeted killings of people associated to Escobar, using notes on corpses that had been previously tortured. For example on February 28, 1993 architect Guillermo Londoño White, accused of having links with Pablo Escobar was found dead and his face had a dozen bullet holes, his hands were tied with duck tape, and a note on his body read “Luis Guillermo Londoño White, faithful figurehead and initiator to Pablo Escobar’s kidnappers. Pepes” (Serrano 2010, p. 66). According to some estimates Los Pepes alone killed about 125 people in Medellin during 1993 (Lamb 2010, p.67) and were predecessors not only to the horror techniques136 used by paramilitaries in Colombia137 but also to the brutal methods used by Mexican’s traffickers to publicize their acts of violence. Visible violence was also used by Escobar’s associates and towards 1990 the press releases of Los Extraditables reflected internal tensions. Thus, visible violence was not only Escobar’s decision. Both the extreme spike of violence in Medellin and the proliferation of visible violent methods reflected the third element that explains the perfect storm of violence, that is, the outsourcing of violence to youth gangs.                                                                                                                 136 For extensive descriptions and accounts of paramilitary violence see www.verdadabierta.com y www.memoriahistorica-cnrr.org.co. 137 In fact the Castaño brothers, paramilitaries who were part of Los PEPES, eventually became leaders of the main paramilitary force in Colombia, the Autodefensas Unidas de Colombia (AUC).   176   4.1.3. Combos, Bandas and Oficinas in Medellin The culture that emerged around drug trafficking, gangs, sicariato, and Pablo Escobar deeply affected an entire young generation in Medellin in the 1980s, this is what scholar and former Mayor of Medellin Alonso Salazar and scholar Ana Maria Jaramillo accurately named “the subculture of drug trafficking” (Jaramillo and Salazar 1992). Such influence has extended to current young generations engaged in violence. Referring to a 20 year-old sicario from Medellin captured in 2010, an intelligence official in Colombia remarked “For Caliche [the sicario] Pablo Escobar is a guardian angel… they [sicarios] adore him and consider him a saint. There you can you see that the influence continues for an entire generation.”138 Escobar outsourced the use of violence to gangs, yet he did not try to inhibit the common criminality produced by sicarios or gangs, as would later be the case with Diego Fernando Murillo, in the early 2000s. The history of gangs in Medellin did not start with Pablo Escobar, but they proliferated and became more sophisticated under the influence of drug trafficking. Unlike the Cali DTO that mainly recruited former security personnel (military and police) to provide for their security, Escobar made local gangs his main armed force, partially because his access to local police forces was not as extended as the one Cali traffickers had. Not all youth gangs worked directly for Escobar, and it is likely that he directly hired only a small proportion of them. Yet, the expectations of the money that could be made working for traffickers created incentives for gangs to proliferate (Ceballos and Cronshaw 2001, p. 118), and as the number of potential workers increased, so did the competition among them to obtain the best and more profitable jobs. Such competition                                                                                                                 138 Author’s interview with criminal intelligence official, Bogota, September 13 2010.   177   became another source of violence in itself, as described by a gang member interviewed by Salazar in the first ethnography of youth gangs in Medellin There have been many gangs in the neighborhood: Los Nachos, Los Montañeros, la del Loco Uribe, los Calos. And as the song goes, there is no bed for so many people. We need to remain active because if we lower defenses other gangs expel us and start to screw people […] We also fight with the police but things are easier with them because they come up shitting their pants and one knows the terrain (Quoted in Salazar 1990, p. 32) As explained in chapter 1, one of the disadvantages associated with outsourcing violence to youth gangs is that discipline is more difficult to maintain. In the case of Medellin, the difficulty to discipline gangs may seem to be at odds with the symbolic power that Escobar had on criminality and with the relative hierarchical, even if not completely regularized, structure of gangs in the city. The crucial point is that high level traffickers did not have direct contact with all gangs and specially with low level sicarios, thus they were not able or willing to control their behavior. An interview conducted by Salazar with another gang member clearly reflects this outsourcing dynamic where traffickers used the services of gangs but did not imposed vertical control on them What the Brigade [the Army] says about sicarios is a televideo. It does not work like that, with a regular structure. They put together the gangs that work directly for Escobar with others that have nothing to do with him. Everybody respects him, because he has been kind with the people and his people handle many threads, but there are also many independent combos. Around the area where I have been working the thing works like this: there is a small group of strong ones who handle the contacts at the highest level, and you don’t get to see them around anymore. They handle relationships with the bandas’ bosses, and these bosses are the ones who handle the selection process when there is a job to be done […] But many of the kids (pelados) who benefited from the business were not as serious as the first ones. Many outlandish little bands emerged. They are abusing people, killing because you did or you did not look at me. That is the people who damage the business, the eccentric ones (visajosos). (Quoted in Salazar 1991, p. 68)   178   Local scholars in Medellin describe the structure of gangs in the city as a three- tiered model composed of oficinas, bandas, and combos (Ramirez 2005, Riaño 2006). Oficinas are broker organizations, in charge of hiring and controlling the relationships with gangs hired by drug traffickers. Bandas refer to relatively sophisticated, usually older gangs, with territorial power, significant armed and economic resources, and direct relationships with traffickers and oficinas. Combos refer to more specialized and territorially limited gangs (usually controlling blocks or parts of a neighborhood) that sell their services to the bandas who in turn sell it to oficinas and traffickers (Ramirez 2005). In this model of outsourced coercion, the socialization of gang members in violence and the use of fire arms, coupled with the excess supply of force also created incentives and capacities for gangs to engage in other criminal activities like extortion, bank robberies, or kidnappings that became an additional source of violence in the city, and a stable source of income when they did not work for traffickers. The dynamics of gangs clearly illustrate that understanding intense periods of drug related violence does not mean to characterize every violent event as a product of a drug trafficking disputes, but to understand how the power dynamics of trafficking organizations reproduce or inhibit other expressions of violence. Outsourcing explained the extreme spike of violence in Medellin because it multiplied the number of potential violent actors in the city and exacerbated conflicts among them. In fact, besides all the forms of visible violence employed by the Medellin DTO, during the narcoterrist period massacres proliferated. For example on July 1 1984, seven heavily armed men opened fire against a group hanging out in a creamery after a funeral of someone who had been recently assassinated, killing five and leaving 6   179   wounded.139 Massacres like this became more frequent as gangs became better armed and more violent at the service of drug traffickers; many events of collective violence reflected the power struggles and retaliation between gangs. Massacres often affected the most depressed areas of the city, but eventually occurred in elite neighborhoods, as was the case of a massacre that took place on 23 June 1990, inside a bar where 17 young people were forced to lay on the floor and were executed, and notes were placed on the corpses announcing that four elite kids would be killed for every sicario killed in the city. A portion of these events of collective violence also reflected an identification problem as it has been defined by Kalyvas (2006). As the enemy became difficult to identify, the possibility of mistakes increased: many events of collective violence occurred when men moving in motorcycles (or cars) arrived to public places such as restaurants or creameries and opened fire against the crowd even if they were aiming at just one person. Some massacres also reflected the action of vigilantes, militias, and police forces. It is crucial to understand that the outsourcing of violence to youth gangs and a greater incidence of violence in young populations does not mean that all young people and gangs were criminals. In fact, the percentage of youth directly involved in criminal activities was small when compared with the entire young population in Medellin. Salazar (1990) argued that only 30% of gangs were related to drug traffickers. Yet, outsourcing to youth gangs reproduced violence to extreme levels, and beyond trafficking disputes, claiming the lives of many civilians. To sum up, the narcoterrorist period in Medellin was characterized by a sharp increase in the frequency of violence, caused by the competitive nature of the illegal                                                                                                                 139 El Colombiano, July 1 1984, p. 14b. Author’s dataset on drug related violence.   180   market. It was also characterized by the emergence and proliferation of visible violence characterized by the use of communiqués to claim responsibility for attacks, the targeting of public officials and cops, and the use of methods that exposed the evidence of violence, such as bombs and massacres. Visible violence reflected the difficulties of a fragmented security apparatus in confronting criminals, and the criminals’ difficulty in obtaining credible protection within the state. Two issues stand out in the characterization of visible violence in this period. First, not all the victims or the perpetrators of violence were drug traffickers; both victims and perpetrators included a mix of civilians, criminals, and state officials, and the motivations intermingled traffickers’ attacks to the state and to competitors; state officials illegally using force, criminals using copy cat techniques to infuse fear; revenge and personal disputes. The mix was even more complicated due to the role that armed guerrilla and paramilitary groups played in urban violence. This complexity highlights the difficulty of defining drug violence discussed in Chapter 2: a broad definition of drug violence would lead to including events where motivations are complex and may trascend criminality, but reducing drug violence only to events that include criminals or that pit criminals against the state, would mean to ignore the larger impact of drug trafficking on violence. In Medellin, and without pretending to reduce all violence to drug trafficking, it is clear that drug trafficking organizations radically transformed the nature of violence. While waging war against rivals and the state, traffickers made arms more easily available, socialized youth gangs into violence, naturalized the use of violence, and fueled competition among many armed groups.   181   The second crucial element is that even though visibility is a key dimension of violence, visible attacks did not cause the majority of lethal victims. As shown in Table 1 most events of violence could be classified as the simple use of firearms, that is, events in which a single killer targets a single victim and shoots at him/her. These events are not the ones that usually get more attention from the public and the media, and thus it could be argued that the acts producing most victims can be routinely ignored as they are not visible, and that a relatively small number of attacks can create an image of extreme brutality. The key point is still that such events do not emerge all the time and in all cities, and thus when these acts emerge they represent a crucial change even if they produce less victims than more routine and less visible violence. 4.2. A market in transition: (1994-2002) In the last years of the narcoterrorist period the state security apparatus in Medellin evolved into a more cohesive structure.140 In 1993, a joint military-police command known as Bloque de Busqueda that facilitated coordination between enforcement agencies (Pardo 1996, p. 448) succeeded in hunting down Escobar, killing him in 1993. The influence of United States support and the collaboration of Los Pepes were also crucial in bringing down Escobar (Bowden 2002), but it is important to recognize that through the narcoterrorist war, the state changed radically in Medellin. The death of Pablo Escobar contributed to a minor reduction of violence, but violence would still be a major concern in later years, and the state security apparatus faced many challenges during the late 1990s. This section shows how in the period 1994-                                                                                                                 140 Author’s interview with former Minister of Defense Rafael Pardo, Bogotá, January 13 2011.   182   2002 violence remained frequent because the criminal market was very competitive, yet violence was also less visible, because of the absence of Escobar, but also because the state security apparatus was not as fragmented as in the years of narcoterrorism. 4.2.1. Militias, paramilitaries and re-accommodation of traffickers in Medellín After Escobar’s death the criminal market remained as a competitive market that involved directly and indirectly, militias, guerrillas, paramilitary groups, and all the traffickers of the Medellin DTO that had not been eliminated, and were starting to re- accommodate. Competition and fragmentation were key causes of the frequent violence that continued besieging the city. Militias had consolidated by the mid 1990s and were, in the most basic definition, self-defense neighborhood groups created to “defend” communities from petty criminals. Yet, their roots and composition were complex and mixed groups derived from guerrilla training camps, criminal bands that defended turf, police forces, and even members of economic elites (Ceballos and Cronshaw 2001, Jaramillo 1994, Medina Franco 2006). A great deal of violence in this period resulted from militia actions against rival militias, militia actions against low-level criminals, and “local wars” between neighborhood gangs, which skyrocketed in the mid 1990s (Riaño 2006). The engagement of militias in drug trafficking was not direct, and it was in fact contradictory. On the one hand in their attempt to regulate behaviors and supposedly control criminality, militias targeted and eliminated drug consumers and petty criminals. However, on the other hand, militias rarely targeted the higher-level traffickers, or the big oficinas and combos that worked for traffickers. Sometimes oficinas supported militias’   183   cleansing activities, as some traffickers considered that the instability brought by delinquents could threaten them (Riaño 2006, p. 33). Competition emerged because different trafficking factions aligned with different sectors of militias; supposed enemies would ally to defeat common enemies (Jaramillo, Ceballos and Villa, 1998) in a pattern of alliance and confrontation that was very unstable and generated frequent violence. The average homicide rate between 1994 and 2002, although lower than in the previous period, was still very high by any standards, and much higher than the national average: 187 deaths per 100,000 inhabitants for Medellin compared to a rate of 63 for Colombia as a whole. The aftermath of the Medellin DTO has often been described as a period of fragmentation between “microcartelitos” or smaller trafficking organizations (Bagley 2012, Gootenberg 2012). While it is true that the market was very fragmented after Escobar’s death, a far more complex process was taking place in this period: former members of the Medellin DTO and members of the Oficina de Envigado were consolidating their power in the city, and in their effort to consolidate power, they extended the scope and strength of alliances with paramilitary groups. A key protagonist of this process was Diego Murillo, who as already mentioned, started as Escobar’s hit man, became founder of Los Pepes, and eventually achieved a monopoly of power in Medellin. The “microcartelitos” were just the surface of a more complex process of re- accommodation. Murillo himself recognized 1994 to be a crucial year in the consolidation of alliances with paramilitaries, and in the expansion of the paramilitary project, which mixed political and criminal motivations in very complicated ways.   184   In 1994 some militia groups demobilized in a process promoted by the national government, but the demobilization was not completely successful. Many demobilized members associated in a “security cooperative” were assassinated, while others simply maintained their activities in the most marginal areas of the city, thus reducing their public importance, but still retaining their influence in the violent and criminal dynamics of the city (Ceballos and Cronshaw 2001, p. 125). Meanwhile paramilitary groups were becoming stronger, partially through the creation of private security groups known as Convivir that were created by then governor of Antioquia (and later President) Alvaro Uribe (Roldan 2003, Valencia 2007). By 1997, the project of paramilitary expansion consolidated at the national level (Romero 2003) and Medellin was an epicenter of the consolidation. Paramilitary groups engaged in a wipeout of militias and guerrillas, through a faction known as Bloque Metro. As paramilitaries consolidated their power, the boundaries between criminal and political violence became increasingly blurred: paramilitaries consolidated their power over drug trafficking activities without abandoning their antiguerrilla motivations, and the money and armed power of criminal groups became vital for paramilitary expansion. This co-dependence between criminals and paramilitaries is illustrated by the story of a gang known as La Terraza, which emerged in the Barrio Manrique of Medellin, grew while working for Pablo Escobar, and then was controlled by Murillo, who used it, among many other activities, to carry out assassinations contracted by paramilitaries against leftist intellectuals and leaders.141 The power of La Terraza grew, and its leaders                                                                                                                 141 This is the case of two academics and activists that worked for a very well recognized Jesuit NGO (CINEP) and were assassinated in 1997, an academic from the University of Antioquia, and famous political comedian Jaime Garzon. Semana “La Entrega de Don Berna”, 2005. Semana Edicion 1204, May 30. The emphasis of this dissertation on criminal violence does not negate the complex motivations behind   185   wanted to get more directly involved in the trafficking of drugs;142 they carried out attacks against Murillo in 2000 hoping to eliminate and replace him, but he survived and assassinated the leaders of La Terraza. The Oficina de Envigado, also commanded by Murillo, became the key criminal structure in the city, and established direct relations with the paramilitary Bloque Metro. By 2002 the Oficina de Envigado successfully mutated into a paramilitary faction known as Bloque Cacique Nutibara, commanded by Murillo, and allied with the Bloque Metro to defeat militias and guerrillas. After the success against the guerrilla and militias, these two paramilitary factions engaged in a battle for controlling Medellin143 that contributed to maintain frequent violence, and that was again won by Murillo. In sum, the illegal market after 1993 was highly competitive and fluid, and this explained the persistence of violence, as armed actors struggled to eliminate each other, and engaged in very unstable alliances. At the same time, the state also started to undergo processes of institutional change that reduced fragmentation and paved the way for more stable collaboration between criminals and state authorities, but also for more efficient enforcement actions that reduced criminals’ incentives to engage in public displays of violence.                                                                                                                 violence in Medellin, yet it is clear that without commanding both political and criminal groups, it was difficult for any armed actor to completely control criminality in Medellin. 142 Semana. “El Fin del Terror de Don Berna”, 2006. Semana, Edicion 1206, June 13. 143 Interview with former peace and security advisor of the Medellin’s Mayor Office. Medellin 26 October 2010.   186   4.2.2. The state and its transformations Two key aspects were crucial in reducing state fragmentation in this period: the increased coordination between the police and the military towards the end of the 1990s, and the consolidation of the paramilitaries’ corrupt networks among large sectors of the political class. This situation started to create interlocking incentives for armed actors in Medellin to evolve into using less visible forms of violence: although disappearances occurred before, they became more prevalent in the 1990s (Uribe and Vasquez 1995, p. 115). Similarly, dynamics of social cleansing became even more pronounced in the hands of militia groups that attempted to impose their own social order. Without a doubt the cleansing practices of militias, directed at anything they considered a threat, from drug consumers to bad neighbors (CNRR 2011) were very visible for the local communities, but not necessarily for the state and the general public, as had been the visible violence implemented by the Medellin DTO. In the mid 1990s, despite increasing competition, traditional political elites of the city managed to dominate elections (Arenas and Escobar 2000) thus contributing to more cohesion in the state, although in different occasions local mayors of Medellin opposed programs advanced by the national government to address the root causes of violence (Moncada 2011). The main source of cohesion within the state security apparatus was the collaboration between enforcement agencies, which consolidated with the implementation of joint military-police actions towards the beginning of the 2000s, aimed at defeating guerrillas (FARC and ELN) especially in areas where they consolidated power in the 1990s, such as the Comuna 13.144 In 2002, there were eleven military                                                                                                                 144 Comuna is a name given to subdivisions within the city that comprise several neighborhoods. There are 16 Comunas in Medellin.   187   operations in this part of the city (CNRR 2011, p. 76), and two of them –Operations Mariscal and Orión- were particularly notable because of their size, the type of arms deployed, and because police and Army forces operated jointly along with the judicial police DAS and the Attorney General’s Office; these operations unlike the urban military operations of the late 1980s, were highly coordinated. In an interview, the Army general who commanded these operations emphasized that teamwork had been key in all the joint operations against the guerrilla in Medellin.145 The successful coordination among enforcement agencies was also facilitated because the national government led by recently elected President Alvaro Uribe, and the city’s Mayor Luis Pérez Gutiérrez, jointly supported these military operations. The operations were successful in undermining guerrilla presence, but generated large human rights violations, and most importantly, were perceived by large sectors of the population as facilitated by paramilitaries, and instrumental in their consolidation (Angarita et al 2008) a perception that some paramilitary leaders later confirmed (CNRR 2011). In fact, the dark side of the cohesion that allowed effective coordination of enforcement in these large-scale operations was precisely that paramilitaries were receiving stable protection from important sectors of enforcement and the political class, and consolidating their domination of the city. As the power of the paramilitaries was consolidating, the increasing protection they received from large sectors of the local political class motivated them to prefer hidden violence such as disappearances (Oude and Rozema 2009) and to use mass graves to hide the remains of people. Paramilitaries used mass graves to hide victims of                                                                                                                 145 El Colombiano. 2003. “Desarticulada en Medellin Red Logística de las FARC: Trabajo conjunto ha sido clave” Entrevista con General Leonardo Gallego, 2003. El Colombiano, September 13.   188   mass disappearance in the Comuna 13 of Medellin.146 Hiding violence would become an even more important strategy to maintain peace during the unprecedented period of pacification between 2003 and 2007. One of the paradoxes of the increasing cohesion of state power in Medellin in the late 1990s is that the state became more sophisticated and efficient in confronting violence, although this also opened the way for more stable patterns of collaboration between state and paramilitaries. But in this period another crucial paradox surfaced: a city wounded by extreme violence also had become more active in confronting it (Baird 2012, Riaño 2006). As one community organizer expressed to me in an interview in the nineties one could see very strong levels of solidarity, of social organization. Grassroots efforts developed and another fiber for the region was created, that of working together. That process activated a society that examined itself critically, through norms and state support. Many groups were created in Medellín, such as musical collectives.147 These community efforts are important for two reasons: first, organized civil responses to violence seem more likely to emerge when cities experience extremely frequent and visible violence; as we will see in the next chapter, civil society efforts appear to be more limited when violence is not very visible, as in Cali or Culiacan. Second, these community efforts are crucial and do contribute to mitigating the impacts of violence, but seem limited to explain the dynamics of violence. Strong community efforts and programs persisted both during the pacification period, and during the subsequent return of violence, that will be explained in the next two sections.                                                                                                                 146 Evidence about these forms of violence became public as part of the demobilization process of paramilitary groups which started in 2002. El Colombiano. 2006. “Don Berna reconoció las desapariciones en la Comuna 13”, 2006. El Colombiano, October 6. 147 Author’s interview with community organizer in Medellin, October 27, 2010.   189   4.2.3. Youth gangs in the transition During the period of transition, the complex links between small youth gangs (combos) and larger armed and criminal actors continued. Youth were both the main victims of the cleansing efforts of militias, and later of paramilitary groups, and also were an attractive source of armed force for guerrillas, paramilitaries, and militias. Combos and bandas engaged in territorially bound disputes that created fronteras (frontiers) marking the localized limits of power within neighborhoods and between gangs. Interestingly, precisely because of this dynamics of “local wars” the Mayor’s Office started to promote local pacts aimed at reducing violence in particular areas of the city. Many pacts succeeded in reducing violence temporarily, but they were often unstable and short lived, to a great extent precisely because they did not acknowledge the strong links that connected youth gangs to larger criminal groups and to broader dynamics of conflict (Alonso, Giraldo and Sierra 2007; Riaño 2006; Velez 2001). This relation between gangs and criminals started to change towards the end of the 1990s: rather than contracting out violence, militias, traffickers, and especially paramilitaries, attempted to discipline and co-opt youth gangs. Because paramilitary structures in the city were relatively new and their knowledge of the urban space was limited, gangs’ familiarity with the geography was thus key for their project of urban consolidation (CNRR 2011, p. 60), and it was necessary for them to use gangs as armed force. But in order to counteract the influence that militias and guerrillas could have on gangs, paramilitaries tried to vertically integrate gangs into their structure and strictly controlled youth behavior. Paramilitaries recruited and sent members of gangs to military training camps (Riaño 2006, p. 202). Social cleansing of drug consumers and low level   190   criminals became more common, as well as norms controlling the times of circulation, dress codes, and drug and alcohol consumption of young people. In the words of a community organizer in the late 1990s “the bandas were crushed with blood, there were fliers distributed in convenience stores showing that there were others coming to exert social control.”148 The vertical integration, cooptation, and repression of gangs, consolidated and had a direct impact in reducing violence when one criminal actor effectively dominated the city after 2002. In other words, conflicts in the criminal market needed to disappear for this pattern of control to reduce violence, as occurred in the following period. 4.3. The criminal monopoly and the Medellin “miracle” (2003-2007) Since 2003 Medellin experienced unprecedented processes: an urban reform promoted by Sergio Fajardo, elected mayor of the city in 2003; the demobilization of paramilitary groups; and a reduction in violence that seemed impossible in a city with such a complicated story of violence and sorrow. In 2003 homicide rates decreased by about 81%, and reached a historic low of 34 homicides per 100,000 inhabitants in 2007. Such a stark transformation bewildered observers and inhabitants of the city, and was happily attributed to the peace process and the urban transformation in the city. Fajardo, introduced a large scale program of investment in marginalized areas of the city based on the premise that inequality and violence were closely connected, and thus urban reform could “close the entrance door for youth into criminality.”149 Yet, over time, and especially when violence returned a few years later, less flattering causes for the decrease                                                                                                                 148 Author’s interview with community organizer in Medellin, October 27, 2010. 149 Author’s interview with Sergio Fajardo, November 6, 2010.   191   became clear. Without denying the benefits of the urban transformation process, the reduction can also be explained by the consolidation of a criminal monopoly under the complete control of Diego Murillo, and by the cohesion within the state that provided the incentives to hide violence. Diego Fernando Murillo -Don Berna- illustrates perfectly the evolution and complexity of violence in Medellin. He started as a member and then enemy of the Medellin DTO, and successfully mutated into a paramilitary leader who consolidated a criminal monopoly by co-opting some gangs, eliminating rivals, and finally winning over an initial ally and the rival paramilitary faction, the Bloque Metro (CNRR 2011; Alonso, Giraldo y Sierra 2007). His success was the result of his ability to wage and win subsequent wars against criminal and political competitors. As one expert defined it, Don Berna “ was the winner of all the wars in the city.”150 The successful monopolization of criminality implied controlling different markets, from extortion to local drug distribution, but also imposing strict control over civilian population. The control Don Berna had on different types of criminality in the city questions arguments that explain upsurges in violence as a result of the diversification of criminal portfolios. In Don Berna’s case, such diversification was in fact a tool to accumulate more power, and in turn control, if not completely eliminate, violence. Don Berna commanded a tight chain of command and control. A clear manifestation of this criminal network’s power occurred in May 2005, when Medellin was paralyzed by a bus stoppage, ordered by members of Don Berna’s criminal organization as a reaction to an order to detain him, issued by the Attorney’s office. This                                                                                                                 150 Author’s interview with former peace and security advisor of the Mayor’s office, Medellin, October 26, 2010.   192   event made evident not only the intimidation power of criminals at the service of Don Berna, but also the control over the extortion market of transportation routes. Bus drivers agreed to stop out of fear. In a similar episode, in 2007 armed men at the service of Don Berna tried to persuade the principal of a local school to fill two buses with people so they could attend a judicial public audience, and support Don Berna.151 Another crucial element for establishing a monopoly was the tight control over gangs, the regulation of their behavior, and the brokerage of their localized conflicts. For Don Berna disciplining the violent behavior of other armed actors in the city was crucial “with the subordination of bandits, nobody could move without Berna’s permission; there was a subjugation of bandits.152” 4.3.1. Disciplining criminal gangs Don Berna’s success in controlling violence depended on his ability to regulate the violent behavior of youth gangs, rather than on outsourcing it (IPC 2003, Ramirez 2005). Such control was based on the strict disciplining of the young population in marginal areas and on the regulation of their behaviors (IPC 2003, p.162, 170; IPC 2004, p. 223). Rather than prohibiting drug use, that was one of the markets don Berna controlled, his people controlled where and how drugs could be used. In an interview to the NGO IPC, a young man asserted that “Paramilitaries in many occasions warned young people that we cannot drink alcoholic beverages or use psychoactive substances in certain places, but at the same time they indicate where they can do it; in other words,                                                                                                                 151 Semana.  “Los Secretos de Don Berna” 2007. Semana, Edición 1315 July 16.     152 Author’s interview with security analyst and social worker of the Mayor’s Office, Robledo neighborhood, Medellin, October 28, 2010   193   there is control not only based on immediate violence, but in a consolidated power that is well-known to the youth” (IPC 2003, p. 170). Combined evidence of interviews and human rights reports suggests that paramilitaries and Don Berna used “exemplary” punishments on those who did not follow their rules, for example obliging them to remain for hours in sewers, forbidding families from approaching them while the punishment lasted (Personeria 2005) and increasing forced displacement within the city. Don Berna and his organization expelled those who did not fulfill the requisite of not using violence unless it was authorized by the high commands. A local police commander told me in an interview “The power don Berna had could be inferred from his power in prison. He ordered not to kill, and nobody killed. I met a kid who knew who killed his mother but he did not do anything because he knew he could be killed as well.153” During this period, sexual violence increased, perhaps reflecting its strategic use as a punishment and intimidation tool for gangs; it was not likely to generate massive reactions or to be very visible due to the often contradictory attitudes of the population towards sexual violence (Riaño 2006).154 The ability of Don Berna to regulate gang behavior in Medellin highlights that the mere presence of gangs, or the existence of large young populations does not determine the involvement of youth on violence and criminal activities. It also underscores that the mechanisms that lead to a reduction of gang criminal activity may sometimes hide dangerous dynamics, such as manipulation by powerful criminals. It is interesting that                                                                                                                 153 Author’s interview with commander of police station, Medellin, October 23, 2010. 154 Wood (2006) has noted the enormous variation in forms of sexual violence that may appear during wartime; strategic sexual violence maybe one of those forms that “appears to occur when an armed group believes it to be an effective form of terror against or punishment of a targeted group.”   194   despite the tight grip that criminals exerted on social behavior, the reduction of violence sometimes led communities to accept this control, thus highlighting the complex interaction between criminal behavior and its social acceptance. For example, Don Berna had a reputation for settling disputes between gangs, which led him to use the nickname of Adolfo Paz (Adolf Peace) for the negotiation that preceded paramilitary demobilization. Not surprisingly, as Oude and Rozema (2009) note, social imageries about Don Berna mixed both fear and admiration for his ability to bring peace. Of course, the overall reduction of violence in this period (which is undeniable even when considering the hidden violence at work) brought an enormous relief to the population, but the criminal arrangements that sustained the equilibrium became the basis of the return of violence. 4.3.2. Controlling and hiding violence: the strategic coincidence of interests to maintain peace The cohesion in the state security apparatus that started in the late 1990s was consolidated by 2003, and an additional factor emerged to contribute to this cohesion and facilitate the pacification of Medellin, the initiation of a peace process between paramilitary groups and the government of President Alvaro Uribe in 2002, right after Uribe took office. The negotiation was materialized with the signing of a demobilization agreement in July 2003. Don Berna became a paramilitary leader and entered the peace process. In November 2003, as commander of the Bloque Cacique Nutibara he   195   demobilized along with 868 men, many of which were members of local combos rather than paramilitary soldiers (Balbin 2004, HRW 2010, Vivanco 2004).155 Alvaro Uribe presented Medellin as the first test for the peace process, and in fact, the city received the largest concentration of demobilized soldiers in the country, 3,270 by 2007 (Alonso and Valencia 2008). Mayor Sergio Fajardo, the first independent politician elected in the city, was not officially part of the peace negotiation or a member of Uribe’s political coalition, but decided to support the process that he defined as an opportunity to take advantage of “an exit door from delinquency”, even though he did not really know the terms of what had been negotiated.156 This situation created a strategic concurrence of interests that eliminated conflicts between the national and local governments: the local government did not question the terms of the negotiation, and the national government did not intervene in the local government’s security initiatives.157 The success of the process depended on the collaboration of local and national government: the national government had the political control of the process, but the local government was in charge of the initiatives aimed at the reinsertion of the demobilized soldiers (Alonso and Valencia 2008). The strategic concurrence of interests within the government agencies also facilitated the strategic concurrence of interests between state and criminals: Don Berna benefited from maintaining peace in Medellin because that would allow him to keep the                                                                                                                 155 The demobilized men were concentrated for three weeks in La Ceja, a municipality near Medellin, and then they returned to their neighborhoods of origin. 156 Author’s interview with former Mayor Sergio Fajardo, Bogotá, November 6, 2010. 157 By contrast to Fajardo’s decision, Luis Eduardo Garzón, mayor of Bogotá at the time decided to leave the responsibility of demobilization on the hands of the national government and did not engage with it. His relationship with the national government was then more conflictive on this issue (Author’s interview with former peace and security advisor of the Mayor’s Office, Medellin, October 26, 2010).   196   legal benefits of demobilization; at the same time, even though evidence that Don Berna remained engaged in illegal activities persisted, national and local authorities also converged on the need to have a successful peace process an thus initially condoned some of Berna’s behaviors. In the words of a local police commander in Medellin “The order was not to follow them much, because there was a pact between the President and the Mayor’s Office. There was political resistance to bother them, maybe not because there were pacts, but because there was a peace process ongoing.”158 In several occasions the national government protected the demobilization of Don Berna in the face of increased evidence of his criminal activity. In 2005 don Berna was accused of ordering the assassination of a local politician in the northern department of Cordoba, and the government ordered his capture; as Berna handed himself in to authorities, his demobilization process continued. Then, in response to an extradition request on drug trafficking charges made by the United States in 2004, the Colombian government announced that Don Berna would not be extradited if he continued collaborating with the peace process and abandoned his illegal activities, a decision that was not welcomed by the U.S but that guaranteed the continuation of the demobilization. Don Berna was transferred from the demobilization camp in the north of Colombia, to a security prison near Medellin. In 2007, with recurrent reports of criminal activity, Don Berna was transferred into another high security prison, farther away from his area of operations. Finally, under extreme pressure from the U.S, the government finally extradited Don Berna in May 2008. The key point of this chain of events is that the government tried hard to maintain the process as it was delivering results in terms of                                                                                                                 158 Author’s interview with police commander, Medellin, October 25, 2010.   197   reducing violence in Medellin. The state’s credible promise of not prosecuting criminals within the peace process, facilitated by the cohesion of the state security apparatus, was the key mechanism to maintain the peaceful equilibrium in Medellin. In fact, some people have colloquially termed this relation “Donbernabilidad” referring to the power of don Berna in setting the terms of governance in the city. The pacification of Medellin was real as reflected in homicide rates, yet there is evidence that when violence was necessary it was carefully hidden. Instances of collective violence decreased by 69% between 2002 and 2003 (IPC 2003) while reports of forced disappearance and instances of homicide with blunt objects, suffocation, and knives, increased in the period. According to communal leaders, such methods reflected an interest to hide the responsibility of those committing those crimes, as these methods tend to be quickly dismissed as associated with common delinquency or interpersonal disputes (IPC 2004, p.144). Between 2003 and 2006, the proportion of homicides committed with knives more than doubled from 9.3 to 22%159 and many of these homicides were committed by paramilitaries who wanted to divert attention and elude responsibility on violence (CNRR 2011). During this period of pacification a crucial political paradox emerged, as electoral competition was very high but there was also a strategic concurrence of interests. New civil political forces (represented by Mayor Fajardo) were becoming stronger at the same time that some traditional forces and other emerging politicians were co-opted by paramilitaries. The large-scale alliance between paramilitaries and politicians was carefully documented by political analyst Claudia López (Lopez 2010) and became the                                                                                                                 159 Personal calculations based on DANE statistics.   198   basis of a judicial process that linked 139 congressmen nationally and 22 in Antioquia to paramilitary groups.160 The paradox is that the competition between these two large political sectors –the new civil movement, and the co-opted politicians- did not fragment political action in the city. Each faction has enough political resources to control the city, thus creating a situation whereby protection rackets could be easily established when the corruptible faction is in power because once in power, networks have capacity to work cohesively, but protection is likely to be unstable because the other faction can gain power back, as the two political sectors alternate in power. For example in 2011 the independent network led by Fajardo regained its power back from another network that was closely linked to the parapolitica scandal (Avila and Velasco 2012). In sum, the security apparatus became significantly less fragmented in the pacification period due to the coordination of enforcement agencies and levels of government, and despite high levels of political competition. 4.4. Breakdown of the monopoly and reemergence of violence (2008-2011) The monopoly of criminal power in Medellin started to be under internal and external pressure since July 2006. On July 3rd, Gustavo Upegui, a sports entrepreneur and politician, widely known as a leader of the Oficina de Envigado, was assassinated, and his death was followed by the disappearance of one of Don Berna’s closest collaborators, indicating power struggles within the criminal market. With don Berna’s extradition in 2008, struggles evolved into criminal disputes for the control of the city, which produced                                                                                                                 160 It is also calculated that in total 943 politicians have been involved in the parapolitica scandal between 2007 and 2012, including 41 former mayors, and 20 governors. Verdad Abierta 2012. “Cinco años de parapolítica” Available from: http://www.verdadabierta.com/Especiales/cinco-anios-parapol/parapolitica- 2012.swf [3 October 2012].   199   a notable –and certainly unexpected for many leaders and observers in the city- increase of homicide rates that doubled from 45 in 2008 to 94 in 2009. The first months after Don Berna’s extradition were followed by quick re- accommodations and disputes: some of his successors were captured, killed, or handed themselves to authorities. By 2009, two former sicarios, Sebastián and Valenciano, made it to the upper line of confrontation for power, and the dispute between the two leaders was one of the main drivers of violence between 2009 and 2011. The power of these two criminals was significant, but not comparable to the power of former criminal bosses of the city. A security expert asserted that delinquency today “is very violent and not very organized, before it was very violent and very organized.”161 When violence reemerged in Medellin, national authorities quickly associated it with a struggle for controlling local drug markets (microtrafficking) in the city, and emphasized the criminal nature of violence. These efforts to depolitize the nature of violence in the city, which have been criticized by many activists and experts, once again illustrate the blurred boundaries between political and criminal violence.162 While depictions of violence in the 1990s tended to decriminalize violence, ignoring the increased engagement of political armed actors in drug trafficking, more recent depictions depoliticize violence, ignoring how actors in dispute today are closely linked to previous dynamics and actors of violence. In fact, some of the new actors fighting for control in the city since 2008 derived from structures known as neoparamilitaries or                                                                                                                 161 Author’s interview with former peace and security advisor of the Mayor’s office. Medellin, October 26, 2010. 162 It is beyond the scope of this work to discuss the nature of the groups operating in Medellin, but several reports show that the political dimensions of their work, such as killing social leaders, are still very much present (HRW 2010).   200   BACRIM (criminal gangs), formed by paramilitaries that never demobilized, some demobilized soldiers, and new recruits (HRW 2010, MAPP-OEA 2006). Two of these groups, the Urabeños and the Paisas, which emerged outside Medellin, were also trying to enter the city and achieve control over its criminal markets. Criminal markets besides drug trafficking became more prominent, including local trafficking, extortions, stealing of gasoline, and control of gambling. As illustrated in the previous sections, these markets were not new in the city, although they became more consequential for criminal income. The key aspect of the reconfiguration of illegal markets during the return of violence was that different markets were indeed linked to larger trafficking groups and activities, but unlike in the previous period of monopoly, they were highly contested between the factions vying for control.163 One illustrative case was extortion in public transportation. Extortion was a very old activity in Medellin. Under the former criminal monopoly, public transportation functioned as an effective protection racket: transporters paid their quota and they were effectively protected from delinquents. But since 2010 assassinations associated with this illegal activity proliferated (IPC 2011). According to observers from a very reputed NGO in Medellin, the difference was that as the monopoly of criminality fragmented, no single “protector” could guarantee safety for transporters, and thus, a single bus line could be extorted and threatened by very different actors.164 This example is useful to understand why the diversification of criminal markets is not a source of violence in itself; violence is rather a                                                                                                                 163 This assessment was confirmed in many of my interviews with experts, local cops, international organizations, and local Ngos in Medellin. 164 Interview conducted by Arturo Alvarado, Alberto Concha and the author with NGO focused on violence and conflict issues, conducted as part of a project on juvenile violence in Latin America, coordinated by Professor Arturo Alvarado at the Colegio de Mexico (Medellin, August 24, 2011)   201   product of the type of control exerted on criminal markets and of the connections established between different criminal activities. Controlling local markets was not only a source of income but also a way to gain territorial control. According to a local analyst and social worker, Sebastián and Valenciano competed on the basis of how many combos they controlled, which in turn determined their ability to control illegal commerce, territories, and access points to key strategic drug trafficking routes in the Uraba port and in the drug producing area of the Bajo Cauca.165 4.4.1. Silent traffickers in Medellin The return of violence in Medellin drew concern among different government sectors and the public opinion, yet it was far from causing the same concern that violence caused in the 1990s, and it was quickly diagnosed by some national authorities as the result of very localized disputes for local drug markets rather than as threat to national security. This difference in perception, likely reflected that the violent disputes emerging in 2008 were considerably less visible than those of the 1990s. In the words of a local analyst “one sees an attempt to regain leadership. The difference is in the method for applying violence.”166 Violence, although widespread and deadly, did not create the same level of threat for the wider public of Medellin and of Colombia. As another practitioner in violence prevention programs stated “Deaths nowadays are more selective, more focused. Before we felt involved as possible victims, but there is more indifference                                                                                                                 165 Author’s interview with security analyst and social worker of the Mayor’s Office, October 29, 2010. 166 Author’s interview with local expert, Medellin October 25, 2010.   202   today167” a statement echoed in similar terms by a local councilwoman “the situation today is not as bad, because of what was done back then [in the 90s] the Mafiosi changed their profile. One does not feel scared of going to 70th Street as before. The problem of the narco can even be worse now, but the fact is that the bombs made it a more aggressive period.168”   In this period, attacks against high-level public officials were not common, while methods hiding violence such as disappearances and the destruction of corpses proliferated. Local authorities reported that corpses were increasingly found in the stream of the Medellin river: 48 bodies were recuperated from the river in 2010 only. According to an investigator of the attorney’s office, “criminal organizations throw the bodies with the intention of disappearing them. Thus the evidence, and the possibility of getting to the perpetrators, disappears.”169 Different human rights organizations also reported the increasing use of brutal methods to hide the evidence of violence; one human right worker interviewed in 2011 reported that disappearances were increasing in the Comuna 8, with “terrible methods that eliminated all remains, such as incinerations and the use of machines to destroy the bodies.”170 Similar stories emerged in conversations I had with several members of a human rights group in the Comuna 6; according to one of them “[criminals] torture, murder, chop, burn. People leave because they have been intimidated, there is fear about denouncing because of the infiltration in the police.171”                                                                                                                 167 Author’s interview with violence prevention practitioner, Medellin. October 26, 2010. 168 Author’s interview with local councilwoman in Medellin, October 28, 2010. 169 El Espectador. 2011. “El Rio, la fosa más grande de Medellín” El Espectador, February 8. 170 Interview conducted by Arturo Alvarado, Alberto Concha and the author with Ngo in Medellin, Medellin August 22, 2011. 171 Author’s conversation with members of the Human Right’s Association Comuna 6, Medellin, October 29, 2010.   203   The lower visibility of violence in this period reflects a strategic adaptation of criminals to a state that remains cohesive especially through the more coordinated action of enforcement agencies. As explained in Chapter 3, in Colombia traffickers seem to have learned that when they gain too much attention from the state, they can be more quickly captured. In fact, Sebastian and Valenciano had long criminal careers that dated back to the Pablo Escobar years, but they survived with a relative low profile that did not attract police attention; they were captured relatively quickly after they were identified as the leaders of criminal disputes in Medellin: Valenciano was captured in November 2011, and Sebastian in August 2012. 4.4.2. Revitalizing the combos Sebastián and Valenciano had a long experience with criminality in the city, but not the symbolic power or control over armed force that their predecessors had. Thus, for them outsourcing violence in combos was crucial to wage war and gain control over criminality. As in previous periods, not all combos worked for larger criminal structures, or not all of them were associated with larger criminal organizations in the same way, but the outsourcing carried out by the warring criminal groups marked competition and territorial disputes. According to a security expert and Ngo worker, Valenciano’s relation with combos was “what combo do you represent? You are accountable for everything [in your area] not all of them are aligned, the idea is “you pay me for microtrafficking, for extortion” is a criminal enterprise. But not all of them are part of the same payroll. It’s easier to contract a pelado [young kid] for a job than having him in the payroll, but it is   204   not total control. In fact there is now another señor [guy, criminal] identified.172” In a similar statement, a security analyst and social worker of the Mayor’s Office explained that in many cases the members of combos do not know who they work for “today the narcos are not known. The kids do not know who Sebastián and Valenciano are, but in the oficina they are told who they work for.”173 These statements illustrate the dynamics of outsourcing and the contrast with the previous period: gangs were contracted to carry out violence and there was a hierarchy that associated them with the larger criminal disputes in the city, yet this hierarchy was not strong enough to impose military discipline over gangs. Outsourcing to youth gangs coupled with competition in the criminal market revitalized territorialized disputes within relative small territories that mark the boundaries of a combos’ control area, similar to the “local wars” during the consolidation of militias. These localized wars have led to the creation of what local residents and experts call fronteras invisibles [invisible frontiers], the lines that demarcate areas of control between combos, and that severely limit mobility in local communities. A remarkable consequence of the invisible frontiers has been an increase in intraurban displacement: between 2008 and 2011 reports of urban displacement skyrocketed and victims totaled 17,912 by the end of 2011 (Personeria de Medellín 2011) reflecting that many people left their neighborhoods as a response to threats and territorial conflicts between combos. According to a human rights defender of the Comuna 6 “One cannot move to another neighborhood because one can be killed; those who say that criminals                                                                                                                 172 Interview conducted by Arturo Alvarado, Alberto Concha and the author with Ngo in Medellin, Medellin August 22, 2011 173 Author’s interview with security analyst and social worker of the Mayor’s office in the Robledo neighborhood. Medellin, October 28, 2010.   205   are only killing each other are lying, there are always “astray bullets.” There are many who die by mistake. Overall, the disputes between combos have left more civilian population victimized.”174 Territorial disputes thus illustrate the problems associated with outsourcing: local armed groups and gangs acquire better equipment and money, and engage in competition with each other, and those contracting out their services do not have the ability or interest in controlling them.   The recent period of violence was less intense than the “perfect storm” but yet very high when compared to the national average, and with the reality of many other cities in the world. Criminal groups were probably more fragmented and weaker than in previous periods, but they were not detached from national dynamics and from the international drug trade. They preferred to use less visible, and very localized forms of violence, because they have learned that once they become too visible for authorities, especially for national ones, they can be captured quickly. It is possible however, than if criminals want to expand and consolidate their power they may start deploying visible forms of violence, because as we saw in Chapter 1, visible violence is an instrument to show power. Some events occurring in Medellin in 2012 suggest that this can be the case with some of the criminal groups that have killed and conducted massacres. 4. 5. Conclusion The history of violence in Medellin illustrates the complexities and contradictions that surround the understanding of drug violence, especially when it overlaps with political violence: armed actors that mutate and change sides, political motivations that                                                                                                                 174 Author’s interview with local human rights NGO. Medellín, October 29, 2010.   206   overlap with criminal ones. The theory developed in this chapter to understand violence in Medellin does not pretend to reduce this complexity, but it shows that in the midst of it, dynamics of war -not necessarily the motivations of violent acts- are inextricably linked to the dynamics of control over illegal markets, and by the interactions between those controlling such markets and the state. Table 4.2 summarizes how different interactions between state security apparatus, criminal market, and criminal armed force created different configurations of violence in Medellin, from the perfect storm of high frequency-high visibility violence, to the unstable pacification and low frequency-low visibility violence. As summarized in Table 4.2, the criminal market in Medellin has experienced both moments of monopoly with low frequency violence, and moments of competition with high frequency violence. Likewise, the structure of the state security apparatus has shifted from very fragmented to more cohesive, and this in turn, explains why and when drug violence has become more visible. In a context where state protection becomes unpredictable, criminals have an incentive to use visible violence that can help them signal their power vis a vis rivals. Not using violence would not guarantee that law enforcers do not prosecute criminals because while state officials may be protecting criminals, others are confronting them, and the protectors’ ability to provide credible commitments is low. Thus, when the criminal market is more competitive and the state more fragmented, the High Frequency - High Visibility (HF-HV) and the “perfect storm of violence” follows. In this context, violence has been unusually high in Medellin also because traffickers, have, at different moments and rates, outsourced their armed coercion to youth gangs.   207   Table 4.2. A political approach to the evolution of drug violence in Medellin Period Criminal Market State Security Type of armed force: Apparatus Relation criminal actors- gangs 1984-1993 Competitive: Fragmented: Outsourcing: Medellin DTO Deployment of military Gangs proliferated and dominated, but militias, in urban antinarcotic grew at the service of guerrillas, also operated operations generates traffickers. Common conflicts. delinquency became Lack of coordination more violent as gangs between national and acquired arms and local government. money. Fragmentation of local There were some rules authority. of behavior, but Increased political traffickers did not competition. regulate gang behavior strictly. 1994-2002 Competitive: Increased cohesion: Transition to insourcing: Militias, trafficking Towards the end of the Attempt to co-opt gangs groups and oficinas, period state successfully but also regulate their guerrillas and coordinated military- behavior (especially paramilitaries. police action. paramilitaries). 2003-2007 Monopoly: Cohesion: Insourcing and complete Under the leadership of Incentives created by control of gangs Don Berna. Successful peace process eliminate control of different conflicts between criminal activities and national and local elimination or control of government. criminal competitors. Coordination between enforcement agencies 2008-2010 Competitive: Cohesion: Outsourcing: Successors of Don State remains relatively Criminal actors use Berna strive for control; cohesive in enforcement combos as their tool for reconfigured agencies. control but do not have paramilitary groups. Political alternation enough power or prevents consolidation willingness to discipline of state protection them. rackets. The dynamics of violence in Medellin have been undoubtedly framed in, and influenced by, larger national dynamics of conflict. But this national influence does not mean that the specific dynamics of violence in the city can be explained by power dynamics at the national level; to understand why the trajectory of violence in Medellín differs so markedly from the trajectory of its country counterpart Cali, we need to consider power relationships at the local level and how local state and criminal actors   208   interact. By doing this, I have challenged conventional understandings of criminal power, for example, that Pablo Escobar was an almighty criminal. While his power was undoubtedly enormous, and his control of the illegal market large, he did not have a monopoly of violence and illegality in Medellín as it is often assumed. Several insights derive from analyzing the extreme fluidity of violence in Medellin. First, trajectories of violence might be path dependent but not necessarily cumulative or self-enforcing. Path dependent, because as I described throughout the chapter, today’s criminal actors have stories that connect them to former periods of violence and former criminal actors and it is not possible to understand violence today without understanding the history of violence. Non -cumulative or self-enforcing, because trends can change rapidly and reverse: the peaceful equilibrium during Don Berna years eroded quickly as his power faded-away. The trajectories of violence can also be non-cumulative because rapid changes can occur in the distribution of state power, for example through re-accommodations in the relations between different enforcement agencies and different government levels. This is a crucial insight because we usually think of state power as relatively static and separated from political dynamics. But state power, can in fact be affected by political dynamics. Finally, the trajectories of violence can be non-cumulative because violence can persist despite positive institutional transformations and strong civil society responses to violence. A crucial debate in Medellin revolves around how to assess the urban transformations the city underwent between 2004 and 2007. On the one hand, it would be impossible to deny the positive impact that the city’s transformation had both on the perceptions about Medellin and in the services provided in the city. An analyst and social   209   worker from the Mayor’s office declared “the city is like the Phoenix, today there is a transformed city, that has been humanized in a certain way. Before you could be in any place and you could fall. When could you imagine that tourists would come to Medellin? Nobody thought that people could believe in Medellin.175” On the other hand, the return of violence casts doubt on the impact of urban transformations. As a councilman argued “[violence] is not only lack of opportunities, it is a problem of authority and culture. For example, in the area of the cable car, which is one area that has had more social investment, is very unsafe today.”176 The fact that the ceiling for violence in the recent violent peak was high, but comparatively lower than in the 1990s, may reflect the positive impact of urban transformations on violence. While my argument does not negate the importance that social investment and infrastructure have, and their potential impact on reducing violence, it highlights that they are insufficient to explain patterns of violence consistently. A second insight derived from analyzing Medellin is that explaining the causes of peace is as important as understanding the causes of violence, especially because analytical attention is usually skewed towards explaining extreme violence. Furthermore, Medellin shows that we cannot assess criminality by looking at violence levels and that apparently low violence should not be equated with the absence of criminality, because peaceful outcomes, often result from the dominance of strong criminal powers. A third insight is that the relations between youth gangs and criminal organizations are central to understanding violence; yet, violence is not the automatic result of the existence of gangs or the existence on large young populations. Gangs can                                                                                                                 175 Author’s interview with analyst and social worker from the Mayor’s Office, Medellin October 27, 2010. 176 Author's interview with local councilman, October 25, 2010.   210   sometimes exist without being associated with violence and criminality and other times criminals can use them to conduct violence, as has been the case throughout most of Medellín’s history. In rare occasions criminals can attempt to strictly discipline gangs in order to deter any violent behavior, as occurred during Don Berna’s monopoly. Analyzing how and when these different relations take place, provides important insights to better understand urban violence. Finally, the case of Medellin suggests that extreme, and especially, visible violence, can create a paradox as the need of both civilians and state actors to respond to violence can strengthen civil society responses to the causes and consequences of violence, and force the state to find more creative responses to violence. It is undeniable, for instance, that in its long experience with violence Medellin has more government programs dealing with youth,177 violence, and conflict resolution than any other city in Colombia, and more communal and civil organizations experienced in dealing with vulnerable and at risk populations and victims. This positive side effect of violence is not as noticeable in other cities, even when they have experienced persistent violence, and this is at least in part consequence of not experiencing very visible levels of violence. Cali and Culiacan illustrate this situation clearly and they are the subject of next chapter.                                                                                                                 177 As of 2011, there were more than 100 local programs for attention to youth in Medellin. The local budget for youth programs was about 3.7 million dollars and the local office for youth programs (Metrojuventud) had a larger budget than that of the national office for youth (Colombia Joven).   211   CHAPTER 5. THE “BUSINESSMEN” AND THE “GENTLEMEN”: LOW VISIBILITY VIOLENCE AND STATE-CRIMINAL SYMBIOSIS IN CALI AND CULIACÁN If an external observer were asked to identify the most violent city in Colombia and Mexico, the response would probably not be Cali or Culiacán. These two cities, despite having long histories with violence and drug trafficking, have not been at the center of concern by national or international actors;178 despite having homicide rates that have often surpassed national averages, these cities have not witnessed the civil and institutional responses against violence –however imperfect- that their country counterparts like Medellín or Ciudad Juárez have generated. Furthermore, despite the persistence and deadliness of violence, wide sectors of the population perceived some drug traffickers in these cities, especially older ones, as “gentlemen” who respected “honor codes” in their use of violence. These responses and perceptions of violence and traffickers reflected the prevalence of frequent but low visibility drug violence. How can we explain these strikingly similar histories of violence in two cities imbued in completely different national contexts, with different political and social histories? In this chapter I examine how the same factors – a cohesive state, a competitive market and the insourcing of violence- explain the similarity.                                                                                                                 178 It is important to note that in the early 1990s, and thanks to the leadership of the mayor of Cali, physician Rodrigo Guerrero, the city became internationally known for applying an epidemiological approach to violence and a preventive framework for citizen security, that was then promoted through the Interamerican Development Bank and other international institutions, but that was abandoned when Guerrero left office (Guerrero got reelected as Mayor of Cali in 2011). For a detailed discussion of these preventive security frameworks see Concha Eastman, Espinosa and Guerrero (2002) and Moncada (2009, 2012).   212   The core argument of this dissertation is that variation in drug violence results from the interactions and power relations between the structure of the state security apparatus and the structure of the criminal market. The visibility of violence is determined by whether the security apparatus is cohesive or fragmented, and the frequency of violence is determined by whether the drug market is monopolized or competitive. In this framework, the prevalence of low visibility violence in Cali and Culiacán can be explained by the existence of a cohesive state symbiotic with criminal organizations, which allowed impunity in the perpetration of violence but also created incentives for criminals to hide violence, as visible violence could force an otherwise friendly state to attack them. Although criminal markets in these cities were at times monopolistic they were mostly characterized by competition, and violence became more frequent as the criminal market became more competitive. Violence did not experience extreme spikes179 because for the most part criminal organizations insourced their use of violence. The histories of the two cities are framed in different political and social contexts: Cali has been affected by the broader dynamics of armed conflict in Colombia, while political violence has not been widespread in Culiacán; the politics of Culiacán cannot be understood without considering the broader history of the decades-long hegemony of the Revolutionary Institutional Party (PRI) in Mexico. Furthermore, the cities also differ in aspects such the extent of urban transformation: while Cali has been subject to intense immigration flows that shape the dynamics of social inclusion and exclusion,180 migration                                                                                                                 179 I operationalize spikes as homicide rates that surpass the national homicide rate by three times or more. 180 Since the 1950s Cali has been subject to intense and diverse migration flows (Moncada 2009) that derive both from the mid-century industrial expansion and also from displacement derived from the armed   213   flows to Culiacán tend to be seasonal, responding to labor demand in agriculture, and population growth is not as sharp as it has been in Cali (see Table 5.1. for basic socio- economic statistics). Yet besides being home to powerful criminal actors, the two cities share important features such as the prevalence of strong landowning and agroindustrial elites, relatively weak civil societies, and most strikingly, similar configurations of violence that contrast sharply with those of their country counterparts, and that can be understood through the particular interactions between state and criminal actors. The chapter is divided in two sections. The first analyzes the history of violence in Cali, the capital of the Valle del Cauca department (state) in the Pacific coast of Colombia, a city that was first dominated by the Rodriguez Orejuela brothers, leaders of the Cali Drug Trafficking Organization (DTO). I explore how the extensive network of corruption created by the Rodriguez Orejuela, allowed by the existence of a cohesive state security apparatus, created incentives for them to hide violence. Then I show how after the Cali DTO was dismantled, disputes emerged among its heirs, thus making the criminal market more competitive. Violence became more frequent, but for the most part remained with low visibility due to the pervasive protection criminals founds in the local state apparatus.                                                                                                                 conflict. The migration of black populations is particularly prominent and Cali is the urban center with the largest population of people of African descent in Colombia. This population in turn is concentrated in some of the poorest areas of the city (Moncada 2009, Barbary 1999).   214   Table 5.1. Cali and Culiacán: Basic statistics Cali Culiacán 1985 2000 2010 1985 2000 2010 Population 1350565 2236709 2244536 560011 745537 858638 Pop. growth 65 % 0.34% 33% 15% Pop. 10-29 years 43% 37% 50% 36% Unemployment rate 21.3 12 2.6% (2002) 6.1% (2008) Gini 0.54 0.517 0.43 (2005) (1998) % GDP Agriculture 0.16 0.33 12.3* Industry as % GDP 20 22 8.1* Sources: Author’s elaboration, with data from DANE, INEGI, Alcaldía de Cali, UNDP Mexico. * Based on state GDP The second part of the chapter revisits the history of violence in Culiacán, the capital city of Sinaloa in the Pacific coast of northern Mexico. In Culiacán, as in Cali, the existence of a cohesive state that allowed the consolidation of symbiotic relations between Sinaloan traffickers and state officials, determined the prevalence of low visibility violence. The competitive tendencies among trafficking leaders in Sinaloa contributed to maintain highly frequent violence. I explore how in 2008 violence became both more frequent and more visible, and Culiacán ranked among the three most violent cities in Mexico, as the trafficking organizations in the city experienced the most serious split in its history, and the security apparatus became more fragmented due to the urban deployment of the military by the federal government. I explore how despite these notable changes, the Sinaloa DTO still managed to control the market in Culiacán, insourced violence, and operated within a less fragmented security apparatus than that of other violent cities like Tijuana or Ciudad Juárez. Table 5.2 summarizes the trajectories of violence in Cali and Culiacán and emphasizes that in both cities low visibility but frequent forms of violence have prevailed.   215   Table 5.2. Trajectories of Drug Violence in Cali and Culiacán 1984 2010 Cali LF-LV HF-LV Culiacán HF-LV HF-HV LF: Low Frequency, HF: High Frequency, LV: Low Visibility, HV: High Visibility 5.1. Cali, Colombia 5.1.1. The emergence and demise of the Cali DTO (1984-1994) Cali is the third largest city in Colombia, and by 2011 its population was 2.294.643 inhabitants. In the early 1980s the city and the region witnessed the consolidation of mafias associated with drug trafficking that had a strong nucleus in Cali led by Miguel and Gilberto Rodriguez Orejuela, but that also extended to, and associated with, other regional nucleus in the department of Valle (Betancourt 1998, p.132). Over time, the Cali nucleus came to be known as the Cali DTO and spread a reputation as “businessmen” that preferred not to engage in violence. Such reputation became stronger as the Medellín DTO and its leader Pablo Escobar declared a frontal war against the Colombian state and against the enforcement of an extradition treaty; the Rodriguez Orejuela brothers repeatedly refused to engage in this war. But they did not renounce the possibility to use violence altogether; when it was necessary, they simply deployed violence differently, hiding it (Chepesiuk 2003, p. 63). Figure 1 shows homicide rates between 1984 and 2010,181 and represents the evolution of the frequency of violence in Cali. It shows how homicide rates were closely                                                                                                                 181 I use overall homicide rates as the main proxy for the frequency of drug violence. Of course it would be simplistic to assume that all homicides in a city are drug related. But these data has been triangulated with a   216   below the national average, thus they were not very high, but not low either, during the period in which the Cali DTO controlled the city. During the 1980s, the Cali DTO maintained a market monopoly despite the presence of other non-state armed actors in the city. These actors included urban guerrillas, militias derived from guerrillas, and self- defense militias whose presence dated back to the establishment of demobilization camps in the city in 1982. These camps were part of a peace negotiation between the guerrilla groups FARC, EPL and M-19, and the government of Colombia’s President Belisario Betancur. Until the peace process broke in 1985, the demobilization camps had become a training ground for gangs and militias, in a pattern similar to that occurring in Medellín in the same period (see chapter 4). Yet unlike Medellín, in Cali the presence of these groups was less pervasive and numerous and thus it did not challenge the monopoly of power of the Cali DTO.                                                                                                                 wide range of sources (my own dataset on violent events, reports, and interviews) to create a realistic picture of drug violence. This triangulation suggests that even though not all homicides in a city are drug related, the ebbs and flows of homicides are clearly connected to dynamics in the illegal market where criminal organizations have stable historical presence. For an analysis of the influence of organized criminal actors on homicide rates in Colombian cities see “Caracterización del Homicidio en Colombia 1995-2006”, Vicepresidencia de la República (2009)   217   Figure 5.1. Evolution of homicide rates in Cali Source: Author’s elaboration, with data from DANE and Policia Nacional The characterization of the Cali DTO as a cartel has been criticized as a misrepresentation of an organization that in practice was networked (Kenney 2007) and where leadership was dispersed at least among four main characters: Gilberto and Miguel Rodriguez Orejuela, Jose Santacruz and Pacho Herrera. In fact, in an infamous interview to the Time Magazine in 1991, Gilberto Rodriguez called the Cali Cartel “a poor invention of General Jaime Ruiz Barrera.”182 Yet, if the organization was indeed an invention, then it would not be possible to explain the dense support network that it established in Cali, and that for a long time shielded its members from state action. In the words of a member of the Cali Search Bloc, a group created in the 1990s to dismantle the Cali DTO, “the cartel was a powerful and well structured organization, they had neutralized the authorities. The technical part [of the operation against them] was very complicated, any movement we made, they knew it. There was an incapacity to combat                                                                                                                 182 Quinn, T. 1991. “A Day with the Chessplayer” Time, July 1. Rodriguez went on saying “[Ruiz Barrera] was commander of the Fourth Brigade from 1986 to 1988, if I'm not wrong. He chased Mr. [Pablo] Escobar and his partners persistently, and yet he failed in all his attempts. He didn't succeed in gaining immortality with the Medellín cartel. Thus the Cali cartel was invented, and with it the war over the New York market. Of course this tale about the Cali cartel has been helped along by my differences with Mr. Escobar.”   218   the powerful groups.”183 Of course, just as criminals wanted to deny existence of the organization, authorities were constantly trying to create images of a powerful organization. But there is enough documentation to prove that the Cali DTO indeed had power and a strong support network, which allowed it to keep illegal market disputes under control. Towards the end of the 1980s Pablo Escobar and the Medellín DTO openly confronted the Cali DTO thus increasing market competition, and this confrontation led to increases in the frequency of violence in 1991 (figure 1). Interestingly Escobar did not challenge the control that the Rodriguez Orejuela had in Cali184 in part because he was unable to physically dominate the city (Vicepresidencia de la Republica, 2009) and thus the dispute did not see homicide rates spike. Violence never reached extreme levels in this period because the Cali DTO insourced violence, using armed apparatuses clearly controlled within the organization, and without actively recruiting youth gangs. The decision to insource derived from the limited demand of gang armed force by the Cali DTO, linked to its success in securing other sources of intelligence, protection, and security, in networks that engaged taxi drivers and a great number of former or active military and police personnel. According to an Army colonel who persecuted the organization, and to legal documents of the Proceso 8000, a large-scale judicial process initiated in 1995 that documented the links between the Cali DTO and state officials, about 60% of the personnel at the Metropolitan                                                                                                                 183 Author’s interview with Army Brigadier, Bogotá, September 1 2011. 184 Even though Escobar sent sicarios to Cali, there was never a massive move of force that could really challenge the power of the Cali DTO. A nice contrast to this is the situation of Ciudad Juárez described in Chapter 6 that shows how the power of the Juárez DTO was really challenged by the physical move of force carried out by the Sinaloa DTO.   219   Police of Cali appeared in a record of payments given by the Cali DTO. 185 One document of the Proceso 8000 listed 200 officials, sub officials, and police members that were on the payroll of the DTO in Cali. This penetration of the police guaranteed criminals not only non-enforcement of the law, but also direct provision of security. The extensive recruitment of police and army personnel in the Cali DTO made outsourcing to youth gangs unnecessary (Chaparro 2005, p. 127) and thus made the control and compliance on the use of violence easier for the Cali DTO. In Cali, limited outsourcing was also related to the relatively short supply of youth gangs compared to Medellín, where outsourcing to gangs was widespread (Guzman 1993, p. 24, 42; Universidad del Valle 1998). Existing gangs at the time were weakly organized and armed, usually involving kids that gathered for drug consumption rather than for the commission of crimes and that could sell services to criminals but usually sporadically (Vanegas 1998, Vicepresidencia de la República 2006). Interestingly, the low supply of youth gangs did not simply reflect the existence of a smaller young population in Cali:186 Cali and Medellín had very similar population structures by age brackets, yet the youth gang phenomenon and its relation to drug trafficking was remarkably different in both cities.187 Youth gangs were less organized in Cali precisely because traffickers did not deploy them systematically. Thus the insourcing of violence                                                                                                                 185 In a testimony of Guillermo Pallomari, accountant for the Rodriguez Orejuela brothers, he described to a judicial court that there was a person in the organization’s “security area” whose role was to contact cops and military to work for Miguel Rodriguez. (Fiscalía General de la Nación, Diligencia de indagatoria que rinde el señor Guillermo Alejandro Pallomari Gonzalez). 186 Reuter (2009) has suggested that the decrease of violence in drug markets in the US might be related to a generational change and a reduction of young populations, but this does not seem to be the case in Cali. 187 Between 1973 and 2001 the population between 10-29 years represented 41% of the total population in Cali and 40% in Medellín.   220   of the Cali DTO was mostly the product of an organizational preference, facilitated by the extensive and successful recruitment of alternative sources of armed power. 5.1.1.1. Proceso 8000: how state capture motivated less visible violence The Cali DTO preferred methods that concealed, rather than exposed, the evidence of violence. During the 1980s the press reported the occurrence of mysterious crimes that involved, for example, the finding of bodies in the Cauca River,188 or the purposed elimination of fingerprints in corpses, so they could not be identified. In the words of a military colonel who was part of the Search Bloc against the Cali DTO “In Cali there were silent deaths, without noise, without paraphernalia.”189 An interview conducted by Betancourt (1998) to a local woman in Cartago, a municipality close to Cali, summed up the low visibility, and yet the brutality, of the violence used by criminals: dead people started to show up, young people, some of them very humble, farm laborers, disfigured, burned, tortured; people said they were employees of the narcos that had snitched or robbed them; others were poor people, beggars, fagots, sluts, because it was necessary to accommodate the city to the requirements of progress and that people were inconvenient. […] A kind of dead people that was not talked about were women, young, girls, between sixteen and eighteen, pretty and well dressed. The curious thing is that they hanged out with Mafiosi; people say they were lovers that had cheated and so they were killed [muñequiar]; others say that it was a punishment, because they did not want to go out with them; the truth is that nobody wants to think about it or remember it. There is a law of silence here, of suspicion; every person who arrives is watched. (Betancourt 1998, p.165)                                                                                                                 188 The Cauca River borders the Eastern part of Cali. 189 Author’s interview with Army Colonel, Bogota, January 31 2011.   221   Along with the use of methods that concealed evidence, high profile victims, attacks against public and enforcement officials, or collective attacks with more than 3 victims were not common in Cali in this period. Figure 5.2 shows that most victims of violence were merchants and non public figures (street vendors, non-professional workers such as security guards or mechanics, and presumed delinquents) but there were no attacks on high profile political or public figures. Yet, while they avoided prominent attacks, the members of the Cali DTO promoted social cleansing, that is, the elimination of sectors of the population considered as undesirable such as petty criminals, drug consumers, and prostitutes.190 In the 1980s, social cleansing groups such as “Cali Limpia”, “Amor a Cali”, and “Palmira Eficiente” proliferated in Cali. Although the emergence of such groups was not limited to the action of drug traffickers191 it is clear that they played a role in this form of violence, which probably represented a mechanism to legitimize their presence both among the population and among authorities, but that given the profile of the victims, was of less concern for authorities (Stannow 1996).                                                                                                                 190 The UN defines social cleansing as the elimination of impoverished and marginalized sectors of the population (Stannow 1996). 191 Guzman and Camacho (1990), Guzman (1993), and Vanegas (1998) describe the complex dynamics and multiple actors involved in social cleansing in Cali, including state authorities, business actors, militias, neighborhood groups. They also describe how social cleansing was related to political violence and especially in 1985, with an effort of eliminating the M-19, a major guerrilla group that started to expand its presence in the city since 1984.   222   Figure 5.2. Occupation of victims of violence in Cali 1984-1989 Author’s dataset on drug related violence, based on El Pais newspaper, for the years 1984 and 1989. In about 50% of the news reports the occupation of the victim could not be identified The Cali DTO’s success in penetrating the state created the incentives to maintain this low profile violence. Such penetration was not limited to local politicians, included national level Senators, and was so extensive that it reached the coffers of the presidential campaign of Ernesto Samper who became president of Colombia in 1994. The Proceso 8000 documented such links and revealed how the Cali DTO built a powerful protection network throughout the Colombian state, and how the basis of such a network was carefully crafted in Cali. Local politicians indicted during the Proceso played a variety of crucial roles for the Cali DTO; the Mayor of Cali, Mauricio Guzman (1992-1995) created façade businesses to launder money for the criminal organization, and senator Gustavo Espinosa supported favorable legislation for traffickers. Other prominent local politicians indicted included a former Secretary of Government in Cali, a former Governor of Valle, the General Comptroller of the nation, Manuel Francisco Becerra, and local senators like Alvaro Mejia Lopez and Armando Pomarico.   223   The Rodriguez Orejuela preferred bribing state officials rather than killing them, but their success in obtaining state sponsored protection was not only a personal choice, it was allowed by a cohesive security apparatus in Cali, which provided secure protection from the state, lighter sentences, and cover-ups for money laundering; in exchange, state actors obtained less violent behavior, campaign funding, direct financial support (Betancourt 1998, p. 136) and even collaboration in dismantling, the Medellín DTO (Bowden 2002). As explained in chapter 1, a cohesive security apparatus depends on the absence of intergovernmental conflicts, the absence of interagency conflicts, and the relative stability of time horizons for elected and enforcement officials. These characteristics were present during the existence of the Cali DTO. Cali displayed powerful, ingrained, and stable traditional political elites whose power was very difficult to challenge.192 Even though at different times political competition threatened the cohesiveness of the security apparatus (for example with the first popular election of a mayor in 1988) a powerful political and economic elite dominated politics in the city (Camacho and Guzman 1990, p. 185) blocking opposition and reformist efforts. At the same time, interagency conflicts were limited because authorities other than the police were not involved in antinarcotics operations in the city. The military did carry out massive urban antiguerrilla operations in 1985 in Siloé, one of the poorest and marginalized neighborhoods of Cali, causing human rights violations (Camacho and Guzman 1990, 139). These operations however, did not cause conflicts                                                                                                                 192 According to Rodrigo Guerrero, the first mayor elected outside traditional parties in the city in 1990, Cali’s main problem is the persistence of a strong political class that controls the city and blocks every reform effort. Author’s interview, Cali, November 11 2010.   224   between the Army and the Police as they were implemented as joint operations that were short-lived and well coordinated. In this context, accessing the state became easier for the traffickers of the Cali DTO. The following testimony of Miguel Rodriguez Orejuela before Colombian authorities encapsulates both his ideas about the use of violence and how these were linked to the collaboration established with the local security apparatus Mr Pablo Escobar […] sent us a letter asking for our collaboration, physical and economic if you can say that, to start a war without barracks against the State and its institutions and representatives, including the Justice sector. Our response, because of our beliefs and because of our respect not only towards institutions but also towards citizens, and because of our families’ peace, was absolutely negative, and Mr Escobar’s reaction did not wait, sending people to Cali and Bogota to spy on us. […] Mr Escobar became our enemy […] That led us to focus on protecting our families, the city of Cali and ourselves, as many people can testify, especially the institution of the police, to which we always provided information and informants, with the results that happily, not only for us but for all the country, that institution can have. (Las Confesiones Secretas de Miguel y Gilberto Rodriguez Orejuela, 2005, p. 58) Eventually, the Proceso 8000 and the pressure it generated from the United States193 forced the government to act against the Cali traffickers and to involve the military in urban antinarcotics operations in Cali. The national government created a joint military-police elite force known as the Cali Search Bloc, which was mimicked after the elite force created to persecute and capture members of the Medellín DTO, and responsible for killing Pablo Escobar. Yet, according to members of the Cali Search Bloc and other high level politicians I interviewed, in Cali the Search Bloc relied more on the Army and focused on intelligence gathering rather than on direct military actions because                                                                                                                 193 In 1994 the US government approved the use of a certification process to condition aid to Colombia on the progress on the War on Drugs. In 1995 Colombia was for the first time decertified, and in the same year the visa of Colombian President Ernesto Samper was cancelled.   225   the core of the Cali DTO’s power was precisely its penetration of the police and the political establishment: “the crucial difference between the Cali and Medellín Search Blocs” a former Minister told me “was that we understood that the police had stronger links with the narco in Cali, thus the Bloc focused on the army, and the objective was to cut their communication networks in the city. We knew there were difficulties given the social acceptance these guys had.”194 The Cali DTO’s successful penetration of the state provided them with important benefits and protection, yet the emergence of the Proceso 8000 scandal illustrates how state protection rackets can be prone to instability, in this case due to pressures from a competitive electoral system at the national level. In fact, the first person accusing President Ernesto Samper of receiving drug money and then revealing evidence that led to the initiation of the Proceso 8000 was Andres Pastrana, the main competitor who lost the Presidency to Samper (Vargas et al 1996). The most notable leaders of the Cali DTO were captured in 1995 and even though they managed to operate business from prison until the extradition of the Rodriguez Orejuela brothers to the United States in 2004 and 2005, the structure of the organization faded away. A period of re-accommodation in Cali’s state security apparatus and in the illegal drug market followed. The perception among the population about the traffickers that followed the Rodriguez Orejuela was that they were less respectable than their predecessors, as echoed in several interviews and conversations I had. A taxi driver told me “Today’s traffickers are not like their predecessors. Yesterday’s were businessmen, respectable people. Today’s are simply matones, mafiosos...” while a woman who started                                                                                                                 194 Author’s interview with former Defense Minister. Bogotá, January 13 2011.   226   conversation in a public bus with me said “Mafiosi were not bothered by the government before and there was employment for everybody, today there is only unemployment.” These statements help illustrate that the network and power of the Rodriguez Orejuela was critically tied to their connections and acceptance within civil society; these connections with civil society, as Arias (2006) has illustrated in Rio de Janeiro, are crucial for the power of criminal networks. Yet, the fact that social acceptance faded away for the successors of the Cali DTO and yet their violent behavior remained similar, lends more support to my argument that the interaction between state and illegal organizations determines the latter’s violent behavior. The massive corruption of police forces in Cali during the existence of the Cali DTO motivated major reforms in the police in Colombia195 including rotations and purges, such as the dismissal of all station commanders in Cali in 1995.196 But despite these transformations the security situation in Cali would not change radically in the years to come.                                                                                                                 195 As seen in Chapter 3, according to Casas Dupuy (2005), the operational changes that took place in the police between 1995 and 1998 were actually a counter reform of a reform process initiated in 1993 aimed at decentralizing the structure of police forces, strengthening their civilian nature and the community’s oversight, and professionalizing it. By contrast, the process occurring between 1995 and 1999 recentralized Police forces (for example by giving the national Director of the Police discretionary authority on firings and dismissals) and gave more autonomy to the institution. 196 El Tiempo. 1995. “Bajo la presión de E. U. Asume comandante de policía de Cali”. El Tiempo March 9.   227   5.1.2. Fragmentation of the criminal market: everything changes but everything remains the same (1995-2010) With the dismantling of the Cali DTO, criminal groups based in rural areas of the Valle department gained preeminence, the battles for turf between factions of the Norte del Valle DTO got more intense, and the market more competitive. According to a local intelligence agent in Cali “Since 1996-1998 the dynamic has always been the same, a fight for territorial control.197” Cali remained as an important center for criminal activity because it represented a hub for money laundering and a crucial connection from urban centers to smuggling routes in the Pacific coast of Colombia. Drug violence in Cali became more dependent on dynamics occurring in the northern rural areas of Valle, but still had a distinct character. Violence remained with low visibility because the security apparatus in Cali remained cohesive despite the massive firings of police, criminals continued penetrating the state, and thus, with some exceptions, they preferred to limit the use of violence that could force state attention. In the words of a military member of the Cali Search Bloc: “The Cartel del Norte del Valle assimilated the experience of Cali and devoted more emphasis to bribing at the municipal level.”198 5.1.2.1 Disputes in the post Cali DTO era The disputes of the late 1990s involved high-level members of the Cali DTO that had not been captured and leaders of the Norte del Valle DTO such as Orlando Henao and Ivan Urdinola. As these leaders were captured or assassinated, people who worked                                                                                                                 197 Interview with intelligence agent of the attorney’s office, Cali, November 12 2010. 198 Interview with Military Colonel of the Search Bloc, Bogotá, January 31 2011.   228   for them gained prominence and became protagonists of subsequent disputes, not only for market control, but also for retaliation against criminals who collaborated with US authorities by handing them information about their own associates, a practice that became very common among Norte del Valle traffickers.199 Between 2002 and 2007 competition in the illegal market heightened as two prominent traffickers struggled to eliminate each other: Diego León Montoya, who became the leader of the Norte del Valle DTO, and Wilber Varela, who started as a bodyguard, and ascended up in the organization. This dispute increased the frequency of violence in Cali as reflected in an average homicide rate of 84 for those years. Both leaders created their own armies, known as the Los Rastrojos that worked for Varela, and los Yiyos and Machos, that worked for Montoya. The war between Varela and Montoya, and the frequency of violence decreased after a short truce between them in 2005.200 The truce illustrated the possibility of reducing violence in a competitive illegal market through criminal arrangements, but the aftermath also reflected the instability of criminal pacts. After the Army captured Montoya in 2007 and Varela was assassinated in 2008, their armies became more powerful: Rastrojos and Machos expanded operations to several areas of the Pacific Coast of Colombia and entered more aggressively in urban areas. Since 2008 the confrontations between these groups kept Cali’s homicide rates high. The dynamic of the                                                                                                                 199 This prominent dynamic of dilation is captured in the book “El Cartel de los Sapos” authored by a former member of the Cartel del Norte del Valle. The book has become the basis of numerous soap operas in Colombia. 200 The truce was apparently mediated by a paramilitary leader who considered that the war was starting to be noisy and detrimental for illegal business and armed action in the region. See: El Pais. 2007. “El Ocaso de los narcos” El País, January 15, and Revista Semana. 2005. “La Jugada de Macaco” Revista Semana, November 22.   229   criminal market thus remained highly competitive; the names and protagonists of turf disputes varied when a leader was captured or killed, but the constant dynamic of confrontation and accommodation persisted, thus maintaining violence on a highly frequent equilibrium. 5.1.2.2. Fuzzy trafficking-gang connections and limited outsourcing The protagonists of turf disputes in the 1990s and 2000s occasionally recurred to sicarios [hitmen], and oficinas [offices], groups that acted like brokers between trafficking organizations and sicarios (Duncan 2005, p. 45). Yet, these traffickers did not engage in a sustained practice of outsourcing violence to youth gangs, thus maintaining violence in Cali at high, but not extreme levels. According to a police commander I interviewed, sicarios were usually deployed to carry out single operations, yet traffickers relied on core groups of insourced force. During the 2000s outsourcing was more prominent than during the Cali DTO period, but it was still limited. Insourcing was not a reflection of a short supply of youth gangs, as the presence of these organizations was significant: By 2010 the Metropolitan Police of Cali and the Mayor’s Office estimated that 103 gangs existed in the city (Observatorio Social Alcaldía de Cali 2010).201 It was rather the result of a lack of demand, and of the existence of other sources of armed force. Traffickers like Montoya and Varela established complex relationships with guerrillas and paramilitaries in rural areas, which supplied force and reduced the need to outsource to youth gangs in Cali.                                                                                                                 201 This calculation included both organizations devoted to large-scale crime and connected to drug trafficking and oficinas, as well as “corner groups” usually dedicated just to drug use and occasional petty robbery.   230   Varela initially recurred to guerrillas202 and later on, as paramilitary groups consolidated in the north of Valle, different criminal factions established closer relationships with them.203 An indication of the fuzzy relations between traffickers and gangs204 was the difficulty I noticed during my fieldwork and interviews in Cali to identify connections between gangs and traffickers. This situation stood in contrast to the relative clarity that observers, gang members, and authorities had to identify connections between gangs and larger criminal organizations in Medellín. Interviews conducted with youth in conflict in Cali by Alberto Concha-Eastman205 further captured the episodic nature of the relations between gangs and traffickers in Cali. According to one young man with legal antecedents “no, one ever knows the patron, somebody looks for you “look, I have a vueltica [a job, killing or stealing]” […] but one does not know the patron” and another young man reinforced the idea stating that “there is a law of silence here, in small neighborhoods they control everything.” A report by the Observatorio Social de la Alcaldía de Cali echoed the same observation about gangs “They are also linked to selling and distributing hallucinogens and fire arms; it is possible that they have relations                                                                                                                 202 Interview with Police Commander, Bogota, August 30 2011 203 Multiple reports refer the effort that both Varela and Montoya did to create alliances with paramilitaries that would allow them to demobilize as paramilitary leaders within the peace process that the government started in 2003. They did not succeed in posing as paramilitaries partially because their trafficking past was too evident. Their failure contrasted with long time criminal Diego Fernando Murillo’s success in posing as paramilitary in Medellín, which allowed him to demobilize, and monopolize all sectors of criminality (Chapter 4). 204 The fuzziness of links between gangs and criminals created a situation in which common criminality acquired a higher level of visibility while criminal disputes became less identifiable in the city, leading authorities to claim at different points that most violence in Cali was related to common crime or “social intolerance” See for example El Tiempo. 2010. “Ante violencia urbana, Cali promueve el desarme” El Tiempo. 31 August. This is not to negate the fact that social intolerance can cause homicides but more to emphasize that the relation between common disputes and criminal disputes is blurry. 205 In the context of a project on youth and violence carried out by El Colegio de Mexico and directed by professor Arturo Alvarado.   231   to “oficinas sicariales” and illegal groups, but as a whole they do not devote themselves to committing homicides although their members can oscillate on jobs like sicariato” (Observatorio Social, Santiago de Cali 2010). 5.1.2.3.Threats to democratic institutions in the aftermath of the “big cartels” The criminal disputes that followed the demise of the Cali DTO were deadly, long and destabilizing for Cali, yet they did not cause the same public outrage caused by trafficking disputes in the 1980s. Even the war between Montoya and Varela, which was particularly bloody, was, according to a local agent of the Attorney’s Office “not noticed, the society was indifferent because it did not touch us, the clubs [bars] were packed.”206 This perception of security in Cali reflected the persistent low visibility of violence. A press report declared that the war between Montoya and Varela left more than 3000 victims, yet most of them were reported as disappeared people.207 Throughout these years the local security apparatus remained as a cohesive structure prone to protect criminals. In 2008, a regional development report produced by the United Nations Development Program identified the penetration of drug trafficking in the public and political life of the region and the criminalization of the political class as key challenges the department of Valle required to address to promote human development (UNDP 2008).                                                                                                                 206 Interview with an intelligence agent of the General Attorney’s Office (Fiscalía General), Cali, November 12 2010. 207 Revista Semana. 2007. “Don Diego permanecerá en el bunker de la Fiscalía antes de ir a Combita” Revista Semana, September 10.   232   Powerful local political elites maintained a cohesive state apparatus and successfully participated in national politics, preempting conflicts among levels of government and facilitating the protection of criminals. Some analysts argue that the problem of security in Cali resulted from the difficulty of creating governing coalitions for security reform, as political competition brought in new players and independent leaders won different elections, but no reformist coalition succeeded (Gutierrez et al 2009). Although this difficulty can be interpreted as a sign of fragmentation, in practice it reflected the power that patrimonial political forces with strong illegal links maintained, as these forces ended up penetrating even the campaigns of politicians seen as detached from corruption. Thus, despite conflicts and opposition, there were no forces that could veto the key dominant political players in Cali’s security apparatus208 and thus countermine the strong influence that illegal actors had in the politics of the city. Along with the persistence of powerful elites, the relations between enforcement agencies were relatively stable in the aftermath of the Cali DTO. Yet, in line with the argument posed in this dissertation, when conflicts among enforcement agencies or rotations and purges occurred in Cali, instances of visible violence appeared. For example, compared to other criminal disputes, the conflict between Varela and Montoya produced more instances of visible violence.209 According to some of the intelligence analysts I interviewed in Cali, this reflected the fractured alignment of police forces with Varela and military forces with Montoya and the lack of collaboration between police                                                                                                                 208 A UNDP (2008, 291) report explains that local governments in Cali have been poorly evaluated by the public opinion, and in one case a mayor was impeached. Yet scandals usually emerge out of civil society or media efforts, but not out of publicity efforts by political opposition forces. 209 On October 13 and December 12 2003 there were two massacres one in a dance club and the other in a mall in Cali, both related to the confrontation between Varela and Montoya.   233   and military: “there is a certain struggle between the army and the police […] The war is more difficult because the state is divided.210” One of the worst waves of violence in Cali during the Montoya-Varela confrontation occurred in May 2006 after a military patrol killed 10 members of an elite police force in Jamundí, a municipality located to the south of Cali. The judicial investigation revealed that Montoya hired an Army colonel to eliminate the police group that was in search of a cocaine stash211 and that Army collaboration in Montoya’s organization was widespread.212 The incident not only complicated police-military relations; it also illustrated the possibility that competition in the illegal market may fragment a state protection racket by permeating one agency more than other, or by dividing protection from different agencies among different criminal groups; this fragmentation can produce more violence. The tension between Army and Police after the Jamundí events did not completely fragment the security apparatus in Cali because the tension emerged in rural areas, and the army was less present in the struggle for security in Cali than in the rural areas, where it focused on fighting guerrillas and destroying illegal crops.213 At any rate, the Jamundí episode also illustrated the extent to which criminal and state actors in the aftermath of the “big cartels” established strong networks of collaboration that debilitated state institutions and threatened democracy as much as in previous periods. Although not all allegations of corruption were fully investigated in                                                                                                                 210 Author’s interview in Cali, November 12, 2010. 211 Revista Semana. 2007. “Un año de massacre de Jamundí en la batallón del ejército aniquila la major unidad antidrogas de la policía”. Revista Semana, May 21. 212 Revista Semana. 2007. “Corrupción hasta el tuétano” Revista Semana. August 11. 213 Author’s interview with intelligence official, Cali, November 12, 2010.   234   Cali’s recent past, key members of Cali’s political class were associated with criminal and armed actors. And even though allegations of links between politicians and drug traffickers emerged in other regions of Colombia, the crucial aspect in Cali was that powerful politicians with large political, electoral, and power networks were part of these scandals and sometimes remained powerful despite these scandals. In Cali, borrowing Peter Lupsha’s definition of the relations between criminals and state in Mexico (1991), “the players change but the game continues.” Even if the protagonists changed continuously, the process that connected criminals, state, and violence remained very similar. A case that illustrated the persistent capture of the state in Cali was the imprisonment in 2011 of Senator Juan Carlos Martinez Sinisterra for his alleged links with drug traffickers of the Norte del Valle DTO and paramilitary groups in Valle. Martinez started his career214 under the protection of prominent politician Carlos Herney Abadia (investigated within Proceso 8000) and maintained close links with Abadia’s son and then governor of Valle (2008-2010) Juan Carlos Abadia.215 He experienced a meteoric rise in politics thanks to his links with criminals who provided him key assets for sustained electoral success: votes (obtained when necessary through the coercion of paramilitaries on civilians) and money. In exchange Martinez provided favorable public                                                                                                                 214 Martinez was born in a small municipality of the Pacific department [state] of Cauca, and his power extended to the departments of Cauca, Choco, and Valle. 215 Revista Cambio. 2008. “Senador Juan Carlos Martínez se enreda por presuntos vínculos con narcotraficantes”. Revista Cambio, November 2. Revista Cambio. 2011. “El hombre que maneja medio país desde la cárcel” Revista Semana, September 17.   235   decisions (for example assigning public contracts to illegal allies) and tolerated drug trafficking in the Pacific Coast of Colombia. Martinez’s case reveals important aspects of the connections between criminals and state. First, it illustrates how powerful corrupt political elites perpetuate themselves in power, through the support of illegal actors. Martinez kept his influence even after he was imprisoned and prominent members of Cali’s political class courted him in prison. He also managed to keep influencing elections: in 2010 he formed the political party PIN that won 20 seats in Congress becoming the fourth political force in the country, despite the shadow of doubt that existed for its links with paramilitaries. For the 2011 elections a new party called MIO (Movement for Inclusion and Opportunities), which included former members of the PIN and politicians linked to Martinez, consolidated as a political force in Cali’s Council and Valle’s Assembly, and reflected that Martinez’s power remained alive and strong (Corporación Nuevo Arco Iris 2011) even in spite of some electoral defeats (Avila and Velasco 2012). Second, Martinez’s story shows that symbiotic relations between criminals and state cannot simply be reduced to a story of subordination of politicians to illegal forces. Politicians depend on illegal actors for their electoral success and power, but at the same time the power of these politicians provides illegal actors privileged access to the state and incentives to use less visible violence. It is important to note that in the late 1990s and 2000s the incentives for using low visibility violence also derived from a learning process in the criminal world that taught criminals that too much noise forced the state to confront them. This was particularly important in the context of reform that throughout the 1990s and 2000s increased the technical, operational, and tactical capabilities of the Colombian National Police (De   236   Francisco 2005; Llorente 2005)216 and made it more effective in targeting high-level criminals. Thus, using less visible violence became also an adaptation of criminals to operate better without attracting much attention from national level authorities that could capture them more rapidly once they became a target of state action. According to authorities, the “life” of criminals in Colombia has been shortened217 because once identified, the time elapsed before their detention is shorter than it used to be in the 1980s. While such statement ignores the fact that criminals operate for a long time before they become high priority targets, it illustrates that if a criminal becomes too public, it is under higher treat of being imprisoned. Thus low visibility in recent years in Cali has been the result of criminals’ interlocking incentives to keep state protection at the local level, but also to avoid risking such protection by forcing more effective national state forces to go after them. The next section illustrates how criminal competition and insourcing of violence, combined with high levels of state protection in a cohesive state, also explained the persistence of low visibility violence throughout the history of Culiacán. 5.2. Culiacán: the invisible “Chicago” of Mexico The origins of drug trafficking in Mexico, as Luis Astorga (2005) has carefully documented, date back to the late eighteenth century and can be traced back to the state of Sinaloa. Modern drug trafficking organizations took shape in this region between the                                                                                                                 216 Two critiques to this process of institutional change in the Police are notable: first the increased militarization of police functions in Colombia; second, that despite a significantly improved image of the Colombian police both domestically and internationally, and despite its increased effectiveness, corruption cases are still prevalent, specially at the local level. 217 Revista Semana. 2009. “La vida útil de los narcos.” Revista Semana, April 21.   237   1920s and the 1940s as the focus of law enforcement on narcotics shifted from an emphasis on public health to an emphasis on public security (Astorga 2004, Florez 2005, Serrano 2007). As this process took place, Culiacán became the bridge between entirely rural areas where most marihuana and poppy crops were located, an urbanized and commerce oriented economy,218 and the coastal area where drugs could be shipped. By the 1960s and 1970s drug trafficking had consolidated in the region, and the first law enforcement responses emerged. During this period, confrontations among traffickers started to give the city the reputation of “the Chicago of Mexico”, as observers compared it to mafia disputes in Chicago (Craig 1980). In 1977, ten thousand soldiers of the Mexican military with US support and under the command of the Attorney General’s Office in Mexico (PGR) carried out a large-scale illegal drug crop eradication campaign known as “Operación Cóndor”. The operation led to massive displacement of poppy cultivators known as “gomeros” and of civilians in the rural areas of Sinaloa. It also entailed significant human rights violations (Astorga 2005, 118), and caused temporary but major shifts in trafficking organizations (Smith 1997). To avoid further action of authorities, many traffickers relocated from Culiacán to the city of Guadalajara further south in the pacific. Towards the beginning of the 1980s, however, many trafficking operations and traffickers had returned to Culiacán.                                                                                                                 218 Compared to other major cities and to the national average, in Culiacán the weight of agricultural activities on the GDP is very high (about 12% compared to 3.7% nationwide), and is one of the leading cities in the production of corn, tomatoes, peppers, and also in livestock. Yet commerce and service also represent a big proportion of the GDP (21% compared to 15% nationwide), and Culiacán headquarters companies of national importance like Coppel, Casa Ley and Comex. [INEGI 2010]   238   5.2.1. Political protection and criminal power struggles (1984-2008) By 1986, Culiacán had a very high homicide rate of 66 per 100.000 inhabitants and the rate remained constantly above national averages throughout the decade, as shown in Figure 5.3, reflecting the disputes that emerged as traffickers relocated in Culiacán.219 This high frequency of violence contrasted with its low visibility, which was reflected both in the methods employed by criminals, but also in the perception of the population that violence only affected criminals: In the words of a local journalist “the disputes were between them, if you did not have problems with them, nothing happened.”220 According to many civilians, experts and authorities I interviewed in Culiacán,221 the high levels of violence that historically characterized the city and that are tragically reflected in the multiple cenotaphs, or crosses that mark places were people have been shot,222 were not perceived as threatening because they were thought to only affect criminals, and people felt that if they did not get involved with traffickers, they were not likely to be victims of violence.                                                                                                                 219 It is worth noting that the average homicide rate for Cali in the entire period 1984-2010 was 79.6, considerably higher than the average of Culiacán of 34.1, and this difference reflects the impact of long standing armed conflict in Colombia. Yet, these two cities are comparable because in their national context both were considerably more violent than the rest of the country. 220 Author’s interview with local journalist, Culiacán, April 2011. The same statement appeared in most of my interviews in Culiacán 221 Interviews conducted between March 18th and March 26th 2011 in Culiacán. 222 There are so many of these cenotaphs in the city that at some point in 2008 the Mayor’s office discussed the possibility of removing all of them.   239   Figure 5.3. Homicide rates Culiacán 1985-2010 Source: Author’s elaboration with data from INEGI and SINAIS The persistence of high frequency violence in Culiacán throughout the 1990s and 2000s reflected the conflicts that surrounded a criminal market that was under constant threat of dispersion, due to the confrontation between different criminal factions of Sinaloa. The common origin of drug traffickers facilitated certain continuity in the criminal market, but at the same time generated strong disputes among leaders that competed against each other (Astorga 2005, p. 166). In the 1980s, Miguel Angel Felix Gallardo, a prominent trafficker that started his career as bodyguard for a Sinaloan governor, managed to keep some of these oligopolistic tendencies in the market controlled. He engaged in disputes with other traffickers like Hector Palma and Joaquin Guzman (Astorga 2005, p. 145) but consolidated power through his pervasive political and economic networks. Even though Felix Gallardo resided in Guadalajara, he constantly traveled to Culiacán, where he maintained businesses and close relationships   240   with politicians like Governor Antonio Toledo. He also had a reputation as a “discrete” narco.223 On April 8 1989, the Federal Judicial Police (PJF) captured Felix Gallardo; his capture was influenced by his involvement in the assassination of DEA agent Enrique Camarena in 1985. While in jail, Felix Gallardo divided the territorial control of the major trafficking regions in Mexico including the states of Baja California, Chihuahua, Tamaulipas, and Sinaloa (Astorga 2005, p. 167; Blancornelas 2002, p. 54) among the different groups that tended to be opposed in Sinaloa. Such decision had a double effect: on the one hand it decentralized violent disputes moving some violence away from Culiacán as occurred in Cali after the demise of the Cali DTO; but at the same time it contributed to make the market more competitive. Disputes among factions, especially between the serranos [rural, from the mountains] like Ismael Zambada and Joaquin Guzman, and the urban Arellano Felix brothers, leaders of trafficking in Tijuana, increased throughout the decade, and so did homicide rates in Culiacán. These disputes always had a manifestation in Culiacán because even those traffickers who had moved away from the city, like the Arellano Felix brothers, strived to retain properties, businesses, power, and access to drug producing areas in the Sierra [mountains] through Culiacán. These disputes however did not create extreme spikes in homicides because drug traffickers in Culiacán preferred to insource violence. On the one hand, most Sinaloan                                                                                                                 223 Ortiz Pinchetti, F. 1989. “El hombre más buscado del mundo.” Proceso No 650, April 17 1989, reprinted Edición Especial No, 32 February 2011.   241   traffickers were from rural origins224 and gained the allegiance of older peasants who acted as their armed muscle (Astorga 2005, p. 178): using older peasants as sicarios guaranteed discipline and allegiance to the traffickers. On the other hand, youth gangs never took off in Culiacán. In the 1980s youth gangs and cholos225 became common but they never reached the strength they had in border cities like Ciudad Juárez; furthermore most cholos were not armed or violent.226 An interesting indication of insourcing to old and loyal workers of traffickers, can be seen in the proportion of homicides by age group presented in Figure 5.4: compared to Ciudad Juárez and Tijuana, Culiacán had smaller proportion of homicides among population 24 or younger in the period 1985-1996; by contrast, the proportion among an older bracket of the population was significantly higher in Culiacán than in other cities. As we will see in the next section, this is a pattern that persisted even after a “war” exploded in 2008.                                                                                                                 224 With the exception of the Arellano Felix brothers, who were born in Culiacán’s middle class, and were more educated and urban. 225 Cholos is a term used to characterize a youth cultural expression that emerged in the 1960s mixing American and Mexican identities especially in border areas, and associated to particular dress codes (baggie pants, vandanas). Cholos are linked in the public imaginary to criminals, although in most cases they are not linked to any form of criminality (Valenzuela 2009). 226 Author’s interview with academic expert in Culiacán March 19 2011; author’s interview with local journalist March 18 and 19 2011.   242   Figure 5.4. Proportion of homicides by age group 1985-1996 Source: Author’s elaboration, SINAIS data 5.2.1.1. The power of Sinaloan governors as a motor of state protection The persistent violence of Culiacán was also characterized by its low visibility. The dataset of violent events that I constructed for Culiacán (which includes the years 1984, 1992, 1996, 2002, 2008 and 2009) provides important elements to characterize this low visibility of violence. First, events of collective violence were not common: before 2009 there were only two events with more than two victims, and a high percentage of violent events (62%) did not produce lethal victims but only left people wounded. Second, most victims of homicide were unlikely to attract significant public attention: peasants, non-professional workers, and delinquents. Finally, most violence was geographically located in rural areas of Culiacán, especifically in the sector known as Tierra Blanca, a poor neighborhood that connects the city with the mountains [La Sierra]. A higher percentage of homicides and bodies were found in the outskirts and rural areas of Culiacán (37%) than in the other cities studied in this dissertation (the average from all   243   other cities was 16%). This spatial distribution reflected that a significant amount of violence was hidden from public attention. In fact, there were places in the city known as “botaderos” (dumpsters) for corpses, for example the “Cerro del Tute” in the 1970s and 1980s, and in the 2000s “La Costerita” road.227 It is understandable that traffickers preferred these locations to hide violence because few people mingled around these locations, usually connected to vast areas of land owned by traffickers. In fact, while I was traveling around one of these locations I had my first experience being followed by a car, as a reminder that these areas are not normally transited by inhabitants of the city. If the confrontations within a strong, but competitive group of criminals explained the persistence of violence in Culiacán, their ability to secure stable protection by the state reduced their incentives to use highly visible forms of violence. The concentration of political power facilitated the consolidation of close relations between traffickers and authorities. The historic power of governors in Sinaloa, which shadowed local mayors and contributed to concentrate power in Culiacán, helped reduce intergovernmental conflicts, and along with the stability of the PRI party in elections in the region, contributed to a cohesive security apparatus. Since the 1980s powerful governors like Antonio Toledo (1981-1986) were accused of protecting traffickers (Lazcano y Ochoa 1992); mayors in Sinaloa were seen as “submissive to the governor”228 and considered invisible229 while governors were seen as “almighty”. Sinaloan governors were also important in national politics, as many of them were part of presidential cabinets                                                                                                                 227 Author’s interviews with an academic and local expert from Culiacán conducted in March 19 and July 1 2011. 228 Author’s interview with State deputy for Sinaloa, Culiacán, March 23 2011. 229 Author’s interview with official from the Governor’s office in Sinaloa, Culiacán, June 28 2011.   244   (Alvarado 1997), and this also reduced sources of conflict with the federal government. The power of governors was reinforced by the stability of the PRI domination both in gubernatorial and mayor’s elections. Sinaloa was a PRI stronghold (Alvarado 1997), and even though in the early nineties opposition parties PAN and PRD gained more votes, those were not enough to challenge the PRI domination. Gubernatorial power and PRI dominance explain how, despite the undeniable reality of violence in the city and the increasing accusations of corruption of its political class, none of the two issues became top priorities for authorities in Culiacán and Sinaloa. By contrast, as documented by Astorga (2005) at different points in the 1980s and 1990s, state level authorities tended to downplay the extent of violence in the state and in the city; for example a state’s justice attorney declared that violence in Sinaloa was not very different from violence in any other part of the world, and Governor Renato Vega in 1993 declared that the existence of drug traffickers in the state was “pure hallucination.” Vega’s position stood in sharp contrast to the position of his competitors for the gubernatorial election in 1992, who presented drug trafficking and violence as two priorities of their political platforms (Alvarado 1997, p. 285-289) At different moments the cohesion of the state in Culiacán was challenged by the intervention of the military in antinarcotics policies, and by occasional anticorruption efforts in state and municipal police. These interventions, as predicted in the argument I propose, led to temporary increases in the visibility of violence. Public figures were assassinated in Culiacán in 1977 (Alfredo Reyes, underchief of judicial police in Sinaloa and Gustavo Samarco military advisor of the Operación Cóndor), 1983 (Municipal police Commander Jaime Cota Felix and chief of Secretariat of Public Security Ceferino Ojeda)   245   and 1990 (human rights defender Norma Corona, who had publicly spoken about the protection that security officials and governor Toledo Corro provided for Felix Gallardo). These cases followed law enforcement operations such as the removal of commands in the municipal Secretariat of Public Security in 1989, which challenged the protection criminals had in the state. According to a former state Attorney, the Army during the 1980s and 1990s was a strong institution, the only one that “destroys plantations and confronts gunmen [gavilleros]” (Lazcano y Ochoa 1992, p. 250). Yet, despite its historical presence, the Army did not control directly public security in Culiacán and this mitigated the fragmenting effects of the “cleaning” operations of police forces that occurred throughout the 1980s until the early 2000s. Thus, short of major political transformations, these temporary shake-ups were not enough to challenge the high frequency-low visibility configuration of violence in the city. In sum, between the 1980s and 2000s the situation in Culiacán was relatively stable: high levels of violence derived from disputes in the criminal world; low levels of visibility as criminals strived to maintain protection within the state, and stable power structures that maintained a cohesive security apparatus. The situation changed radically in 2008, as a “war” exploded in Culiacán. 5.2.2. “The War started on April 30th, 2008” On April 30th 2008, a battle erupted in the Colonia Guadalupe, an upper-middle class neighborhood in Culiacán. Around the corner from the Governor’s house, a joint   246   military-police group engaged in open fire with armed gunmen at a “safe house;”230 the balance of the operation was 13 people detained, the seizure of an extensive arsenal, munitions, police uniforms, $379,980 dollars, and enormous bullet holes in the façade of the house that remained in place and were still observable three years later, when I conducted fieldwork in 2011. At the same time this battle ensued, two members of the State police, that were supposedly attending a back-up request made by the soldiers and federal police battling in the “safe house”, were intercepted and killed by sicarios. Soon after, two members of the state judicial police (ministeriales) were killed as they attempted to capture the sicarios who attacked the state police.231 The original official version of the events was that the battle started when federal police attended a call to conduct a search on the safe house. Many other unofficial versions emerged in the following days about what in reality created the confrontation; the most widespread was that the Beltran Leyva family was retaliating against other members of the Sinaloa DTO, Joaquin “Chapo” Guzman and Ismael “Mayo” Zambada, for a betrayal that led to the capture of Alfredo Beltrán Leyva, aka, “El Mochomo,” on January 20, 2008. Violence existed prior to April 30th but the visible events that occurred that day marked the initiation of the war in the public imaginary. All my interviewees in Culiacán pointed to that day as the day that signaled a radical transformation in the “honor codes”                                                                                                                 230 A safe house is a place used to hide arms, drugs, hostages, and in many cases to conduct violent acts. 231 For journalistic accounts of violent events that day Noroeste. 2008. “Arroja hora violenta cuatro muertos en Culiacán” Noroeste, May 1. Noroeste. 2008. “Son doce los muertos por balacera en la Guadalupe”, Noroeste, May 1.   247   of violence. Visibility increased, as well as the sense that violence was more indiscriminate. In the words of a local journalist We were used to death before, there was constant violence but it was hidden, the victims appeared but it was not high impact violence […]. After April 30th 2008 a collective psychosis started. On May 5th five federal police were killed. We went from hidden violence to high impact violence, in which there were shootouts downtown. The drain finally overflowed.”232 The words of a local transit police I interviewed also confirmed this perception that the “codes” of violence were fundamentally broken in 2008 “when I started (as a police) there was a different mentality of the people dedicated to illicit jobs and they respected the cop, as they respected women and children, of course there have always been limitations if they tell you not to go to that colonia [neighborhood]. That respect does no longer exist, they kill women, children; the command that existed in those years has been lost and that’s why violence has proliferated.”233 This perception does not consider how innocent people were also targeted in the past, or how the interests of traffickers corrupted the action of authorities, but the interesting thing is precisely this idea that previous violence was more acceptable for the public. Between 2007 and 2008 homicides increased in Culiacán by about 180%, and the methods deployed in violence multiplied rapidly, including forms of killing aimed at exposing the evidence of violent attacks. Within a few days of the events of April 30, the number of “mantas” or banners appearing in the city increased, and their content provided some clues to explain the change occurring in the city’s violent dynamics. Most of the mantas that appeared within the first week of the battle in Colonia Guadalupe were                                                                                                                 232 Interview with local journalist, March 18th 2011, Culiacán 233 Author’s interview with transit police, Culiacán, March 13 2011.   248   directed to traffickers Chapo Guzman and Mayo Zambada, and to members of police corporations and the military. Members of the Beltran Leyva group signed many of these mantas accusing el Chapo and Zambada of collaborating with state authorities in the capture of Alfredo Beltran, and promising to take revenge. In one of them the threat read “this is the reverse for you Chapo, the UKA will fall on you; hold on tight Mayos and Zambadas you have awaken the monster of the skies.”234 Although some threatening mantas against army officials appeared back in 2006 in 2007, after April 2008 mantas and notes on corpses became daily routine: both frequency and visibility proliferated. The change was the result of an extreme rupture in the trafficking market of Culiacán, in the words of one of my informants, this dispute, unlike former disputes, was between “family members”, one that would generate more extreme retaliation: the five Beltran Leyva brothers allegedly introduced Guzman in to the secrets of drug trafficking, and later as he accumulated power, became his close associates.235 The recourse to visible forms of violence reflected the rupture in the protection racket provided by the state: corruption did not decrease, but became less predictable. The open threats from members of the Beltrán Leyva organization towards military and police reflected that they did not believe in having protection, as the security apparatus had fragmented after the involvement of the military in urban operations in Culiacán.                                                                                                                 234 The translation of these messages is extremely difficult given the extensive use of slang and grammar mistakes. The original message in this case read “ay te va la reversa Chapo, te va a caer la Uka (la verga sin peluka) agarrense Chapos y Zambadas, despertaron al mostro de los cielos.” Another message read “Dinastía Zambada, toda llena de cagada cuando no traicionan, matan a su misma gente como hicieron con Mario Quijote Deda Niño” (Zambada dynasty, all full of s.., when you don’t betray, you kill your own people, as you did with…). See Noroeste. 2008. “Más Narcomantas” Noroeste, May 8. 235 Ravelo, R. 2008. “Historias de Familia.” Proceso No. 1630 , January 27, reprinted on Special Edition 32, May 2011.   249   5.2.2.1. The declaration of war against drug traffickers and the fragmentation of the security apparatus in Culiacán The Army’s urban operations started in Culiacán after President Felipe Calderon declared war against trafficking organizations in 2006, and instances of visible violence proliferated. In September 2006 a corpse was tossed in front of the military compound with a threatening message for General Rolando Hidalgo Eddy, who had been commanding operations against “El Chapo” Guzman.236 If protection was no longer guaranteed, then the use of visible violence to demonstrate power and scare away the state and competitors acquired more value. The changes that followed the deployment of the military in urban operations illustrate that the structure of the security apparatus can change rapidly, especially when the military is suddenly deployed at a large scale in urban operations. On January 21st 2007, Felipe Calderón started the Operation Triangulo Dorado (Golden Triangle) that covered the region known by the same name and comprised by the states of Sinaloa, Durango, and Chihuahua, one of the major drug producing areas in Mexico. The operation, which implied the mobilization of 9,000 military troops with the aim of reestablishing “conditions of peace and tranquility in the region”, was the first step to extend into urban areas the historic presence of the military in Sinaloa. The military expansion was further extended in May 2008 when the government implemented the Joint military-police Operation Culiacán-Navolato as a response to the April 2008 events, which mobilized 2,733 forces including soldiers, federal police, marines and attorney                                                                                                                 236 Rio Doce. 2007. “El Otro frente: copan al General Hidalgo Eddy y se le vence el plazo.” Rio Doce, June 11.   250   general’s police.237 These operations had a fragmenting effect between enforcement agencies, as the military saw and treated local law enforcement as corrupt and as part of the problem and interagency conflicts ensued. The words of a local state cop reflected the tension that permeated relations between security agencies: “now the military are harsh, very strict with the cops.”238 As the security apparatus fragmented so did the protection rackets criminals secured within the state. And as the powerful trafficking organizations split, their networks of support within the police and the political class did too. This was reflected in the messages sent by members of the Beltran Leyva organization to members of the state and the federal police. On January 20th 2008, after the capture of Alfredo Beltran Leyva, national newspapers revealed that a list found in Beltran Leyva’s residence detailed local enforcement in Culiacán connected to him and to the Sinaloa DTO; soon after, most of the list members had been assassinated.239 All these changes occurred in the context of a decision made by the federal government (declaring the war on traffickers), therefore it is clear that national dynamics were inextricably linked to the local situation, and this in turn reflects the overlapping jurisdictions and sovereignties that coincide in a city (Davis and Libertun de Duran 2011, Davis 2011). Yet, the mobilization of the military created different dynamics in Culiacán compared to other cities in Mexico, and this emphasized the centrality of local state power in understanding security outcomes at the local level. In Culiacán the increase of                                                                                                                 237  El Sol de Sinaloa. 2008.  “Operativo Culiacán-Navolato; 2 mil 723 elementos federales.” El Sol de Sinaloa, May 14. 238 Author’s interview with state cops, Culiacán, March 25 2011. 239 Noroeste. 2011. “Vuelve el pasado: Regresa Alfredo Mejía Pérez a la PEP.” Noroeste, March 30.   251   homicide rates was significantly lower than in Ciudad Juárez (Chapter 6) and the visible forms of violence that were clearly directed against the state and rivals in 2008, seemed to change radically in 2009 and 2010, targeting victims that were a lesser source of public concern, such as common delinquents. This difference reflected that although the radical changes created by military intervention fragmented the state, one of the main pillars of cohesion in the previous decades still remained relatively intact: the structure of political power. In Culiacán, the disruptive effect of the military was mitigated by the fact that the military was a constant presence in the city at least since the Operación Cóndor, but most importantly, it was mitigated because the conflicts between different levels of government (President, Governor, and Mayor) and the fragmentation of electoral power was limited. The Governors during the explosion of violence, Juan Millan (1998-2004), Jesus Aguilar (2004-10), and Millan’s protégé Mario Lopez (2011-2016), appeared in all my interviews as powerful figures controlling the politics of the region. Meanwhile, the mayor of Culiacán was not mentioned as a prominent figure in security decisions in the city. The political allegiances of Mayors and Governors remained aligned, as had been the rule in the state,240 except in 2010 when PAN’s Mario López Valdez241 was elected as governor while a PRI mayor was in power. This concentration of political power maintained incentives for criminals not to defy the state through visible attacks against its members, and the visible violence that persisted in 2009 and 2010 seemed to serve a different purpose: to discipline or eliminate local criminality. This objective was aligned                                                                                                                 240 Except in 1996 when Sadol Osorio from PAN was elected as Mayor of Culiacán. 241 Although Lopez Valdez had been a PRI politician and won with the support of the PRI machinery   252   to the other characteristic that explained why the increase of violence in Culiacán was not as sharp as it was in Ciudad Juárez: the continued insourcing of armed force by traffickers. 5.2.2.2.Vertical control and “disciplining” of armed forces in trafficking organizations The armed apparatus deployed by the Sinaloa DTO in Culiacán since 2008 was insourced and remained clearly controlled within the hierarchy of the organization. Even though the age of both victims and perpetrators of violence seemed to decrease after 2008, reflecting an increasing involvement of young people in violence, youth gangs did not proliferate in Culiacán. In the words of a prominent local journalist “There are no gangs here, the last gangs existed in the 80s and were basically cholos. What we see here is cells within the organization, commanded by experienced killers at the service of capos… these cells are identified by the name of the killer who commands them, for example sicarios would say “I am the Machoprieto’s people” or “I am the Ondeado’s people.”242 The same local journalist further explained “there is a monopoly of crime in the hands of the Sinaloa Cartel. If I work for the cartel and I kill without permission, they kill me, because thus you contribute to heat up the plaza, you are too noisy. The bosses say “I don’t want to see you pistiando (drinking) or stealing”, and in the sectors they control there are less robberies and assaults […] the strategy of the Sinaloa cartel is not to make noise, if there are robberies, they attract attention.” The insourcing of violence was tied to a tight attempt to discipline low level                                                                                                                 242 Author’s interview with local journalist, Culiacán, March 18 2011.   253   criminality. Referring to the relations between killers and traffickers, an informant told me “there is a pyramidal structure in the Sinaloa cartel, identity is constructed face to face, respect for the boss is huge, because they know closely who is who. Now the Ondeado is starting to bother el Chapo precisely because he is too crazy.”243 The historic roots of traffickers facilitated their control over the population and explained their preference for insourcing violence in Culiacán, but it was not only a personal choice. In fact, as seen in Chapter 6, in Ciudad Juárez Chapo Guzman outsourced violence to youth gangs because he did not have any historic roots with the population in Juárez when he “invaded” the city. An important indication of the Sinaloa DTO’s attempt to discipline its armed force in Culiacán was the systematic targeting of petty robbers, which resembled the “social cleansing” that became common in Cali in the 1980s. In 2009 out of 35 notes that accompanied corpses, 30 were directed against common delinquents and car robbers (robacarros), the bodies were generally mutilated, and the remains exposed publicly with a note and a toy car placed on the body. The notes usually threatened others with receiving the same treatment if they continued robbing. In 2010 the content of mantas and messages on corpses continued reflecting the existence of a pattern of visible violence that did not target police forces or criminal competitors as in 2008, but seemed aimed at “cleaning up” delinquents. Out of 39 messages on corpses found in 2010, 7 were directed to “rats”, and 12 more to kidnappers, car robbers, rapists, 3 to people who “messed up with married people”, and another seven made a direct reference to the criminal groups confronted in the region: Beltran Leyva, Arellano Félix or Zetas. All                                                                                                                 243 Author’s interview with informant, Culiacán, July 1 2011.   254   these aspects suggest that in Culiacán criminal leaders attempted to control common delinquency and everything that resembled a social malady, an aspect that reflected, and at the same time contributed to keep, their legitimacy among elite sectors of the population, but above all, their control over their armed force. 5.3. Conclusion Cali and Culiacán are different in many respects, yet they share a common pattern: the persistence of violence that is very frequent, yet not very visible. This situation reflects both the existence of criminal organizations that drive up violence as they compete for control of a market, and the successful collusion between criminals and state structures at the local level. The protagonists of violence have constantly changed; as Hansen and Stepputat (2006) and Weinstein (2008) argue, in urban spaces, arrangements between the state and “informal sovereigns” like criminal actors are renegotiated frequently, and that is the case in these two cities. Yet, the process and the mechanisms that connect state and criminals in these cities have remained constant, thus explaining the persistent low visibility of violence. The histories of Cali and Culiacán illustrate that even when violence is not perceived as extreme and as a threat to national security, it is extremely damaging for societies. Even when the perception is that violence only affects criminals, it is likely that civilians suffer, because in order to maintain their turf criminals not only eliminate their competitors but whoever may represent a threat to their power. But even if violence were only limited to criminals, the legitimization of certain forms of violence contributes to weaken even more the state’s claim to the monopoly of force.   255   One interesting aspect that emerges out of the stories of Cali and Culiacán is that eventually, hidden violence can become visible: for example the disappearances that occurred during the dispute between Montoya and Varela in Cali were eventually unveiled. Yet, using less visible forms of violence seems to pay off in the short run as an effective way to avoid state and media attention. Another crucial aspect highlighted by Cali and Culiacán is that the mechanisms that sustain less extreme violence can be even more damaging for democratic institutions than those that generate extreme violence. For example, Colombian authorities claim that trafficking organizations were significantly weakened after the demise of the Cali and Medellín DTOs. But as the cases of the Norte del Valle traffickers and their successors illustrate, the nexus of criminals and the state still runs deep in state institutions even if the Colombian traffickers’ power in the global commodity chain of drug distribution has decreased, specially vis a vis Mexican traffickers. Their choice to use less visible violence is not only driven by their changing position in the trade, but rather by the mechanisms that connect them to the state. A troubling implication of this analysis is the unintended consequences that can emerge as states implement anticorruption purges or large scale military operations; this does not mean that states should abandon the possibility of combating corruption and violence, but rather that the conditions of these operations need to be carefully considered before they are implemented. Furthermore, short of transformations in power structures, these changes may only create short-term re-accommodations. The other side of the story is that for a cohesive state to become a source of effective policing rather than a source of corruption, checks and balances, justice, and reduction of impunity need to be in place.   256   The cases of Cali and Culiacán suggest that national contexts matter in different ways: homicide rates in Cali are significantly higher than in Culiacán because of the existence of armed conflict in Colombia; Culiacán’s explosion of violence in 2008 cannot be separated from the federal decision to declare war against trafficking organizations; criminal-state links in Cali are more unstable partially because national level institutions and media play a more active role in initiating high level judicial processes (impunity rates in general are similar in both countries, yet Colombia has seen a significantly higher number of high level judicial processes). More politicians have been condemned in Cali than in Culiacán, where no significant political figure has been officially investigated. Yet, these national dynamics cannot explain why criminal violence in Cali is so different from Medellín, why Culiacán’s is so different from Ciudad Juárez’s or Tijuana’s, and why dynamics of violence are so similar in Cali and Culiacán. Pervasive links between criminals and civil society have been crucial for criminals in both cities: the power of the Rodriguez Orejuela in Cali cannot be separated from the extent to which they gained legitimacy both among businessmen and among popular sectors of the city; similarly, the power of the Mayo Zambada and the Chapo Guzman are clearly linked to their connections within wide sectors of the society in Culiacán. Yet, these links are not sufficient to explain changing patterns of violence: in Cali, the Norte del Valle traffickers seemed to be less popular among the population, yet their violent tactics resembled those of their predecessors; in Culiacán, criminal’s legitimacy among civil society seems to have decreased, but it appears to be more a result, than a cause, of violence. Finally, the prevalence of low visibility violence in these two cities may explain   257   why systematic efforts to control violence have been weak or unsuccessful, because low visibility violence is less likely to generate strong civil society or state responses. This is not say that anti-violence programs have not existed, because significant efforts as those conducted by Mayor Rodrigo Guerrero in Cali in the 1990s are very relevant. Although it is not fair to assess anti violence efforts here, it is important to highlight that different forms of violence may determine the type of civil society and state responses. Less visible violence may lead to more limited civil society and state responses as in Culiacán, where for over twenty years one woman, Mercedes Murillo, has been the only visible face of human rights and civilian anti violence efforts in the city.   258   CHAPTER 6. BEYOND A BORDER TALE: THE DIVERGENT VIOLENT TRAJECTORIES OF CIUDAD JUÁREZ AND TIJUANA Ciudad Juarez and Tijuana have been marked by their status as cities bordering the United States. From the Mexican-American war in 1846, through the emergence of “vice districts” during the prohibition of alcohol and prostitution in the US in the 1920s, to the expansion of drug markets fueled by the proximity to the largest market for illegal drugs in the world, the evolution of illegality in these cities has been shaped by border controls, transborder movements and exchanges, and by the effect of US legislation and foreign policy. Yet, the histories of drug violence in Ciudad Juárez and Tijuana have not been completely parallel and their border status cannot completely explain the different trajectories of illegality and violence they have experienced over time. How cities within the same country, similarly shaped by their status as border cities, by the powerful influences of economic globalization, and by diverse migrant populations, can nevertheless experience diverging patterns of drug violence? I explain these diverse patterns as the result of different interactions between the state security apparatus and the structure of the illegal market, and as the result of different strategies of armed coercion used by drug traffickers. When competition in the illegal market has increased, the frequency of violence has increased too, because criminals try to eliminate their rivals. And when the state has become more fragmented, criminals lose the key incentives to refrain from using visible violence. Within a fragmented state criminals do not have predictable protection benefits that they fear to lose, or do not fear being   259   effectively dismantled by the state. A fragmented security apparatus could neither become a reliable provider of protection or an effective fighter of criminals. The cross-city comparison presented in this chapter challenges the rhetoric, especially widespread among US enforcement agencies, that border cities are inevitably dangerous and prone to violence, and a source of national security threat to the United States.244 The comparison also allows me to challenge explanations of violence that focus on the effect of national government decisions, such as the declaration of war against trafficking organizations by President Felipe Calderón in 2006. While the national government’s mobilization of the military deepened state fragmentation and violence in the two border cities, I show that the local security apparatus and local criminal strategies mediated the effect of national policies: Insourcing and a more cohesive security apparatus led to more limited and short-lived increases in the frequency and visibility of violence in Tijuana than in Ciudad Juárez, where outsourcing and fragmentation became prevalent. The first section of the chapter focuses on Tijuana. It describes how the Arellano Felix trafficking organization (AFO) established a criminal monopoly in the second half of the 1980s, and wove networks of protection within the state security apparatus, and these conditions created a low frequency-low visibility pattern of violence. As competition to the AFO grew in the 1990s, frequency also increased, and visibility remained relatively low because the security apparatus remained cohesive and criminals retained protection and thus, their incentives to hide violence. This situation changed radically in 2008 when conflicts within the AFO became an all-out-war and the state                                                                                                                 244 See Payan (2002) for a discussion of the imaginaries constructed in the United States about security in the border.   260   became more fragmented due to the deployment of military troops and massive changes in enforcement, thus creating a situation of high frequency and high visibility violence. Yet, I analyze why the situation in Tijuana in 2008 was less extreme than that of Juarez. I document how traffickers preferred to insource violence, and even to use elite young men as part of their armed apparatus, rather than gangs composed of marginalized youth. Such decisions determined the relative absence of extreme spikes of violence in Tijuana. The second section focuses on Ciudad Juarez and shows how an illegal market controlled since the 1980s by the Juarez Drug Trafficking Organization245 (DTO) created the conditions for relatively infrequent violence. I also show how a cohesive state that provided stable protection reduced the incentives for the Juarez DTO to use visible violence. In 2008, the situation changed radically as the Sinaloa DTO invaded the territory of Juarez, and extreme conflicts among levels of government and enforcement agencies exploded, leading to a sharp increase in violence that transformed Ciudad Juarez into the most violent city in the world at the time. I describe how even though youth gangs always existed in Ciudad Juarez it was not until 2008 that they were actively deployed by trafficking organizations. Outsourcing is crucial to understand why Ciudad Juarez experienced a more extreme and longer spike of violence than Tijuana. 6.1. Tijuana 6.1.1. The cool guys move in: consolidation of the Arellano Felix Organization (1984- 1996) Tijuana is a relatively young city, located in the westernmost tip of Mexico, bordering the city of San Diego, in the United States. Founded in 1889, Tijuana quickly                                                                                                                 245 Also known as the Carrillo Fuentes organization   261   grew and expanded due to growing migration flows and to the emergence of the maquiladora industry246 in the 1960s, which attracted even larger masses of migrants in search of economic opportunities. Currently, it is the fifth largest city in Mexico with a population of 1,559,683 inhabitants, and is also the busiest border crossing in the world, with about 80,000 people crossing daily. The history of illegal markets in Tijuana dates back to the 1920s, when alcohol and drug prohibition in the US created profitable opportunities for rumrunners and drug traffickers in Mexico. The modern era of drug trafficking, however, is tied to a family originally from Culiacán, Sinaloa, that arrived to Tijuana in the early 1980s (Blancornelas 2002), the Arellano Felix family.247 Upon their arrival, the Arellano Felix allied with Jesús Labra Aviles, a long time trafficker in the region, and started to build up their power by carefully crafting connections with the state, and with the society. Relatively low levels of frequency and visibility in drug violence characterized the period since the arrival of the Arellano Felix brothers until 1987. In 1985, the homicide rate in Tijuana was only 1.5 per 100,000 inhabitants (figure 6.1). In 1989, the power of the Arellano Felix organization increased when, as described in chapter 5, the judicial police captured trafficker Miguel Angel Felix Gallardo, considered the leader of trafficking groups in Mexico. Felix Gallardo divided trafficking territories along geographic lines in order to keep the trafficking business afloat, thus giving more power to the Arellano Felix brothers in Tijuana. In 1993, when Felix Gallardo was transferred into a high security prison thus losing the ability to keep                                                                                                                 246 Maquiladoras are the assembly plants of transnational companies that relocate production sites to places where labor costs are low. 247 Composed by 10 siblings (seven men and three women) and the mother Alicia Felix, all of whom, to different degrees became crucial protagonists of drug trafficking in Tijuana.   262   controlling the business, competition between different regional leaders increased, especially the Arellano Felix on one side, and Joaquin “El Chapo” Guzmán and Ismael Zambada in Culiacán, on the other. Controlling territories was becoming more consequential for traffickers, as drug trafficking routes and the relation between Colombian and Mexican traffickers was changing. Inland trafficking routes through Mexico had become crucial in the 1980s as the United States cracked down on aerial routes used by Colombian traffickers to move drugs through the Caribbean. Mexicans thus became increasingly important for Colombians, and the financial stakes of drug trafficking grew. Competition struggles not only involved an interest in controlling territories, but also a personal difference in character: As Astorga (2005) puts it, Guzman and Zambada represented the rural image of the Mafioso, whereas the Arellano Felix, born in Culiacán’s middle class and educated, represented the image of the urban, businesslike trafficker. Between 1987 and 1996 violence associated with inter and intraorganizational disputes increased the frequency of violence and homicide rates grew to be slightly above the national average (see Figure 6.1). None of these disputes, however, implied the physical invasion of a rival DTO, or a breakdown of the power of the Arellano Felix. Violence had low visibility, and the frequency remained relatively stable at around 20 homicides per 100,000 inhabitants. Violence in this period was thus used to protect the organization from the growing threats posed by competition with traffickers from other regions aiming to gain access to Tijuana. The Arellano Felix also employed violence to discipline improvised traffickers that refused to pay “derecho de piso”, that is, the right to pass drugs through the territory   263   of the organization. According to Jesus Blancornelas, a journalist from Tijuana that dedicated a good part of his life to document and denounce the actions of narcotraffickers, the organization would charge a fee for small time traffickers that wanted to pass drugs through Tijuana and other neighboring municipalities in Baja California, although it would not tolerate attempts to cross larger amounts of drugs (Blancornelas 2002, p. 187). Figure 6.1. Homicide rates in Tijuana 1985-2010 Source: Author’s elaboration from SINAIS and INEGI 6.1.1.1. The narcojuniors: an uncommon strategy of armed coercion The competition between the Arellano Felix and other regional trafficking groups increased frequency but did not create spikes in violence because, for the most part, the Arellano Felix maintained their armed force within the organization and recurred to a different strategy to recruit armed force, using elite young kids rather than local youth gangs. The Arellano Felix recruited young men from Tijuana’s elite families as their hitmen and operators (Astorga 2005, p. 178). These young elite kids came to be known as   264   the narcojuniors,248 and according to journalistic accounts, by 1996 were responsible for many killings. But, why and how did traffickers recruit members of the elite to be their soldiers? Although there is not a good analysis of how this process started, it is likely that the ability of the Arellano Felix to penetrate the highest ranks of the society allowed them to establish contacts with elite families. The phenomenon of the “narcojuniors” became increasingly public as more members of elite families were linked to homicides and to the Arellano Felix DTO in the mid 1990s.249 Perhaps the most visible case was that of Alfredo Hodoyan who was captured in San Diego in 1996, and was linked to several murders. His declarations to Mexican and US authorities revealed the extent to which young members of the elite in Tijuana had become some of the most prominent killers for the Arellano Felix organization.250 This case gained even more attention as Alfredo’s brother, Alejandro, who was first detained in 1996 and made similar declarations, then disappeared in 1997 after being captured apparently by police officials. There is some evidence showing that the Arellano Felix did contract out youth gangs not in Tijuana but in San Diego. Specifically, traffickers recurred to services of the Barrio Logan gang,251 which origins date back to the 1940s in San Diego (Weisel 2002).                                                                                                                 248 In other places, such as Sinaloa, people use the label narcojuniors to refer to the children of drug traffickers, the second generation. 249 Blancornelas, J., and Gonzalez, H. 1996. “Juniors mataron mas que narcos” Zeta October 4 -10, p. 26A. and Blancornelas, J., and Gonzalez, H. 1997. “ Amigo de Xico Jr possible víctima del “Lobo” Hodoyán y Fabián “El Tiburón”. Zeta. October 11-17, P. 26A. 250 Blancornelas, J. 1997. “Descobijaron a los narcojuniors” Zeta, November 14-20. 251 Author’s interviews with academic from Tijuana, September 13, 2011 and human rights worker September 14, 2011.   265   However, this type of collaboration was sporadic;252 as a member of the Secretariat of Public Security in Tijuana observed “There [were] gangs, but they [were] not that involved in the business. That is because of the type of leaders, with the Arellano everything was managed through the narcojuniors. […] Here the enforcers of the Arellano Felix were the Logan gang, and the local gangs were only graffiteros, cholos, slackers, but that was not a problem.”253 The lack of outsourcing to youth gangs in Tijuana resulted from the combination between a short supply of gangs and a short demand of outsourced violence. The supply of youth gangs was limited despite a large proportion of young population in Tijuana. In the 1980s and 1990s there were very few reports of violence or petty criminality associated with gangs: in 1984 while 23 out of 131 violent events in Juarez were attributed to youth gangs, in Tijuana there was no mention of gangs linked to violent events.254 In general, homicides in Tijuana tended to occur among older and better-educated sectors of the population than in Ciudad Juárez (Martinez and Howard 2006). Between 1985 and 1997, 35% of homicides in Tijuana occurred among population 25 years old or younger, whereas in Ciudad Juárez the proportion was 44%255 thus illustrating the different composition of both victims and perpetrators of violence. There is some evidence that the lack of supply of gangs may also have resulted from the Arellano Felix’s attempt to eliminate gangs, seen as social malady, through social cleansing. As a testimony of this, in 1997 the mother of two narcojuniors who                                                                                                                 252 Members of Barrio Logan participated in an attempt on the life of journalist Jesus Blancornelas in November 1997 (Blancornelas 2002, 305). 253 Author’s interview with official from the Secretariat of Public Security. Tijuana, September 13, 2011. 254 Author’s dataset on drug related violence. 255 Author’s calculation based on SINAIS data.   266   disappeared in Tijuana, wrote a letter in which she denounced that Ramón Arellano Felix started his career as a ruthless killer when he asked the judicial police permission to “clean up” the city from the cholos that were killing each other in neighborhoods of the city (Blancornelas 2002, p. 301). Two factors limited the demand for outsourcing. First, the Arellano Felix strived to create the image of educated businessmen, resembling the strategies the Rodriguez Orejuela brothers used in Cali in the 1980s, as we saw in Chapter 5. Perhaps this interest in appearing as members of a business elite led them to employ the uncommon strategy of armed coercion based on elite kids rather than lower class operators. Furthermore, involving members of elite families in the organization provided crucial advantages, as authorities were likely to protect potential killers given their social status, and many of them had dual citizenship, and thus they could easily cross the border into the United States to smuggle drugs or to escape. Second, demand was limited because, as we will see below, the Arellano Felix DTO secured a stable source of protection in enforcement agencies, which mitigated the need to recruit large masses of young people. 6.1.1.2. Securing protection in a democratizing state In the 1980s and early 1990s a cohesive state security apparatus in Tijuana allowed the Arellano Felix organization to expand its connections with political and enforcement actors. Cohesion was possible because the power structures of enforcement agencies were very centralized in the Attorney General’s Office and allowed the establishment of stable protection, even in the face of growing democratization pressures in the city and the state of Baja California. State protection explained the low visibility of   267   violence: criminals hid violence in exchange for the license to operate and smuggle drugs. As can be seen in Table 6.1, during the early years of the Arellano Felix, there were few methods deployed to carry out violence and no high profile killings: the majority of victims whose occupation could be identified were either delinquents or presumed delinquents, non-professional workers, and violence rarely targeted prominent public figures such as cops or public officials. Violence was less visible also because most violent acts did not cause fatalities, or produced only one victim: in 1992, 53% of violent attacks produced no victims, while 43% left only one fatal victim. Table 6.1. Methods used in violent attacks, Tijuana 1984 % 1992 % 2002 % 2010 % Banner (without victims) Car bombs and explosions Combat with fire arm or explosives 3 4.05 3 1.8 8 2.0 Corpse with a note Corpse wrapped in blanket 2 2.7 9 5.4 24 6.0 Fire Head in cooler/frozen corpse 3 0.7 Levantón o paseo (forced 2 8.3 1 1.3 15 9.1 12 3.0 disappearance) Mutilation or incineration with note 1 0.2 Mutilation or incineration without 9 5.4 14 3.5 note Sexual violence 3 3.0 Sicariato (Drive by shootings) 5 20.8 12 16.2 27 16.4 97 24.5 Simple use of fire arm 7 29.1 35 47.3 45 27.4 161 40.7 Simple use of knives 7 29.1 16 21.6 15 9.1 19 4.8 Strangulation 2 2.7 1 0.6 5 1.2 Torture 3 12.50 3 4.0 34 20.7 37 9.3 Total number of events 24 74 164 395 Source: Author’s dataset on drug related violence, information compiled from El Sol de Tijuana newspaper (Tijuana).   268   An exception to the low visibility privileged by the Arellano Felix DTO occurred on May 24, 1993 when a catholic priest, Cardinal Juan Jesus Posadas Ocampo, was gunned down at the airport of the city of Guadalajara. Although the specifics and motivation of attack were, and remain, unclear to this day, the official version is that the attack was carried out by the Arellano Felix to eliminate rival Joaquín “El Chapo” Guzman, who was at the airport at the time, but the attack failed and the cardinal died accidentally in the cross fire.256 Because, as emphasized in Chapter 1 and 2, the visibility of violence can attract enforcement attention, the assassination of Cardinal Ocampo, even if accidental, increased enforcement pressure on the Arellano Felix Organization, and eventually led to the capture of Francisco Arellano on December 4 1993, and of Joaquin Guzman on June 9 1993.257 However, with the exception of this confusing episode, the Arellano Felix DTO’s preferred strategy was hiding violence. Hiding violence was crucial for traffickers to benefit from the police protection that allowed them to move drugs relatively freely, to influence the appointment of security officials, and in many cases, to obtain credentials as security personnel.258 The extent of the protection provided by state officials to the Arellano Felix DTO was evident in different episodes that linked members of the army, state and federal police, with drug traffickers in Tijuana.259 For example in 1985 the Army accused an agent of the state Attorney’s Office of protecting a marijuana cargo, and the situation was reported in the Zeta Magazine. Soon afterwards, the State Judicial Police bought the entire production of                                                                                                                 256 Aguirre, A and Cobián, F., 1993. Quienes mataron al cardenal? Proceso No. 867, June 14. 257 Guzmán escaped prison in an infamous cinematographic episode in January 2001. 258 Ravelo, R., 2002. “Estructura intocada” Proceso No. 1324, March 17. 259 Blancornelas, J., Gonzalez, H. 1997. “Dicho por el propio procurador Anaya: policias corruptos, no tontos” Zeta, August 15-21.   269   the Zeta magazine, thus revealing its interests in maintaining the corruption case out of public sight (Blancornelas 2003, p. 330). The cases that eventually became public revealed that protection networks were not just sporadic interactions, for example, in November of 1992, a Major of the Mexican Army, Juan Jose Sanchez Gutierrez, was detained and accused of serving as liaison between Benjamín Arellano and many army officials. The centralized structure of power within the state facilitated the large protection networks crafted by the Arellano Felix brothers. Yet in 1989 a path-breaking political and electoral event took place in Baja California, threatening the cohesion of the security apparatus. For the first time a member of an opposition party won a Governorship in Mexico after 70 years of rule of the Revolutionary Institutional Party (PRI). Ernesto Ruffo Apel, from the Partido de Acción Nacional PAN, won the governorship for the state of Baja California challenging the decades-old monopoly of power held by the PRI and marking a watershed in the democratization process in Mexico (Magaloni 2008). Ruffo’s election at the state level represented the peak of a process of increased competition between the PRI and the PAN that occurred in Tijuana between 1986 and 1992 (Rodriguez and Ward 1994, p.41). The election also initiated a pattern of confrontations between the PAN dominated state government, and the PRI dominated federal government. However, the Arellano Felix brothers managed to keep their protection within the state security apparatus despite the process of political transition. In fact, there were allegations that Ruffo himself had established a pact with the Arellano Felix, whereby the state guaranteed that traffickers would be protected from prosecution in exchange for not using the state of Baja California as a battlefield (Astorga 2005, p.   270   175). The allegations were never confirmed, but at least there is some clear evidence that Ruffo avoided getting directly involved in anti trafficking operations. Why didn’t growing electoral competition translate immediately into more fragmentation in the state security apparatus? Two major reasons explain the paradox. First, the PAN had little experience governing and a limited membership base thus many members of Ruffo’s cabinet were still PRI officials, notably the first attorney general260 and the state director of transit and transportation (Rodriguez and Ward 1994, p.58-61); thus, the electoral transition did not translate immediately in a bureaucratic makeover. Second, the new competitors at the subnational level were reluctant to take on drug trafficking enforcement as a local responsibility. According to journalists I interviewed, the power of the Arellano Felix was so ingrained when the PAN came to power, that the party did not have the power or the experience to confront them, and rather focused on the administrative challenge of governing without having much prior experience.261 While the political transition process did increase the independence of the state and local governments in areas such as budgeting, public works, and urban policies such as trash collection and transportation (Hernandez and Negrete 2002, Rodriguez and Ward 1994) it had a more limited impact on public security. The changes that the transition process created in governing and administering areas related to social and economic policy, were not immediately translated into governing in areas of public security.                                                                                                                 260 The Attorney General was then replaced by a member of PAN, Juan Francisco Franco, who was also later accused (but not investigated) of providing official identification cards of the State Attorney’s Office to hitmen of the Arellano Felix involved in a shootout at a club in Puerto Vallarta, and of supporting the Arellano Felix organization. Author’s interview with journalist, Tijuana, September 15, 2011. 261 Author’s interviews with journalists in Tijuana, September 15 and 16 2011.   271   The tension between Governor Ruffo and the Attorney General of Mexico Ignacio Morales Lechuga regarding violence in the state of Baja California, illustrated the reluctance of PANista politicians to crack down on trafficking. In a letter addressed to journalist Jesus Blancornelas in 1994, the Attorney General highlighted that “Ernesto Ruffo Appel minimized or undervalued the grave problem of local insecurity and attempted to explain the complex local phenomenon of drug trafficking as a federal responsibility.” Many years later, Ruffo responded to the letter alleging that he tried to coordinate actions with the federal government but “sadly, our federal government was extremely involved in the terrible problem of drug trafficking and its consequences” (Blancornelas 2002, p.83-86). This correspondence illustrates the growing tension that followed the political transition in Baja California, but also shows why in the short run the tension did not fragment the state security apparatus. The local government, at least initially, left the control of enforcement actions in the hands of the federal government and members of the PRI, while PAN members such as Tijuana’s mayor Carlos Montejo (1990-1992) avoided public actions against drug trafficking. The distribution of power in the state security apparatus continued to change incrementally throughout the mid 1990s without fragmenting it. Yet, during 1996 and 1997 a series of purges, rotations, and reforms in law enforcement led to momentary fragmentation in the security apparatus.   272   6.1.2. High visibility and low frequency violence in 1996-1997: Holding up corrupt arrangements Between 1996 and 1997 Tijuana experienced a pattern of violence that is not very common, when the visibility of violence is very high, but frequency remains stable. In those years, a series of high profile killings captured the attention of enforcement authorities both nationally and in Tijuana. These high profile attacks occurred as the frequency of violence and homicide rates remained relatively stable (20 in 1996 compared to 21 in 1995 and 19 in 1997). The targets of the attacks included at least 8 high-ranking government officials, 12 police officers, and journalist Jesus Blancornelas.262 The General Attorney’s Office was a notable target of these attacks, and eight members of the agency were killed, including two former delegates in Tijuana, two commanders of the Federal Judicial Police, and one former commander of the Federal Judicial Police.263 Most of these visible killings emerged as a retaliation for anticorruption policies and enforcement actions that destabilized protection arrangements between the state and criminals, and also as a way to deter enforcement agencies from continuing them. These anticorruption policies started with enforcement operations that directly affected properties and merchandise of the Arellano Felix DTO. In March 1996, the General Attorney’s Office conducted two operations (Operations Alacrán and Cancer) with support of the military that led to the seizure of large amounts of drugs and an intensive search of the properties of the Arellano Felix and their associates. Even though there                                                                                                                 262 Moore, M. 1997. “Drug Traffickers Thrive Amid Tijuana Terror,” Washington Post, 16 February, sec. A, p. 1. 263 Gonzalez, H., and Ortiz, F. 1996. “Tijuana: las ejecuciones alrededor de la PGR” Zeta May 24-Apr. 3. P. 28A.   273   were no captures in these operations, high profile attacks ensued.264 Then, in July 1996, a member of the National Institute for the Combat of Drugs (INCD) made an unprecedented announcement declaring that there were high levels of corruption in Mexican antinarcotics agencies. Such announcement led to the appointment of a member of the opposition party PAN in the General Attorney’s office, who, in turn, purged about half of the members of the Tijuana Attorney’s Office. The newly appointed Director of the Federal Police in Tijuana echoed denunciations of corruption in the Tijuana federal police in September 1996, and not surprisingly he fell victim of violence only a few days after swearing in for his position. These operations, rotations, and purges, fragmented the state security apparatus, and temporarily reduced the incentives of traffickers to keep their use of violence hidden, because the changes in enforcement agencies were increasing the risk that they could be captured or busted by authorities. Previous enforcement actions, like the one that led to the capture of Francisco Javier Arellano by the Attorney’s office in 1993, did not lead to retaliation because they were not part of a sustained enforcement campaign, and because the structure of the security apparatus had not been changed with rotations and purges265 as occurred in 1996. This case thus illustrates how in the context of an illegal market with relatively low levels of competition, visible violence can emerge in retaliation for state actions that threaten the structure of criminal organizations while disrupting the terms of interaction between traffickers and state officials.                                                                                                                 264 Blancornelas, J., Ortiz, F., Gonzalez, H. 1996. “La Mafia sigue adentro de la PGR.” Zeta, November 20- 26 . P. 26A 265 Blancornelas, J., and Gonzalez, H. 1996. “Juniors mataron mas que narcos” Zeta 4 -10 October, p. 26A.   274   Judicial investigations did not fully clarify if the authorities killed had been targeted because they refused to collaborate with traffickers, or because they did not abide to previous arrangements made with traffickers to protect them from enforcement operations.266 In each of the two circumstances, the killings were a form of retaliation to actions that affected the traffickers. The strong control that the Arellano Felix Organization had in Tijuana, guaranteed that at least in the short run, the frequency of violence would not escalate. The visible acts of 1996 and 1997 were effective in reestablishing arrangements between traffickers and criminal actors in subsequent years. 6.1.2. Re-accommodation and reorganization 1998-2007: silent deaths and low visibility violence After the episodes of highly visible violence in 1996 and 1997, confrontations between factions of the Arellano Felix DTO, and with outside competitors, continued growing. Homicides increased by 70%, from 19 deaths per 100,000 inhabitants in 1997 to 32 in 1998; up to 2007 the average frequency of violence was similar to the levels seen in 1987-1995 (a homicide rate of about 20 per 100,000), that is, frequent, but not extreme. Criminal competition grew with the elimination of two main leaders of the organization in 2002, which caused instability in leadership: Ramón Arellano died in February and the police captured Benjamín Arellano in March. These leadership changes, however, did not fragment the control of the Arellano Felix DTO completely because there was no outright territorial invasion by competing trafficking groups. The Arellano Felix DTO survived under the leadership of Fernando Sanchez Arellano, aka El Ingeniero, nephew of the former leaders. The remaining Arellano Felix brothers,                                                                                                                 266 1996. “La Mafia sigue adentro de la PGR” Zeta, 20-26 Noviembre. P. 26A   275   Francisco Javier aka El Tigrillo, and Eduardo, were later captured in 2006 and 2008 respectively (Ramirez Sanchez 2009, p. 376). Between 1998 and 2007 the armed coercion of the organization remained insourced, with some cases of violence still attributed to US gangs.267 The engagement of narcojuniors with drug trafficking became less prominent, yet there is no evidence that the narcojunior phenomenon was replaced by outsourcing to youth gangs. This explains why violence remained high, but relatively stable and without major spikes up to 2007. 6.1.3.1. The act of disappearing: hiding the evidence of violence In the late 1990s and 2000s, the Arellano Felix organization deployed violence to maintain control of the market, but became even more careful in hiding attacks. Perhaps the most illustrative example of the efforts to hide violence was the use of acid to destroy corpses of people killed by traffickers. The remains of the bodies would then be dumped into sewage systems or buried in the outskirts of the city by employees of the criminal organizations. Such techniques to hide the evidence of attacks only became public in 2009 when Mexican authorities captured criminal Santiago Mesa Lopez, known as “el pozolero268 del Teo”. Mesa confessed that over 10 years he and other collaborators, like El Cris, had disintegrated more than 300 bodies as they worked for the Arellano Felix, and especially their enforcers, el Efra, and then for el Teo.269 Mesa’s capture and his methods were quickly merged into the wave of brutality that characterized Mexico’s drug violence since 2006, yet his testimony revealed a more                                                                                                                 267 For example, on November 5, 2002, alleged members of a US based gang assassinated a local lawyer in Tijuana. 268 Pozolero can be translated as the stewmaker. Pozole is a traditional Mexican soup. 269 Declaración ministerial MESA III AP/PGR/BC/TIJ/217/09-M-III   276   intricate puzzle. For almost a decade, Mesa’s job was to eliminate the evidence of violence; later in 2008, when disputes exploded between leaders of the Arellano Felix organization, he was specifically instructed to, instead of burying the remains of the bodies, place them in the middle of a well-transited street, along with a note that read “this is what happens to El Ingeniero’s people. We are going to make pozole with you.”270 This example powerfully illustrates that the brutality that is sometimes seen as a marker of recent violence in Mexico, has in fact older roots. Brutality and visibility are not the same: traffickers were very brutal in their perpetration of violence, yet, before 2008 they were not very visible. Disappearing bodies was an old technique used by traffickers to hide their responsibility on deaths, and it became known to the public only when traffickers changed their strategy and decided to expose the evidence of violence, as we will see later. A few human rights organizations concerned by the tragedy of disappeared people thought to be victims of drug traffickers in Baja California denounced this hidden violence. Yet, there is no reliable information about the number of people disappeared throughout the late 1990s and 2000s. Existing information comes from relatives of victims who individually, or in weakly organized groups, tried to find out about the fate of their disappeared relatives. This is the case of Cristina Palacios de Hodoyan, whose son was first detained in 1996 and then disappeared in 1997. Since 1997 she has been trying to find out the whereabouts of her son, albeit unsuccessfully. In her quest, she met other victims of disappearance and according to the information she collected, about 20 disappearances were recorded between 2002-2007 and about 132 between 2007 and                                                                                                                 270  La Jornada. 2008. “Hallan en Tijuana tres cadáveres en tambos con ácido” La Jornada October 1.     277   2011.271 Yet, according to another activist and victim of the forced disappearance of a relative, the official number of investigations for disappearances initiated between 1989 and 2007 was 430, and it is likely that some of these victims could have been disappeared by “pozoleros.”272 According to him “The stuff of the pozoleros has a long history, the Arellano Felix in the 80s brought a group from Israel to teach them how to disintegrate bodies.”273 The key point is that although disappearances became public in the late 2000s, they had been occurring for a long time in Tijuana and were clearly aimed at eliminating the evidence of trafficker’s violence. Other evidence of press reports in the dataset of violence I constructed for this project documents traffickers’ efforts to hide violence in this period of criminal re- accommodation. In 2002, newspapers reported the increasing finding of bodies with signs of torture, usually naked, and wrapped in plastic. Out of 45 bodies found with signs of torture and attributed to criminal disputes in 2002, 27 were found in the outskirts of the city, in abandoned lots, rivers, or private residencies, as opposed to public areas, and streets. The brutality of violence seemed to be increasing, but for the most part, criminals tried to hide such brutality. For example on October 20 2002, the press reported the finding of a corpse in the sewage system of a neighborhood in Tijuana. The corpse had been mutilated, and the body parts with the exception of the head, were distributed in three plastic bags.274 In another similar case, in November a woman found the remains of                                                                                                                 271 Author’s interview, Tijuana October 21 2011. 272 Author’s interview with the relative of a victim of disappearance in the mid 1990s, Tijuana October 21 2011. 273 Author’s interview with activist and relative of a victim of disappearance, Tijuana October 21 2011. Similar information appeared on Blancornelas, J. 1997. “Entrenado en Israel: Comando narco mató y remató a Federales” Zeta 21-27 November. 274 Author’s dataset on drug violence, reported in El Sol de Tijuana, October 20 and 21, 2002.   278   an incinerated body at the base of a hill275 and the following day workers found another incinerated body in an abandoned lot.276 In most of these cases, bodies were disposed of in private, or abandoned spaces, as if the intention was that the bodies would not be found. The attention to take extra steps to hide violence resulted from the persistence of protection networks that criminals had within a cohesive state. Between 1998 and 2007, the PAN consolidated its domination of elections, winning local and state governments throughout the period, but remained relatively isolated from anti trafficking efforts. At the same time, traffickers managed to maintain protection within the state. A notable case is that of PRI politician and owner of Mexico's largest sports betting company, Jorge Hank Rhon, son of Carlos Hank, a powerful stalwart of the old PRI elite. He became mayor of Tijuana between 2005 and 2007. Since the late 1990s Hank Rhon allegedly appeared in DEA files as launderer for the Tijuana DTO (Bowden 2002, p. 258) then in 2009 the US consul in Tijuana expressed in a leaked cable to the Secretary of State that “Hank is widely believed to have been a corrupt mayor and to be still involved in narco- trafficking.”277 The case of Hank Rhon, who was captured on arms trafficking charges and then released by the Army in 2011, exemplifies the extent to which the Arellano Felix maintained state protection up until the 2000s. In 2007, the situation changed radically.                                                                                                                 275 El Sol de Tijuana, November 4, 2002. 276 El Sol de Tijuana, November 12, 2002. 277 Petrich, B. 2012. “Librarse de la sombra de Hank Rhon, el mayor reto para el PRI, según EU” La Jornada June 6.   279   6.1.4. The explosion of violence and the return to a peaceful equilibrium (2008-2010) In 2007 Tijuana experimented a historic increase in the frequency and the visibility of violence, caused by the explosion of conflicts for the control of the trafficking market, and by the increased fragmentation in the security apparatus accelerated by the deployment of the military in January 2007. Homicide rates increased by 245% going from 13 in 2007, to 42 in 2008, and 72 in 2009. Violence proliferated as conflicts within the Arellano Felix DTO finally split up the organization between two factions, one led by Teodoro Garcia Simental aka El Teo, a high level killer for the Arellano Felix Organization, and the other led by Fernando Sanchez aka El Ingeniero (Ramirez Sanchez 2009). The split became even more consequential for violence as El Teo allied with Joaquin Guzman and the Sinaloa DTO, which had initiated an aggressive strategy of territorial expansion.278 The visibility of violence also increased as the number of methods employed in violent attacks proliferated (see Table 6.2.) including shootouts in the middle of the street, the mutilation of bodies exposed in public spaces with notes, and the use of banners where criminals sent messages to their rivals and to the state. Public authorities became more frequently targeted, and in 2010 eight cops and soldiers died in violent attacks.279 Some of the banners that appeared throughout the city revealed the nature of the confrontations between criminal organizations, for example most of the banners where the content was                                                                                                                 278 Some versions identify April 20 2007 as the day when the war exploded because El Teo refused to account for his actions to the AFO (Sanchez Ramirez 2009). This date, however, is not a clear marker of violence. In Culiacán, by contrast, April 30 2008 clearly marks the initiation of violence. 279 Author’s dataset on drug related violence   280   made public in 2010280 were directed against “Aquiles,” a former member of the Arellano Felix DTO who allied with El Teo and the Sinaloa DTO. Visible violence ensued after the deployment of the military in the city fragmented the security forces. In January 2007, the federal government deployed 3,296 members of the Military, Marine, and Federal Police, for the Joint Operation Tijuana.281 At the onset of the operations, there were confrontations between the governor Eugenio Elorduy and the commander of the military zone, Sergio Aponte Polito. Aponte Polito was publicly critical of corruption in state police, and in a couple of occasions he paid advertisements in newspapers, where he denounced corruption of the state police (Valenzuela 2009, p. 317).282 Thus, the mobilization of the military generated tensions between levels of government, and between enforcement agencies that complicated coordination and protection. These tensions however, were controlled relatively quickly. 6.1.4.1. A fragmented state quickly recovers In 2008, Jose Osuna Millán was elected Governor of Baja California. As he took office, he declared that General Aponte was in charge of public security, and informally gave him the power to coordinate enforcement in the city. In turn, the Mayor of Tijuana, Jorge Ramos, publicly supported this decision. These declarations signaled that political authorities were jointly granting military authorities the command of public security in Tijuana, and the political decision contributed to improve the relations between the                                                                                                                 280 El Sol de Tijuana only revealed the content of 9 out of 10 messages in 2010. 281 La Jornada. 2011. “Operativos conjuntos detonaron homicidios en seis entidades” La Jornada, May 8, p. 9. 282 Author’s interview with journalist, Tijuana, September 15, 2011.   281   Governor, the Mayor, the Chiefs of Public Security, and the Commander of the Military Zone. General Aponte Polito named current and former military officials in key public security positions; the most notable appointment was that of Liutenant Colonel Julián Leyzaola, who became Chief of Public Security in Tijuana. Leyzaola had been commander of the State Preventive Police, and this experience allowed him to also establish collaboration with state police when he was appointed at the municipal level. In Leyzaola’s own words “I had complete support from the state police, because I trained them myself.”283 Leyzaola became the most visible face of what would be known as the “Tijuana model”, which referred to the relative quick demise of violence in the city. Leyzaola also became infamous due to his controversial, but very public, discourse regarding criminals as “the dirty ones”, who did not deserve any consideration of individual rights.284 The paradox is that while military presence prompted state fragmentation, military power also catalyzed renewed cohesion, as described by an official of the Secretariat of Public Security “In 2007 there was a substantial change in the municipal police because of the people who came to direct the agency, the military. […] We removed security chiefs and placed commanders with military formation […] part of the success here had to do with the ability of the municipal police to initiate actions through a hard command which was that of Colonel Lieutenant Leyzaola. The Mayor made a commitment because he let the chiefs [of public security] work […] With                                                                                                                 283 Author’s interview with Julián Leyzaola, Ciudad Juárez, July 20, 2011. 284  Cave, D. 2011.  “A Crime Fighter Draws Plaudits, and Scrutiny.” The New York Times. December 23.   282   General Aponte Polito everything started, he named Leyzaola and everything started there. The Governor had no other choice but to ride along.”285 When General Aponte Polito was relieved of his command in August 2008,286 General Alfonso Duarte Mujica replaced him and became the de-facto commander of all enforcement agencies. When Duarte arrived, governor Osuna Millán again declared publicly that the security situation was in the hands of the General and created a group of interinstitutional coordination headed by the General. Military influence continued with General Duarte, and in fact, most people in Tijuana would grant Duarte, rather than Aponte Polito, the recognition for facilitating enforcement coordination. In any case, military influence was crucial to induce coordination in enforcement agencies, creating a de-facto single command. As one expert put it, the single command “has not been created constitutionally, but it exists de-facto because there is a subordination of the civil government to General Duarte.”287 Besides military influence, political coordination between government levels was crucial to induce cohesion in the security apparatus because Mayor, Governor, and President were all from the same party, the PAN. As a businessman put it “there has been political will of the governor and the municipal presidents, the proof is that all of them accepted a former military in each Secretariat of Public Security at the municipal level.288” Additionally, the limited role of the federal police in Tijuana as well as the reduced number of municipalities in the region (there are only five municipalities in Baja                                                                                                                 285 Author’s interview with official from the Secretariat of Public Security, Tijuana, September 13, 2011. 286 Marosi, R. 2008. “Antidrug General Ousted” LA Times, August 9. 287 Author’s interview with academic and public security expert, Tijuana, September 12, 2011. 288 Author’s interview with businessman, Tijuana, October 19, 2011.   283   California Norte) facilitated regaining cohesion and made it easier for politicians and commanders to coordinate. Under these circumstances, by 2010, homicide rates had decreased by 50% and instances of visible violence had decreased considerably. 6.1.4.2. Continued insourcing The cohesion in the security apparatus facilitated the successful implementation of high profile enforcement actions that had an immediate impact on violence, for example the capture of some of the most violent criminal leaders. El Teo was captured on January 12, 2010 and his associates Raydel Lopez “El Muletas” and Manuel Garcia “El Chikilin” were detained on February 8, 2010; Juan Francisco Sillas, main enforcer for El Ingeniero, was captured in November 2011. These high profile captures had an immediate and notable impact on violence because, to a great extent, even at the height of the violent peak, the warring factions in Tijuana continued insourcing violence, and thus the capture of the main enforcers of the organization had a greater impact on reducing the violent behavior of those under their command. Lack of outsourcing, even in the context of a war, partially reflected a path- dependent effect, as the legacy of the Arellano Felix’s use of narcojuniors prevented the reproduction of gangs in the city289 and those gangs that existed were not directly linked to drug trafficking organizations;290 this situation remained true after the explosion of violence in 2008. There were reports linking the US based gang Mexican Mafia to                                                                                                                 289 The feedback effect of outsourcing contributing to the proliferation of gangs can be seen in the case of Medellín in the 1980s, described in Chapter 4. 290 Zeta. 2007. “Pandillas: una amenaza en Tijuana.” Zeta 1709. December 29 2006 to January 4 2007.   284   trafficking organizations in Baja California (OAS 2010) and also a gang known as Los Sureños based in San Diego291 but gangs and trafficking organizations were not clearly aligned. In 2009, and in line with the tendencies seen in other years, the proportion of homicide victims between 15 and 24 years was lower in Tijuana than in Ciudad Juarez, whereas the proportion of homicide victims ages 35 to 44 was far higher in Tijuana.292 Liutenant Coronel Leyzaola described to me the lack of outsourcing to youth gangs as follows “In Tijuana crime was organized, specialized […] there were gangs, but not linked to crime, they [the traffickers] had their own killers and gangs were used only sporadically.”293 The capture and elimination of violent leaders thus had an immediate impact on violence because the killers were under more direct control of the traffickers. As expressed by a human rights worker “We have gangs in the periphery, but they have not been used by organized crime, there has not been a need for an army like in Juarez where they [gangs] are linked as labor force. Here we have 50,000 unemployed and a cell of the Familia Michoacana [a trafficking group] but once the cell of the Teo was captured, they calmed down.”294 Insourcing was crucial to explain why Tijuana experienced a less pronounced spike of violence than Ciudad Juárez, why enforcement actions disintegrated criminal’s armed coercion more rapidly, and why outsourcing is an important variable to understand drug violence.                                                                                                                 291 Author’s Interview with DEA agent, El Paso, Texas, March 28, 2012. 292 And this difference cannot be readily attributed to a simple difference in the proportion of young population in each city. In fact, according to the 2010 census statistics, the proportion of the total population 24 years or younger is 47% in Juarez and Culiacan, and 57% in Tijuana. 293 Author’s interview with Lieutenant Colonel Julian Leyzola, Chief of Public Security in Ciudad Juarez, and former Chief of Public Security in Tijuana. Ciudad Juarez, July 20, 2011. 294 Author’s interview with human rights activist in Tijuana, Tijuana September 14, 2011   285   6.1.4.3. The return of low visibility violence The reduction in homicides experienced in 2010, has been dubbed the “Tijuana Model,” and it did in fact reflect important successes in law enforcement. Other factors played a key role, such as an increasing involvement of the business sector in monitoring violence trends, in shaping the image of city as a safe place for business and tourism, and in shaping the public representation of violence. However, the reduction of violence is still subject to two major issues of debate. The first issue of discussion derives from evidence suggesting that other less positive factors also drove the reduction of violence, chiefly a criminal pact between warring factions. The pact ensued as the Sinaloa DTO gained control, and the most violent opponent faction of the Arellano Felix DTO was dismantled. In these circumstances, the remaining leaders of the Arellano Felix DTO, renamed as the Tijuana or the Sanchez Arellano DTO, conscious of their weakness, preferred to strike a pact that would still allow them to operate. They preferred to return to the old system in which they charged other criminal organizations, such as La Familia Michoana, for the right to pass drugs through their territory. The second issue of debate is the extent to which violence has really decreased. There is evidence suggesting that despite the reduction, violence still prevails, but mostly hidden. According to prominent human rights workers, the apparent success of authorities in reducing violence in 2010 was no more than a change in the nature of attacks, which are not as brutal and visible as they were at the beginning of 2008, and rather occurred in the outskirts of the city, involving low level drug dealers rather than high ranking criminals. These killings were usually carried out with simple gunshots,   286   rather than using methods that could expose the evidence of violence such as beheadings or the use of notes and banners on dead bodies.295 In the words of a photojournalist “there has always been violence […] there will always be assassinations, but what changed were the high impact crimes. The 12-people conboys that you used to see became 3 guys on a Sedan that shoot three times. It became more discrete. And you don’t see the violence anymore; it’s difficult to see it on the street. I was very surprised by the amount of sirens you could hear before. The high impact crime, which is the one that worries the society, has decreased a lot, and we believed that sensation.”296 An official of the State Attorney’s office echoed this statement “currently we have no homicides at broad daylight, but clandestine homicides. The weapons also change, they use strangulation but not high power weapons.”297 These statements reveal the importance of assessing the visibility of violence because apparent peace does not equate an absence of violence. Although positive, the relapse of violence in Tijuana masks a reality of less visible violence, the militarization of public security, and the interest in maintaining a positive image of the city, even at the cost of not reporting persistent forms of violence. In sum, the reduction of violence in Tijuana resulted from the ability of authorities, especially military ones, to reduce fragmentation in the security apparatus, and this in turn led to more successful enforcement operations. The reduction also derived from the adaptation of traffickers, who, after suffering some crucial blows, preferred to return to a more peaceful arrangement, that would still allow them to operate                                                                                                                 295 Author’s interview with human rights activists in Tijuana, Tijuana September 14, 2011; Mexicali October 20, 2012. 296 Author’s interview with photojournalist, Tijuana, September 12, 2011. 297 Author’s interview with official of the State Attorney’s Office. Tijuana, October 17, 2011.   287   without forcing the state to act, but which required them to hide their use of violence. Finally, the lack of connections between traffickers and youth gangs facilitated both the effectiveness of state actions and the strategic adaptation of criminals. These three elements -reduced state fragmentation, strategic adaptation of criminals, and insourcing- as we will see in the next section, were not existent when violence skyrocketed in Juárez in 2008. In 2010, Ciudad Juárez became the most violent city in the world. The meteoric rise in violence that the city started to experience since 2008 was unexpected but not unpredictable. The next section, describes how during most of the 1980s and 1990s one major organization –the Juarez DTO- dominated the illegal market in Ciudad Juárez, and such domination persisted despite some leadership changes. The Juarez DTO also managed to maintain stable networks of protection, and such combination created a relatively stable pattern of low visibility and low frequency violence. Towards the late 1990s competition in the illegal market started to increase, and thus the frequency of violence. In 2008, three sharp changes combined to create the perfect storm of violence in Juárez: a war exploded between the Juarez DTO and the Sinaloa DTO, both groups outsourced violence to the gangs that had for years existed in the city independently from trafficking groups, and the militarization of the city, along with confrontations between government authorities, immersed Ciudad Juárez in the worst episode of violence in its history.   288   6.2. Ciudad Juarez Ciudad Juárez is located in the northern Mexican state of Chihuahua, bordering the U.S. city of El Paso, Texas. It is also strategically close to the tri-state intersection between Chihuahua, Texas, and New Mexico. Since the 1960s, life in Ciudad Juárez has been marked by the proliferation of the export maquiladora industry that attracted large immigration flows to the city and made it the poster child of globalized production processes and their negative consequences (Jusidman et al 2008; Barraza et al 2009; Bowden 2010). In Juárez, the maquiladora industry, which employed women almost exclusively, radically transformed family dynamics and created social clashes and transformations (Ampudia Rueda 2009, Swanger 2007). The flow of jobs that made Ciudad Juárez, a city of “full employment” for many decades, did not improve the quality of life in the city as most of the employment was low skilled and labor conditions were precarious (Ampudia Rueda 2009). The city has also been characterized by lack of infrastructure, ingrained corruption in the political system, and what Herrera (2008) calls government abandonment. These political and social factors have been often identified as crucial elements to understand the extreme public security crisis that Ciudad Juarez experienced since 2008. While these elements are essential to understand the social reality and the structural causes of violence, they are not enough to explain the specific timing of violence, which has varied depending on the interactions between states and criminals, as we will see in the following pages. For example between 2000 and 2002 the city experienced an increase in unemployment from 2.0% to 3.7% but no increase in violence (Quintana   289   2008). By contrast, the spike of violence in 2008 coincided with an increase in unemployment from 2.5% in 2005 to 3.2 in 2007, 4.9 in 2008, and 8.4 in 2009298 (Quintana 2008, Ampudia Rueda 2009). Thus, while the profound socio-economic problems Juarez faces can foster violent dynamics, the timing and the proximate causes of violence in the short run can be better understood by looking at the changing interaction between the state security apparatus and the structure of the drug market. The story thus should starts with the consolidation of the Juarez DTO in the early 1980s. 6.2.1.The emergence and consolidation of the Juarez DTO: years of peaceful coexistence 1984-1994 Ciudad Juarez has a long history of smuggling (Campbell 2009; Grayson 2009, p.23) and a trafficking tradition that seems to be older, more ingrained and more territorialized than in Tijuana (Poppa 2010). This tradition can be traced back to the 1920s with traffickers like La Nacha, later towards the 1970s and 80s with Pablo Acosta Villamizar, aka El Pablote, and aka El Arabe. In the early 1980s, a commander of the Federal Security Directorate (DFS), Rafael Aguilar Guajardo, and Rafael Muñoz, organized previously dispersed drug smuggling groups in what came to be known as the Juarez DTO.299 In 1985, Ernesto Fonseca Carrillo aka “Don Neto”, an old time trafficker, sent his nephew Amado Carrillo to take control of the “plaza” of Ojinaga, another municipality in the state of Chihuahua.300 Carrillo301 allied with Rafael Aguilar, Rafael Muñoz, and El                                                                                                                 298 Unemployment figures at the city level are not available, these numbers correspond to state figures, but expert consider that the trends are similar at the city level. 299 Perez-Espino, J, and Paez, A. 2009. “Los Carrillo llegan a Juárez”. El Universal April 3. 300 Gutierrez, A. 1996. “Hombre Inteligente, respetado y querido” In Proceso, No. 1003, January 22.   290   Greñas (Campbell 2009, p. 97) to establish an almost monopolistic control of illegal markets in Ciudad Juárez, and such control allowed the maintenance of low frequency violence between 1984 and 1993. In this period, homicide rates were significantly lower than the national homicide rates, at an average of 8.63 homicides per 100,000 inhabitants (figure 6.2.). Figure 6.2: Homicide rates in Cd Juárez 1985-2010 Source: Author’s elaboration, data from SINAIS and INEGI Amado Carrillo consolidated a network of aircraft, aerial routes, and landing strips that gave him the nickname of The Lord of the Skies (El Señor de los Cielos). Carrillo also mastered relationships with Colombian trafficking organizations. In the words of the former Commissioner of the National Institute to Combat Drugs (INCD) for Ciudad Juárez, the control Carrillo achieved in Ciudad Juárez was comparable to a tyranny, as he didn’t let anyone “smuggle a gram that [was] not his.”302 A former official                                                                                                                 301 Who, as many other older traffickers in Mexico, was originally from Sinaloa. 302 Gutierrez, A and Ramirez, I. 1997. “El Escurridizo”, in Proceso, No. 1054, January 13.   291   of the Dirección Federal de Seguridad (DFS) who arrived in Ciudad Juárez in the late 1980s described the control of the Juarez DTO as follows: “When I arrived in Juárez everything was peaceful, it is the most important frontier and everything was coordinated, there was only one group there.”303 6.2.1.1. Gangs proliferate but outsourcing remains limited During these early years, armed coercion by the Juarez DTO remained within the organization, even though there was a significant supply of non-violent youth gangs in the city. In the 1980s, gangs proliferated in Juarez, providing young kids who adopted dress styles and shared cultural norms, and that were known as cholos, with a source of socialization and identification. As Valenzuela (2009) notes, even though cholos were rarely violent, the public, and especially the media, often stigmatized them as perpetrators of crimes; crime suspects were usually described as “males with cholo appearance” (Valenzuela 2009). Despite these stigmas, in this period there was no systematic connection between gangs and criminals. An analysis of violent events in newspapers in 1984 illustrates that there was no outsourcing of violence to youth gangs in the 1980s. Street gangs were mostly involved in minor offenses like street fights and robberies, and rarely in violent homicides. In fact, deaths emerged more often as a result of the population’s reaction to cholos rather than as a result of the gangs’ own violence. Thus, for the most part youth gangs, also known as barrios, were not involved in violent crimes despite widespread perception in the                                                                                                                 303 Author’s interview with former judicial police in Juarez (DFS), Mexico City, 2011.   292   opposite direction. Some instances of young gang members working for traffickers existed but they usually worked as lookouts or carriers of drugs304 but not as assassins. In Ciudad Juarez during this first period, the lack of outsourcing mostly resulted from a lack a demand, rather than a lack of supply of youth armed forces. There was a supply of youth gangs, but they were not connected to trafficking organizations. Even more important, youth gangs were not violent in this period. Traffickers in Ciudad Juárez, unlike their peers in Tijuana, did not try to eliminate gangs through social cleansing, but they did not demand their labor either. This is because the relative monopoly of the Juarez DTO made violent disputes with competitors sporadic, and made it unnecessary to garner massive numbers of soldiers for war. Additionally, as we will see below, the protection that the Juarez DTO secured in enforcement agencies also reduced the demand for outsourced violence. These networks of protection also explained why violence, if employed, was not very visible. 6.2.1.2. State sponsored protection in Ciudad Juárez Amado Carrillo and the Juarez DTO gained protection from broad sectors of the security apparatus, and the cohesion of the security apparatus made such protection very stable. The Juarez DTO established a thick network of support among police forces and enforcement officials; in fact Army General Jesus Gutierrez Rebollo, head of the INCD, was convicted in 1997 for protecting Carrillo, and for selectively enforcing the law against his Tijuana rivals. When Gutierrez Rebollo was detained in February 1997, soon after he had been named head of the INCD, and eights days after White House drug-                                                                                                                 304  Rodriguez, S. 2011. “Truncaron sus estudios de secundaria… y se volvieron el ‘terror’ de 3 Colonias.” El Diario, December 26.       293   policy chief, Gen. Barry R. McCaffrey, described him as ''an honest man and a no- nonsense field commander,''305 the Mexican Minister of Defense declared that for seven years the General had been protecting cocaine shipments for Amado Carrillo, in exchange for vehicles, real state, and cash.306 In the 1980s, the cohesion of the security apparatus in Ciudad Juarez depended on the power of the General Attorney’s Office (PGR), and its police branch, the Federal Security Directorate (DFS- Dirección Federal de Seguridad), which had the ability to reduce conflicts and confrontations among enforcement agencies and created incentives for criminals to hide violence. In the words of a former member of the DFS who worked in Ciudad Juárez “The General Attorney’s Office coordinated everything, sometimes even the Army. There was order even in the relation with the State’s Attorney. There were executions, but the bodies disappeared, they were not thrown out in the streets.”307 Protection derived also from state and municipal police who controlled, and sometimes even extorted, criminals; according to a human rights worker “the weight of the judicial police was enormous, they controlled delinquency.”308 Limited political competition and the domination of the PRI, which as in other cities in Mexico created an extraordinary level of political coordination, also facilitated cohesion. Yet, in 1983, the PRI experienced one of its earliest electoral defeats, when Francisco Barrio from the PAN won elections and became the mayor of Ciudad Juárez.309                                                                                                                 305 Golden, T. 1997 “Misreading Mexico: How Washington Stumbled” The New York Times July 11. 306 Reed, S. 1997. “Certifiable: Mexico’s corruption, Washington’s indifference” The New Republic. Available at http://www.pbs.org/wgbh/pagesAfrontline/shows/mexico/readings/newrepublic.html Accessed [9 November 2012] 307 Author’s interview with former member of the DFS. Mexico City, June 13 2011. Emphasis is mine. 308 Author’s interview with human rights worker, Ciudad Juarez, May 17 2011. 309 In the same year the PAN also won the Mayor of Chihuahua City, the state capital.   294   As in the case of the first PANista government in Baja California, over time Barrio’s election became consequential for the democratic transition in Mexico, but in the short run it was rather limited in its ability to challenge the PRI domination at the state and federal level (Rodriguez and Ward 1992, p. 60). In fact, after this PAN victory, the PRI returned to power in 1986. The early PAN victory had a rather limited effect on challenging the cohesive power structure. Yet, it contributed to create a pattern of active intervention by the Mayor that was not common to other localities of Mexico that did not have early PAN or opposition victories. An opposing Mayor was more likely to challenge the governor than a Mayor from the same party, and thus clashes between PAN mayors and PRI governors proliferated (Rodriguez and Ward 1992, p.74). Just as the electoral success of the opposition in the Presidency of Mexico in the 2000s created a new relationship between Presidents and Governors providing more power to the latter (Ochoa-Reza 2004), in Juárez the early electoral transition at the municipal level started to transform the relationship between mayors and governor. However, the structure of power in the security apparatus did not change radically in the next period, when violence started to be more frequent and the Juarez DTO faced more competition. 6.2.2. Increased competition in the criminal market and upsurge in violence (1994-2007) In 1993 Amado Carrillo eliminated one of his associates, Rafael Aguilar. While Aguilar’s assassination cleared up the way for Carrillo to become the sole leader of the Juarez DTO, it also generated more grievances and competition within the organization and did not eliminate competition with rivals from Tijuana and the Gulf DTO (Lupsha   295   and Pimentel 1997). Homicide rates almost doubled from 14 to 25 between 1993 and 1994. Throughout the period 1994-2007 homicide rates remained at an average of 21 homicides per 100,000 inhabitants (figure 6.2). In 1997 Carrillo died in mysterious circumstances, while he was getting facial plastic surgery. After his dead, his brother Vicente Carrillo took over the control of the organization and although he managed to maintain the dominance of the Juarez DTO, he dealt with opposition from leaders like Rafael Muñoz, who wanted to gain control, until his death in 1998.310 On September 10 2004, Rodolfo Carrillo, another member of the Juarez DTO, was killed in Culiacán, allegedly by Joaquin el Chapo Guzman, leader of the Sinaloa DTO, as part of his effort to take over the control of trafficking routes in the border.311 This episode generated confrontations between traffickers that used to operate as business partners, and these confrontations in turn, maintained frequent violence. Yet, the confrontation did not become and all-out war until late 2007, when the Sinaloa DTO physically moved forces into Ciudad Juárez, as we will see later. The frequent violence remained relatively bounded without major spikes, because, as in the former period, there was no outsourcing of armed coercion to youth gangs, although gangs continued proliferating. During the late 1990s and early 2000s the Municipal Public Security Secretariat reported that there were 300 gangs with more than 15,000 members in Ciudad Juarez (Pineda and Herrera 2007), some of them involved in local drug dealing, but for the most part there was no evidence of systematic connections between traffickers and gangs, specially regarding the outsourcing of criminal coercion. Violence also remained strategically hidden as the networks of state protection and                                                                                                                 310 Dillon, S. 1998. “Mexican Drug Lord Found Slain.” The New York Times, September 12. 311 Ravelo, R. 2004. “La hora del desquite.” Proceso No. 1455, September.   296   control persisted. 6.2.2.1. Disappearances in Ciudad Juárez Between 1994 and 2007 disappearances became a prevalent tool to eliminate the evidence of violence. In 1997, Sam Dillon, journalist of the New York Times, wrote an article revealing many crucial aspects of the disappearances in Ciudad Juárez. He described how in many cases victims appeared to have been detained by police or soldiers, who were likeky contracted by traffickers to eliminate their rivals and debtors.312 The participation of federal police in disappearances was even acknowledged by Governor Francisco Barrio (1993-1998) and by a special envoy of the Attorney General who declared that municipal police in Juárez had told him “the desert surrounding Juarez is a vast cemetery, a huge mausoleum full of corpses.”313 Police participation in disappearances revealed the systemic links between state and criminals, and illustrates the strategic value of hiding violence. Available evidence suggests that victims of disappearance could have been killed and buried in mass graves. Thus, both the evidence of violence and the perpetrators were almost impossible to track. A relative of a victim interviewed by Dillon stated: “they [perpetrators of disappearance] use a large oven to cremate the disappeared”. In fact, in November 1999, a protected witness led the FBI to locate two places were mass graves supposedly existed. The FBI and the Attorney’s office initiated a joint enforcement                                                                                                                 312 Dillon, S. 1997. “A toll of disappearances in Mexico’s war on drugs.” The New York Times. October 7. Also Hervella, J. 2006. “Ciudad Juárez: Residencial narcofosas”. Jesus A Blanco: El Paso, TX. In some cases military or police could use dissapperances to hide their excessive use of force in interrogating suspects of criminal activity. 313 Ibid.   297   operation on the site, but could only recover the remains of nine bodies. In subsequent years, limited discoveries of mass graves314 kept revealing that there was a silent perpetration of violence in the city that could only be uncovered when members of criminal groups became informants of enforcement agencies. It is not clear how many victims disappeared during the late 1990s and early 2000s in Ciudad Juarez. Jaime Hervella, an accountant and resident of El Paso, who created an association to search for his friend’s disappeared son and daughter-in-law, started to count cases of disappearance. In 1997 Hervella published an add inviting relatives of other victims of disappearance to report their cases; in the course of a few days he received more than 50 reports, and by 2003 he had compiled 196 cases. By 2003 the authorities officially acknowledged 181 cases.315 Even though Mr. Hervella and other victims’ efforts forced the Attorney General to send special envoys to investigate disappearances, there was never an official resolution to any case. Rumors abounded, but as Bowden puts it, for the most part, “the entire matter will exist outside the talk of governments” (2004, p. 259). Violence thus remained hidden, first because evidence was only occasionally recovered, but also because the status of the victims as possibly involved in criminal activities, led authorities and even the families, to pursue these cases less seriously. According to Mr. Hervella, the documentation of cases became more difficult over time because people would “bring information but then would tell us that they did not want anything published […] the                                                                                                                 314 One of the most notorious was the discovery in 2003 of 11 buried bodies in a house that belonged to a leader of the Juarez DTO. This case was particularly memorable as some of the killings carried out in this house had been witnessed by a DEA undercover agent. 315 Dillon, S. 1997. “A toll of disappearances in Mexico’s war on drugs” The New York Times. October 7. Also Hervella, J. 2006. “Ciudad Juárez: Residencial narcofosas”. Jesus A Blanco: El Paso, TX.   298   families did not want to get involved because they feared retaliation, but also because they had doubts [about their relatives] and preferred not to get involved.316” Many of these cases of disappearance lost primacy vis-à-vis the systematic killing and torture of women known as feminicides, which took center stage and made the city sadly famous since 1993. Feminicides were characterized by the deliberate use of sexual violence and atrocity aimed at making evident the gender-related dimension of the attacks (Monarrez Fragoso 2009, CEDAW 2005, Washington Valdez 2006).317 Even though the investigation of feminicides was characterized by impunity and corruption, it encouraged the organization of civil society to demand justice to an extent that was far greater than in the case of disappearances supposedly associated to drug trafficking. The discursive construction of victims of disappearance in Ciudad Juarez as possible criminals thus contributed to keeping the evidence of violent trafficking disputes hidden. Hidden violence also allowed traffickers to maintain networks of protection, which despite political changes remained ingrained in political and enforcement agencies. 6.2.2.2. Persistent state protection After the election of PAN’s Francisco Barrio as Chihuahua’s governor in 1992 political challenges to PRI rule grew (Alba Vega 1997, Guillen Lopez 2002) but networks of protection to criminals remained strong. The political change fell short of a real challenge to the old structure of power, and in fact Governor Barrio was criticized for                                                                                                                 316 Author’s interview conducted in El Paso, Texas, March 27, 2012. 317 The nature, perpetrators, and reasons for this horrendous violence are still subject to debate. Investigations were plagued by impunity, and the cases solved were perceived to rely on highly partial investigations. Existing analysis suggest that feminicides were perpetrated with varied motivations by a multiplicity of actors including state officials, drug traffickers, serial killers, and relatives of the victims, and permitted by the government’s complicity or lack of real commitment to stop the tragedy.   299   maintaining cozy relations with the PRI federal government (Rodriguez and Ward 1994, p. 123). Although Barrio carried out massive firings that disjointed the judicial police, the new police forces quickly reinstalled relations with criminal forces. In fact, during the period of mayor Gustavo Elizondo (1998-2001) the Juarez DTO formed an enforcement branch that comprised former and current transit police, municipal and state police, and some federal police, known as La Linea.318 Over the years, the criminal group La Linea recruited mainly from among the ranks of former and current police, and established tight networks of support that combined judicial officials, cops, taxi drivers, and contract killers.319 Cohesion in the security apparatus also survived because the power of the Attorney’s office was still crucial to prevent conflicts between enforcement agencies, which could derive, for example, from military presence in enforcement operations in Ciudad Juárez. Military presence was usually covert and did not employ the massive numbers, the street operations, and the raids that became routine years later in 2008. In 1995, the Attorney General’s Office carried out an antinarcotics operation targeting the Juarez DTO, which deployed 72 personnel that operated as temporary agents of the Attorney’s office (Artz 2011). According to a former official of the Attorney General’s Office, the operation “was coordinated with the commander of the military zone and with the Attorney’s office.”320 The operation lasted about four months and although initially it was successful in forcing the Juarez DTO to reduce smuggling and hide, eventually it                                                                                                                 318 Author’s interview with human rights worker, Ciudad Juarez, May 17 2011. 319 Sullivan, J, and Logan, S. 2012. La Línea: network, gang, and mercenary army. Free Republic, August 12, Available at http://www.freerepublic.com/focus/f-news/2762765/posts Accessed [26 November 2012] 320 Author’s interview with former high level official at the Attorney General’s Office (PGR) in Mexico. Mexico City September 9 and 26 2011.   300   failed because the soldiers ended up involved in corruption scandals. The same official of the Attorney’s office described “the cartel did not have informants so they tried to co-opt our soldiers; then they threatened, and then they started to execute people. […].”321 In this operation, the cohesion in the security apparatus, materialized in the leading role of the General Attorney’s Office, allowed the institution to prevent the conflicts of operation that could derive from the simultaneous presence of military and judicial police, and to quickly cover-up the operation’s failure, that was not excessively publicized. In this context, state officials had the upper hand in regulating the terms of drug trafficking, as Astorga has intelligently put it (2004), but the Juarez DTO also had power to push the state to withdraw when the enforcement operations became too costly for them. Of course, the limit of such push was determined by the organization’s fear of losing state protection, and thus the criminal’s reaction to this operation was not nearly as violent as the reaction caused by military operations since 2008. Thus this operation illustrates why not all enforcement operations cause similar reactions among criminals and why it is crucial to consider the distribution of power within the security apparatus in which particular enforcement operations take place. A cohesive state can organize and control enforcement operations more easily (in this case troops were quickly withdrawn after problems emerged), and thus it creates incentives for criminals to regulate their violent reaction to these enforcement actions. Had the Juarez DTO’s reaction been too violent, it would have created more incentives for the cohesive state to crack them down. In sum, the persistent cohesion and protection within the state limited the scope of violent                                                                                                                 321 Author’s interview with former high level official at the General Attorney’s Office (PGR) in Mexico. Mexico City September 9 and 26 2011.   301   reactions to enforcement, and created incentives for criminals to hide violence, for example through the disappearances described earlier. 6.2.2.2. Momentary spikes of visibility Towards the end of the 1994-2007 period, political changes and enforcement changes that reduced the power of the Attorney’s General Office at the national level, generated moments of state fragmentation and short episodes of visible violence in Juárez. Some late episodes of feminicide in 2001, that seemed to be deliberately publicized, could be interpreted precisely as visible violence perpetrated by traffickers in reaction to the state’s withdrawal of protection through rotation and purges. Feminicides constitute an extremely complex phenomenon, and more than 20 years after cases became public, the causes of this tragedy remain far from being clarified and justice remains elusive. Thus any detailed interpretation of the motivations and perpetrators of feminicides transcends the focus of this dissertation, and even remains elusive for the many excellent analyses of this topic. The crucial point here is that existing evidence suggests that some of these killings could have been perpetrated directly by, or with the protection of, drug traffickers and corrupt policemen, and especially the cases in 2001, were deliberately made public to retaliate and even publicize the incapacity of the local government. In 2001 eight bodies were found in an empty property known as the Cotton Field, a property located in a very busy intersection of Ciudad Juárez. Forensic evidence suggested that not all the women dumped in the lot had been killed at the same time, as if the perpetrators had deliberately stored victims at different moments until they decided to   302   visibilize the corpses.322 Some analysts have interpreted this particular case as a strategy to hurt then Governor of Chihuahua Patricio Martinez, as this case drew attention and criticism to the persistent impunity surrounding the city and to the government’s lack of effective investigation (Washington Valdez 2006, p. 40). A prominent human rights defender in Ciudad Juárez considers that some of the feminicides coul have been carried out by former police officers as a reaction to massive firings and rotation of municipal and state cops that took place during the administration of Governor Martinez, who was a very vocal critique of narco-corruption in the state.323 Exposing cases of feminicide publicly, precisely when the government was under heightened scrutiny for the lack of results in feminicide investigations, was a deliberate attempt to publicize the government’s inability and lack of power vis a vis powerful criminal actors in the city, and thus perhaps persuade the government not to be so vocal against drug traffickers. Other events in 2001, prominently a murder attempt on the life of Governor Martinez in January, added to the general sense that drug traffickers were concerned with Governor’s Martinez public stance rejecting drug trafficking. In 2005 similar instances of high visibility violence occurred when prominent public figures including the chief of the State Investigation Agency Jose Antonio Torres, and at least 13 federal, municipal, and state cops, were assassinated in Ciudad Juárez and                                                                                                                 322 In general the extreme cruelty and clear sexual connotations in cases of homicide of women can be defined as visible violence. The bodies found in the most publicized cases of feminicide were usually mutilated, stabbed, strangled, with signs of sexual violence and in some cases with signs carved on the back of the victim’s bodies. Yet, the case in 2001 was different because it deliberately exposed the evidence of the attacks. 323 Lafranchi, H. 2001. “Mexico declares war against drug traffickers”. The Christian Science Monitor. January 30.   303   in the capital city of Chihuahua.324 A member of an extinct state judicial police, Ignacio Sanchez, was kidnapped and later found dead, blindfolded, with barbwire around his neck, and a note. In January 2006 lawyer Sergio Dante Almaraz, who had publicly denounced the involvement of trafficking organizations and state complicity in feminicide cases, was also assassinated. Visible killings also included another commander of the state judicial police, Arturo Nassar Contreras. Interestingly, this visible violence occurred as the frequency remained relatively unchanged at about 16 homicides per 100,000, thus suggesting that the change in the visibility of violence was not the result of heightened criminal disputes. Rather, the events occurred after several members of the state police, including the state police chief, were arrested on charges of collaborating with the Juarez DTO, leading Governor Jose Reyes Baeza to create a new State Secretariat of Public Security in 2004. In 2005 there were also constant rotations among the Directors of the Secretariat of Public Security with two directors in a single year (Pineda and Herrera 2007, p. 117). These police reform efforts were not successful in the long run as the new agencies ended up being equally permeated by corruption (Zavaleta, Chavira and Sanchez 2007), but in the short run the rotations unsettled the terms of interaction between criminals and state actors. In sum, in the period between 1994 and 2007 the frequency of violence in Ciudad Juarez increased due to growing disputes between the Juarez DTO and outside challengers. The visibility, however, remained relatively low as a result of the persistent cohesion of the security apparatus that allowed networks of state protection to function in stable and predictable ways.                                                                                                                 324 La Jornada. 2005. “Cuatro ejecutados en Chihuahua y Sinaloa; uno era jefe policiaco de la primera entidad” La Jornada. December 25.   304   Changes in political competition increased the use of scandals as a tool for electoral competition, and mayors and governors were frequently accused of protecting drug traffickers, as was the case with governors Francisco Barrios (1992-1998) and Jose Reyes Baeza (2004-2010). Purges and rotations in enforcement agencies created momentary changes in the interaction between criminals, cops, and politicians. These changes, nonetheless, fell short of changing structures of power; accusations of wrong- doing never became full-fledged judicial processes (Zepeda Bustos 2010), structures of enforcement remained corrupt, and the state security apparatus remained relatively cohesive, thus reducing incentives for criminals to use visible violence systematically. Despite efforts to clean-up police forces, governors, and mayors remained reluctant to acknowledge publicly the existence of either violence or drug trafficking. In 2004 for example, Governor Reyes Baeza criticized the international attention to the situation of violence in Ciudad Juárez because it damaged the city’s image.325 Despite cosmetic efforts to reform police forces, and despite political competition, the structure of state power did not change radically and corruption remained ingrained and cohesive. The situation changed radically in 2008 when it became impossible for authorities to deny the situation in Juarez, and confrontations between political and enforcement authorities proliferated. 6.2.3. The perfect storm of violence in Ciudad Juárez (2008-2010) Since 2006 Mexico has experienced the worst episode of drug violence in its history, which as of December 2011 had claimed more than 60,000 victims of drug                                                                                                                 325 Amnesty International. 2005. Mexico: Justice fails in Ciudad Juarez and the city of Chihuahua. February 2005. Available at http://www.amnestyusa.org/node/55339?page=show [Accessed November 23, 2012].   305   violence including more than 30 mayors,326 80 journalists,327 and hundreds of military and police personnel. In Ciudad Juárez such violence took on an even more extreme dynamic: the average homicide rate between 2007 and 2010 was 140 compared to a national average of 13 for the period, and a peak homicide rate of 250 occurred in 2010. The most striking aspect was that homicides increased by 713% over the course of only one year (2007-2008). (Figure 6.2). The explosion of the conflict between the Juarez DTO and the Sinaloa DTO led to a radical change in the structure of the criminal market and to an extreme increase in the frequency of violence. By late 2007 members of the Sinaloa DTO decided to “invade” Ciudad Juárez. Until then, the confrontations between the two DTOs implied that there were killings in each other’s territory, but not an actual mobilization of forces. A criminal arrested by the Federal Police, Noel Salgueiro,328 confessed that in 2007 he was commissioned by the Sinaloa DTO to invade Juárez with 500 gunmen, entering the city from the South with the aim of displacing the Juarez DTO. After the Sinaloa gunmen entered the city in January 2008, violence ensued, federal forces moved in, and the city ended the year transformed by unprecedented death tolls, fear, brutality, and activities, that like extortion, were relatively unheard of. Since Amado Carrillo’s death, important changes had occurred inside the Juarez DTO that made it more vulnerable to the Sinaloa DTO’s invasion, and made the conflict                                                                                                                 326 BBC News. 2011. “Mexico’s Drug War: Lessons and Challenges”. BBC News. 31 December. Available at http://www.bbc.co.uk/news/world-latin-america-16337488 327 Committee to Protect Journalists, Accessed on November 22 2011. 328 Comunicado Conjunto Sedena, PGR. 2011. “El Ejército Mexicano detiene a Noel Salgueiro Nevarez (a) “El Flaco Salgueiro”, fundador y líder de “Gente Nueva”, grupo delictivo de la organización criminal “Guzmán Loera.”” October 5. Also, Moore, G. 2011. “The Legacy of Sinaloa Cartel Lieutenant El Flaco”. Insight Crime, October 17. Available at: http://www.insightcrime.org/investigations/the-legacy-of-sinaloa- cartel-lieutenant-el-flaco   306   even more violent. These changes included the relaxation of the organization’s hierarchical structure that strengthened splinter sectors of the Juarez DTO that then allied with Sinaloa, and a heavy investment in armed power329 that was deployed to its fullest capacity during the war. La Linea, the enforcement branch formed in 2004 by former police members, became the main military and most violent branch of the Juarez DTO. Thus, the invasion of Sinaloa traffickers radically transformed the criminal market from a quasi-monopolistic into a competitive criminal market in dispute. 6.2.3.1. Visible violence and the extreme fragmentation of the state The deployment of military forces and federal police in Ciudad Juárez in 2008, as part of the declaration of war against drug trafficking organizations by Mexican President Felipe Calderón, unleashed a process of rapid and profound fragmentation of the state security apparatus: extreme conflicts between elected authorities at different levels of government, (municipal, state and federal) undermined the protection that criminals obtained in previous periods; fragmentation between security agencies became more acute; and constant rotation and purges of enforcement officials deepened the distrust and confrontation between enforcement agencies, and between criminals and the state. In this context, previous arrangements of protection between criminals and state officials became unpredictable; if a security agency protected a criminal organization another agency could persecute it. Furthermore, networks of protection in the state split up, and criminal actors targeted security agencies, in a pattern that confronted not only criminal organizations against each other but also against different enforcement actors. Traffickers                                                                                                                 329 Ravelo, R. 2010 “Los Carrillo, bajo protección military.” Proceso No. 1732, January 10.   307   from both sides accused and targeted enforcement officials they saw as supporting their rivals. Visible violence represented a key tool for intimidation and retaliation, both of competitors and of state officials. Violence in Ciudad Juarez thus became a semantic system where, as Campbell notes, traffickers used violence to signal “who is either in charge or attempting to take charge of a particular plaza and to explain why specific individuals have been or will be slaughtered” (2009, p. 28). The protagonist of the documentary El Sicario, which describes the life of a paid killer in Juarez, explained that each method of killing signaled a different message “throw it face up, that’s one message, face down, it’s another message, cut a finger and put it in the mouth, another message, cut a finger and introduce it in the anis, message, take out the eyes, take out the tongue.”330 Extreme fragmentation in the security apparatus ensued in March 2008 when the federal government initiated the Operacion Conjunta Chihuahua (Joint Operation Chihuahua) deploying 5,000 military troops and 2,700 federal police to Ciudad Juárez as part of the strategy to dismantle trafficking organizations. There was no institution like the Attorney’s Office in the past, with the power to control and coordinate the multiple enforcement actors in the city. The massive numbers of troops deployed in urban operations clashed with the municipal police, as the federal government and the military saw municipal police as part of the problem due to its pervasive corruption and links with La Linea. Human rights complaints multiplied. Furthermore, some sectors of the military                                                                                                                 330 Rosi, G. 2011. Documentary “El Sicario, room 164” Film Forum, 2011. The documentary is based on the article with the same name by Charles Bowden, in Harper’s Magazine 2009.   308   resented being deployed for operations they were not prepared to handle, and that could undermine their prestige, even though they would not say this publicly.331 The Army conducted anticorruption operations in the municipal police, fragmenting its schemes of operation, as well as the traditional schemes for criminals to deal with cops. In the words of a human rights defender “the Army’s logic is to defeat and disarm every citizen in the city, and the first to be disarmed were local cops, not killing them, but demobilizing them.”332 In this context, municipal police forces were reluctant to work with the Army, and mutual distrust grew as traffickers declared a war against police officers who refused to collaborate with them, or that they considered as protectors of their rivals. In February 2009, the Juarez police chief quit after several officers were slain and traffickers threatened that a police officer would be killed every 48 hours until the police chief quit. The attack against local police agencies motivated local authorities to bring in Liutenant Colonel Julián Leyzaola as the Chief of the Secretariat in Public Security, due to his record of action in Tijuana. Yet, even after Leyzaola’s arrival, confrontations persisted. In 2010, a total of 149 law enforcement officials from municipal, state, and federal agencies were killed. During 2010 three federal police and two municipal police appeared killed with notes like “I’m a cop at the service of ____”333, in a pattern of accusations and rumors that linked the federal police with the Sinaloa DTO, and the municipal police with La Linea and the Juarez DTO. The content of the messages found                                                                                                                 331 Author’s interview with a high level official of an international organization. Mexico City February 21 2011. For an analysis of the military’s perception about being deployed in public security operations before the Calderón government see Artz 2011. 332 Author’s interview with Human Rights worker, Ciudad Juárez, 2011. 333 This information is based on a review of the local newspaper El Diario in 2010. The mentioned example occurred on August 13, 2010, and the press clip omitted the name of the organization signing the note.   309   near victims of violence also reflected a more systematic targeting of state officials than in the other Mexican cities I analyze, and more specifically of members of the Federal or the Municipal Police. In 2010, 8% of homicide victims in Ciudad Juarez were members of police or military agencies (55 victims) while in Culiacán the proportion for the same year was 2.8% (16 victims) and in Tijuana 1.9% (8 victims).334   Confrontations between levels of government further fragmented the security apparatus. Initially, both the Governor of Chihuahua Jose Reyes Baeza and the Mayor of Ciudad Juárez Jose Reyes Ferriz supported military deployments. However, soon after, confrontations between the Mayor and the Governor developed. The Governor did not fully support the mayor’s security policies and seemed to do very little to improve the situation335 and the local mayor was perceived to be weak and detached from the reality in the city. In fact, he was widely criticized for taking residence in the border city of El Paso after receiving death threats. When Reyes Ferriz stepped down as Mayor in 2010, he denounced the lack of collaboration of the state government in general, and especially in prosecuting criminals that municipal authorities had captured.336 Such confrontation reflected the pattern of empowerment of local mayors mentioned earlier that became evident when violence skyrocketed, federal operations started, and the major actively confronted other authorities. However, his actual power to do something about the                                                                                                                 334 This analysis is based on the author’s dataset on drug related violence. Here it is important to emphasize that newspapers only identify a small numbers of victims’ occupations. Yet, it is reasonable to think that a cop or a soldier is more likely to be identified than a civilian. It is also important to note that these numbers correspond to 2010 but it is possible that the situation was different in 2009 in Tijuana when trafficking conflicts were far more visible 335 Author’s interviews conducted with two business leaders and a prominent human rights defender in Ciudad Juarez, May 2011. 336 El Mundo. 2010. “Confesiones de un alcalde de Ciudad Juárez: no hay colaboración entre gobiernos.” El Mundo. 13 October.   310   situation was limited.337 The confrontation also illustrated how fragmentation in the security apparatus complicated not only protection to criminals, but also the effective enforcement of the law. The confrontation between Mayor and Governor in Ciudad Juárez at the onset of violence did not appear in all the cities where the federal government started military operations. The impact of the federal government’s decision to militarize was mediated by local power and political conditions. As expressed to me by an official of the Federal Secretariat for Public Security explaining why military deployments had different dynamics in Culiacán (described in Chapter 5) and Ciudad Juárez “Sinaloa (where Culiacán is located) is an entity formed by 18 municipalities [...] and in that sense the relationship between mayors (presidentes municipales) and the Governor is easier than in Chihuahua where there are 67 municipalities. Distances are also shorter in Sinaloa than in Chihuahua, therefore the conditions for the operation of authority are different. Until last year in Sinaloa all local governments were PRIistas. Chihuahua by contrast was one of the states that first experienced political alternation, and although there have been a few PRIista governments, the mayor of Chihuahua and Ciudad Juárez have usually been on PANista hands.” Thus, militarization contributed to fragmentation in the state security apparatus in different cities, but local histories were crucial in determining how the process unfolded. In this context of fragmentation in Juarez, the methods used to expose the evidence of violence proliferated, as illustrated in Table 6.2., which contrasts the variety of methods used in 2010 with the few used in 1984. These methods were used to                                                                                                                 337 Author’s interview with high level official of the Federal Secretariat of Public Security, Mexico City, September 8, 2011.   311   intimidate criminals and state officials. In July 2010, a car bomb killed three people including a federal police officer; even though the bomb did not cause a large number of civilian victims, it was a clear message for authorities. According to the testimony of El Diego, a killer for La Linea who confessed to organizing the explosion, the attack was in retaliation to the capture of another member of the criminal group.338 This particular event illustrates why indiscriminate violence is not the same as visible violence: methods like bombs that could be quickly labeled as indiscriminate, may not be aimed at targeting a large number of victims, but at sending a message even if there are not many victims. Messages found in crime scenes usually reflected the targets of violence as well as the power disputes between the criminal organizations fighting for control over the city. On July 28, 2010, for example, two beheaded bodies appeared along with very explicit notes: “I'm a kidnapper and extortionist. I'm an Azteca" and "I do carjacking and work for La Linea and the Aztecas” [armed branches of the Juarez Cartel]. In another case, in August 2010 the body of a federal policeman was found scattered in pieces behind a mall, and in a sign painted on a wall the group La Linea assumed responsibility for the attack.339                                                                                                                 338 Video Source. Complete Interview with El Diego available at http://www.youtube.com/watch?v=Wd_3cpZgCDk [Accessed November 15, 2012] 339 El Diario. 2010. “Dejan 2 decapitados en Hacienda de las Torres.” El Diario, July 28. Author’s dataset on drug related violence   312   Table 6.2. Methods used in violent attacks, Ciudad Juárez 1984 % 2010 % Banner (without victims) Car bombs and explosions Combat with fire arm or explosives 8 8.05 22 3.51 Corpse with a note 4 0.66 Corpse wrapped in blanket 7 1.15 Fire 2 0.33 Head in cooler/frozen corpse 9 1.47 Levantón o paseo (forced disappearance) 19 3.11 Mutilation or incineration with note 12 1.97 Mutilation or incineration without note 9 1.48 Sexual violence 1 1.03 Sicariato (Drive by shootings) 8 8.25 160 26.23 Simple use of fire arm 47 48.45 315 51.64 Simple use of knives 25 24.77 15 2.46 Strangulation 1 1.03 Torture 7 7.22 36 5.9 Total number of events 97 610 Source: Author’s dataset on drug related violence, information extracted from El Diario de Juarez The profiles of victims of violence became more visible as journalists of the local newspaper El Diario and five human rights activists340 were assassinated. In September 2010 after the assassination of the photographer of El Diario, and the release of a “manta” (banner) where traffickers of La Linea referred to the assassination of the photographer, El Diario published a highly controversial editorial asking “members of the different organizations who are fighting for turf in Juarez […] as information workers we want you to explain to us what is it that you want from us, what is it that you expect us to publish or to refrain from publishing, so we know what we can expect.”341 The editorial reflected how in the process of state fragmentation and criminal competition violence reproduced easily, and the terms of interaction that seemed predictable even for civilians, became                                                                                                                 340 El Universal. 2011. “Balean a activista en Ciudad Juárez.” El Universal, December 2. 341 El Diario. 2010. “Qué quieren de nosotros.” El Diario, September 19.   313   unpredictable and quickly changing. This complexity and the lack of proper investigation in homicide cases, made even more difficult to determine whether traffickers, state forces, or common criminals were responsible for a particular attack. The difficulty to identify victims and perpetrators reflects how a definition of drug violence limited only to cases where traffickers confront each other, or the state, fails to recognize the spillover effects of violence, and the ways in which civilian casualties, even when not directly the result of a trafficker eliminating a rival, are indeed connected to the behavior of criminals and state officials. Visible violence persisted even after the 2010 elections brought a new governor, Cesar Duarte, and a new mayor, Hector Murguía. The new governor seemed to be more forcefully committed to investigate crimes, and to collaborate with other authorities.342 However at the same time, the relationship between the federal government and the mayor became highly contentious. In February 2010 the National Human Rights Commission declared that the climate of violence in Juarez resulted from the lack of coordination among different enforcement agencies343 and the US consul in Ciudad Juarez also expressed concerns about the lack of trust between government levels and enforcement agencies.344 The municipal-federal confrontation heightened due to the increasing role of both the Federal Police and the municipal police under Leyzaola’s leadership. Huge criticisms and multiple complaints about human rights violations by the Army led the federal                                                                                                                 342 Different interviews conducted in Ciudad Juárez by the author. March 2012. 343 La Jornada. 2011. “Violencia en Juarez por falta de coordinación: Comisión Nacional de Derechos Humanos.” La Jornada, February 21. 344 La Jornada. 2010. “En Juarez ejército se dedico a ver pugna entre carteles: consulado EU.” La Jornada, March.   314   government to replace soldiers with troops of the federal police. After April 2010 the Army’s patrolling of the city significantly decreased, and the federal police assumed incremental control over security. The federal police not only carried out antinarcotics raids but even assumed traditional roles of municipal police forces, such as transit control, thus generating a new pattern of confrontation with municipal forces. The Secretary of Public Security Lieutenant Colonel Julián Leyzaola, described the tension between the federal and municipal police as follows There is no coordination in the operational part. I do have communication with the Commissioner of the Federal [Police] but that’s it. For example, one day, here at the Chihuahua District we detained 10 people and seized arms, the Federals were there and seized two arms. When the time to imprison the detainees arrived I told the Federal, let’s make a joint imprisonment, and the Commander said, that is not possible, the central command ordered me to send everything to them, and I said “No, you don’t take my stuff, he is your Commissioner, not mine.” Before I arrived the municipal police was subordinated to the federals.345 The municipal-federal confrontation was extremely public as the tension involved mayor Murguia himself. On May 5 2011, in front of cameras, Murguia engaged in a vivid discussion with Federal Police who held his bodyguards at gunpoint as they crossed a Federal Police’s security checkpoint. During the discussion Murguia yelled that such kind of behavior explained why the local population did not like the federal police. The incident was the third episode that directly involved Murguia and the Federal Police.346 The confrontations reflected the excesses committed by the federal police but also                                                                                                                 345 Author’s interview with Liutenant Colonel Julian Leyzaola, Secretary of Public Security, Ciudad Juarez, July 20, 2011. 346 In January a Federal cop killed a member of Murguia’s personal security, and later in the year his bodyguards were also held at gunpoint as the Mayor’s convoy patrolled the city. See: Borunda, D. 2011. “Juárez mayor confronts feds: 'You don't rule city.'” El Paso Times, May 5. El Diario. 2011. “Pos sí, pero no tiré’: federal que encañonó a los escoltas de 'Teto.” El Diario de Juarez, May 5.   315   Murguia’s influence on public opinion, which derived from the pattern of mayor’s influence described earlier. In sum, the rapid increase in the frequency of violence in Ciudad Juarez in 2008 resulted from heightened competition in the criminal market. In turn, the fragmentation in the state security apparatus explains why violence became more visible. Over time a fragmented state security apparatus also became a source reproducing the frequency of violence as multiple enforcement agencies confronted each other, confronted different trafficking organizations, and abused force against civilians in the name of confronting traffickers. Another factor yet contributed to make the spike of violence even more extreme in Ciudad Juarez, and it was the outsourcing of violence to youth gangs, as each warring trafficking faction garnered force to dominate the city. 6.2.3.2. Cholos, crews, and barrios in Ciudad Juárez In 2008 when the Sinaloa DTO entered Ciudad Juarez, the turf war required extreme firepower if territorial control was to be achieved, and outsourcing was necessary because the organization did not have strong organizational roots in the city. On its part, the Juarez DTO, despite its historical strength in the city and despite having La Linea, began outsourcing violence to youth gangs in order to match and overpower its competitor. The Sinaloa DTO employed large gangs (crews) known as Los Mexicles and Los Artistas Asesinos, brought killers from other areas of the country, and supported splinter groups from the Juarez DTO such as Gente Nueva. For its part, the Juarez DTO employed the already existing crew Los Aztecas and Barrio Azteca, through the   316   intermediation of its enforcement branch La Linea.347 These large gangs had, according to some calculations, at least 3,000 members (Castillo 2009, 309). The large gangs recruited kids from the smaller, usually non-criminal gangs or barrios. Below the large gangs the Police calculated the existence of at least 460 gangs in the city. It would be difficult to claim that all these smaller gangs were in fact working for traffickers, but without a doubt their dynamics and incentives changed radically as traffickers started to recruit them. In November 2010, a state attorney from Chihuahua declared that there were at least 250 gangs dedicated to kidnapping and extortion in the border city348 and even though not all of them were related to drug trafficking organizations, they were using copycat techniques. The control on neighborhood gangs allowed crews and their respective employers to access micro criminal markets such as drug distribution, but most importantly, it provided armed muscle and more tangible territorial control, crucial to displace rivals out of the city. A community organizer (former gang member) considered that “the narco has always been present, but now there is more need to have soldiers; before anybody could make a business out of selling drugs, now they have to sell for a specific group.”349 This statement casts doubt on the hypothesis, advanced especially by Mexican authorities, that violence has increased as criminals have diversified their portfolios into local drug dealing. If anything, the testimonies of gang members suggest that local drug dealing has existed for years in Ciudad Juárez, but only very recently criminals attempted to establish control over these localized markets.                                                                                                                 347 The United States District Court For The Western District Of Texas El Paso Division. Indictment of Joaquín Guzmán Loera et al, Criminal No. EP-12-CR. April 11, 2012. 348 Excelsior. 2010. “Operan en Ciudad Juárez 250 bandas dedicadas al secuestro.” Excelsior November 16. 349 Author’s interview with community organizer and former gang member, Ciudad Juárez, July 19 2011.   317   The active recruiting of youth gangs radically changed the interaction between usually non-violent gangs. On the one hand, there was a pacifying effect as traffickers asked gang members not to “heat up” the plaza and thus to quiet down their small territorial disputes for the control of a few blocks. On the other hand, confrontations between gangs that belonged to different crews and different sides of the trafficking dispute got bloodier. In the words of a community organizer who worked with gangs in Juarez The “kids” [chavos] know if they are “doblados” [from the AA or Artistas Asesinos] or if they are “buddies” [from La Linea]. Violence even receded when the narco arrived, now they tell them “be careful about kicking up a fuss”, now they do fight for being part of one or the other crew, but they do not fight between them. Formerly, a barrio controlled a territory with its identity, conflicts emerged when boundaries were crossed. Formerly they fought with stones, the issue is when they start to articulate with the narco, there are no longer stones because it is no longer about identity but about defending turf, if they belong to similar godfathers violence is reduced. 350 Former gang members and community organizers echoed similar sentiments: Drug trafficking has changed violence a lot. Before violence was a personal matter, now it is a business matter. [...] Violence of the barrio does not exist anymore, because when there are points [of drug sales] the cholos cannot do things they did before, such as painting graffitis [narcos tell them, you cannot heat up the damn point]. In former years the deaths of the barrios were two or three killed in two years, now they kill three or four in two, three hours.351 Violence spiked because gangs became better-armed and more willing to fight violently. The testimonies also reveal that traffickers attempted both to discipline and to make youth gangs more violent. Yet, the ability of traffickers to discipline members in their areas of control was sometimes limited. A local drug dealer interviewed by                                                                                                                 350 Author’s interview with worker from youth NGO, Ciudad Juárez, 2011. 351 Focus group conducted by Arturo Alvarado, Director of the project Juvenile Violence in Latin America, in Ciudad Juárez, March 2012.   318   Campbell (2009) described how the violence he dealt with derived from micro conflicts of territoriality between street gangs The Juarez downtown area is divided into zones, territories or areas in which different groups or godfathers are all -or mostly all- members of the Aztecas prison gang, there´s no problem when these zones are invaded or when they overlap. The problems are all with -and between- the smaller, zealously territorial street gangs. Close to my zone of operation there are two of these street gangs: Los Jodidos (the fucked up ones) and the Teipiados (the Taped ones, a reference to drug murder victims who are often wrapped in tape by their killers]. (Jorge, drug dealer in Juarez, interview conducted by Campbell 2009, p. 106) The crucial point is that the competition between larger criminal organizations and the subsequent outsourcing of violence to youth gangs changed radically the behavior and organization of street gangs. The relationship between drug cartels and street gangs was hierarchical and tiered, but such hierarchy did not facilitate the disciplining or controlling of the smaller gangs because connections were loose and gangs became professionalized in the use of violence. The same gang member interviewed by Campbell described such hierarchy as one in which There is a tiered, hierarchical society, where every godfather has a protector, until you reach the very top of the hierarchy, where there is a guy, the leader who is currently in jail and will probably remain there for a very long time, if not for the rest of his life. I myself have started or initiated several individuals into the world of drug selling primarily youngsters. (Campbell 2009, p. 105) The largest gangs constituted the brokers and the link between trafficking bosses and the gang bases352 but there was no direct connection between traffickers and armed force. Thus, outsourcing also made it more difficult for authorities to disintegrate trafficking organizations. According to Colonel Leyzaola, his work in Ciudad Juarez was more difficult than in Tijuana because “in Tijuana from one detention we were able to get                                                                                                                 352 Author’s interview with human rights defender, Ciudad Juárez, May 17 2011.   319   five more. Here [in Juarez] from one detention we only get one detainee, because they don’t know each other.”353 The large-scale involvement of youth gangs generated instances of collective violence with three or more victims because criminal soldiers tried to eliminate cells of opponents. Outsourcing also generated identification problems as defined by Kalyvas (2006): as more people became involved in violence it became difficult for warring factions to identify their opponents. Thus, massacres were not a reflection of indiscriminate or random attacks, but rather the result of misidentification or mistaken efforts, and sometimes poorly prepared operations to eliminate rivals. Ciudad Juárez experienced a larger number of collective murders and massacres than its country counterparts. In 2010 while Culiacan and Tijuana witnessed only 19 events with 3 or more victims, Juárez witnessed 103 violent attacks with three or more deadly victims.354 Some of these events had a crucial impact in the government’s response towards violence, as was the case with the “Villas del Salvarcar” massacre that occurred in Ciudad Juárez on January 2010, which forced the government of President Felipe Calderon to mobilize a vast social program into the city, known as “Todos Somos Juárez.” In this massacre, eighteen young students were assassinated when they attended a party in the Villas de Salvarcar neighborhood. President Calderon’s first reaction was to affirm that the victims were members of a “criminal gang”, but as criticism to his quick labeling mounted, he retracted. In July 2010 the confession of criminal El Diego, suggested that students were killed in a case of mistaken identity, as members of the                                                                                                                 353 Author’s interview with Lieutenant Colonel Julian Leyzaola, Ciudad Juarez, July 20, 2011. 354 Based on analysis of the author’s dataset on drug related violence   320   Azteca gang thought members of the rival Artistas Asesinos gang were at the party.355 Other examples of such acts of collective violence included the Colonia Horizontes massacre (with 14 dead and 15 wounded) on October 2010, and massacres within rehabilitation centers for drug addicts. El Diego also claimed that he ordered massacres in rehabilitation centers he thought to be safe houses356 for Sinaloa rivals. The control and provision of arms and sophisticated “jobs” for street gangs created large opportunities for reproducing violence, because poor kids with little professional or educational opportunities were attracted by the money provided by criminal organizations, or simply because they could not risk the violent retaliation of traffickers. As another gang member expressed when referring to local drug dealing “Now the Aztecas are in control. Before, everybody worked independently, now if you don’t work for them they don’t let you work.”357 The outsourcing of violence spread out the abilities to use violence among a larger pool of people. Before the explosion of the trafficking disputes, kids were recognized within their barrio for their particular skills, and only a handful were recognized as the brave or violent ones, but this situation changed with outsourcing. In the words of the community organizer This is how I see things. In the barrios there has been a similar transformation to the one that occurred in the maquila. In the maquila there is a universal operator, who can operate any machine. There is something like that in the barrio. Before there were specialized jobs, before you were the dealer and you did not have anything else to do, or you were the post [watchman], or you were the raper, or you were the assassin and you went and fuck somebody and that was it, you didn’t have to do anything else. If you were an observer, you only had to be wachando                                                                                                                 355 In November 2010, Arturo Gallegos Castrellon, a member of the Azteca Gang confessed that he ordered the massacre because Aztecas believed members of a rival gang were at the party. El Paso Times. 2010. “Expert: Azteca leader's arrest won't end violence”. El Paso Times, December 1. 356 A “safe house” is a term used in Mexico to refer to places used by traffickers to hide arms, drugs, victims of kidnapping, and to conduct tortures and even hide corpses. 357 Author’s interview with former gang member, Ciudad Juárez, July 19, 2011.   321   [looking], the falcons only watched. Each one had a role. As the situation started to get burning, or the risks became greater, the universal narquito was born, the one that is a falcon and at the same time is fucking you, it stopped being fordist to become toyotist. And for the sake of adapting to a market where there are more risks. Before, you entered the barrio, you worked and you left. Now it is not so easy to leave because while being there you get to know the structure.358   The articulation between gangs and traffickers that fully developed in 2008 was unprecedented but not unpredictable. The outbreak of the confrontations between the Juarez and Sinaloa DTOs and the fragmentation of protection networks increased the demand for outsourced armed coercion. Such demand met the large supply of gangs and marginalized youth that existed in the city. The supply of gangs had been nurtured by the binational character of gangs such as Los Aztecas, which extended the source of supply of guns, soldiers, and money to El Paso. Additionally the extremely poor conditions within the Juarez prison system facilitated the reproduction of the large crews within prisons, and massive deportations created an additional supply of poor youth without social, economic, or family roots (OAS 2010). Even if not all members of gangs and barrios engaged with crews, their modes of interaction changed radically after 2008. 6.2.3.3. Return to normality? By the end of 2011 homicide rates had significantly decreased in Ciudad Juárez. Although the reasons for this pacification were not completely clear at the time of writing, in line with the political economy framework for explaining violence presented in this dissertation, they seemed to reflect a reduction in the fragmentation of the security apparatus, and a reduction of competition in the criminal market. On the one hand, the                                                                                                                 358 Author’s interview with worker from youth NGO, Ciudad Juárez July 18 2011.   322   security apparatus became more cohesive as federal forces reduced their presence since October 2011, and the collaboration between state authorities improved. Many inhabitants of Ciudad Juárez coincided in saying that extortion decreased as the federal police left the city.359 At the same time, confrontations in the drug market receded as the Sinaloa DTO gained control,360 and allegedly established a pact with the weakened Juarez DTO and La Linea. Other factors could have played a role in the reduction such as the Federal Police’s strategy of targeting mid-level operational members of criminal organizations, and the security strategies of Julián Leyzaola, inspired by the “broken windows” theory, including an iron fist approach and the concentration of forces in certain areas of the city, attacking even the littlest crimes. The changing situation in Ciudad Juárez after experiencing extreme violence and suffering, without a doubt brought relief to the population. However, in the long run the sustainability of the process could be endangered for various reasons. First, if criminal re- accommodations indeed played a role in reducing violence, the reduction can be unstable and vulnerable. Second, the indication that violence decreased when Federal forces left the city casts serious doubts about the behavior of these actors in enforcing the rule of law. Third, even though the municipal police is empowered and has taken the lead in public security in a way that did not occur before, its tactics seem to be extremely militarized and human rights complaints about its actions and its constant targeting of marginalized populations have increased. Marginalized youth report that the municipal police forces have stigmatized them even more than federal forces. As in Tijuana, reasons                                                                                                                 359 Interviews and focus groups conducted by the author, Ursula Alanís, and Arturo Alvarado in the context of project on juvenile violence in Latin America, directed by Arturo Alvarado at the Colegio de Mexico. 360 Author’s interviews in Ciudad Juárez and El Paso.   323   for improvement seem to do less with improvements in socio-economic conditions, or the elimination of trafficking activities, and more with changes in the terms of interaction between states and criminal actors.   6.3. Conclusion The realities of drug trafficking and violence in Ciudad Juárez and Tijuana are interconnected. Both cities have been affected by national events such as the institutional transformations in federal enforcement which started in the mid 1990s, the declaration of war on trafficking organizations by President Felipe Calderon in 2006, or the interconnection of trafficker’s strategies such as the attempt of the Sinaloa trafficking organization to invade northern territories in the late 2000s. Furthermore, similar strategies have characterized traffickers’ actions, such as the use of disappearances in the late 1990s. However, the timing and evolution of violence has differed in the two cities, affected by the micropolitics of criminal competition, democratic alternation, and institutional transformation. The contrast was particularly notable after the federal’s government decision to militarize the war on drugs in 2006. A fragmented security apparatus and competitive criminal markets characterized the two cities at the onset of military deployments, yet fragmentation and competition were even more pronounced in Juárez thus creating a more extreme scenario of violence. The trajectories of violence in the two cities show that even though their status as border cities has made them prime locations for drug trafficking, the border does not constitute an inevitable marker for violence.   324   The comparison between Ciudad Juarez and Tijuana also highlights the importance of armed coercion in the dynamics of violence. In Tijuana, traffickers always preferred to use alternate sources of force such as elite young kids, thus preventing the reproduction of youth gangs and their connections with trafficking groups. By contrast, in Ciudad Juárez, traffickers did not prevent the reproduction of gangs, and in the late 2000s actively outsourced armed coercion to multiple youth gangs, explaining the extreme spike of violence. The diverse interactions between state security apparatus, the structure of criminal markets, and the type of armed coercion deployed by criminals, explain the diverse, yet similar, trajectories of violence in the two cities, summarized in Table 6.3. Two aspects stand out in Table 6.3. First, the mechanism that sustained low visibility situations in Ciudad Juárez and Tijuana was that of state protection rather than effective state enforcement. In other words, cohesive state apparatuses provided credible state sponsored protection (Snyder and Duran-Martinez 2009a) that helped regulate the terms of interaction and the violent behavior of criminals. State authorities had the power to enforce the law more effectively if necessary (as occurred after the assassination of Cardinal Posadas Ocampo) but only did it selectively. This highlights that a state’s power to enforce the law does not equate the effective enforcement of the law because of the potential tradeoff between state capacity and state autonomy discussed in Chapter 3. The second aspect is that high visibility-low frequency situations can emerge sporadically when enforcement changes such as rotations and purges effectively disrupt the interaction between state and criminals. Whether or not these sporadic situations move back into peace or evolve into more violence depends on whether other state structures change   325   enough to make the state more fragmented, and on whether criminal organizations have enough power to control the frequency of violence.   As has been widely recognized in the literature on democratic transition in Mexico, the political dynamics of the northern states of Baja California and Chihuahua were crucial in setting the stage for the transition of power in the presidency in the 2000s. Yet, as illustrated in this chapter, early electoral transition in these states had a rather limited effect on the interactions between states and criminals. Although important, formal electoral dynamics are not the only key to explain fragmentation or cohesion in the security apparatus. Electoral alternation may not necessarily create rotation of powerful and corrupt elites, and electoral coincidence may not necessarily guarantee coordination among levels of government. For example, while in Tijuana in the late 2000s the coincidence of PAN affiliations in Presidency, Governorship, and Mayor facilitated coordination, in Ciudad Juárez the most notable confrontations at the onset of the explosion of violence occurred between a Mayor and a Governor that shared PRI affiliation. Thus, any assessment of cohesion in security apparatuses requires not only an analysis of formal dynamics of political competition, but also of informal dynamics of conflict and cooperation, and a careful consideration of the power relations between law enforcement agencies.   326   Table 6.3. A political economy approach to the evolution of drug violence in Ciudad Juarez and Tijuana Period/City Criminal Market State Security Apparatus Type of armed coercion: Relation criminal actors- gangs 1984-1996 Quasi Monopoly: Cohesion: Insourcing: Tijuana Arellano Felix Coordination among Narcojuniors (elite LF-LV Organization consolidates enforcement actors through young kids) work for the power, Attorney’s Office. Arellano Felix DTO. No conflicts between national HF-LV after 1987 some struggles and local government. for control with other Increased political competition regional organizations does not transform radically power structure. 1996-1997 Quasi Monopoly: Increased fragmentation: Insourcing: Tijuana Arellano Felix Enforcement operations of the Narcojuniors LF-HV Organization remains in Attorney’s Office and rotations control and purges 1998-2007 Increased competition: Cohesion: Insourcing: Tijuana Arellano Felix Coordination between state and Incidence of HF-LV Organization in federal actors and enforcement narcojuniors seems to competition with Sinaloa agencies recede but no systematic traffickers use of gangs 2008-2009 Competitive: Fragmentation: Insourcing: HF-HV Arellano Felix Military operations create No systematic use of Organization splits, conflicts between enforcement youth gangs splinter faction allies with actors Sinaloa DTO 1984-1993 Quasi Monopoly: Cohesion: Insourcing: Ciudad Juarez Amado Carrillo and Juarez Coordination among Availability of youth LF-LV DTO in control enforcement actors gangs but no demand of armed coercion 1994-2007 Increased competition: Cohesion: Insourcing: Ciudad Juarez Within factions of the Coordination among No systematic HF-LV Juarez DTO enforcement actors and deployment of youth government levels, political gangs competition grows but power remains stable Increased fragmentation due to 2005-2006 rotation and purges in municipal and state police 2008-2010 Competition: Fragmentation: Outsourcing: Ciudad Juarez Sinaloa DTO invades Military deployments create Juarez and Sinaloa DTO HF-HV Juarez, and confronts conflicts between enforcement deploy youth gangs in Juarez DTO for control actors the city to wage war. Confrontations between mayor Sinaloa deploys Artistas and governor, and then Asesinos, Mexicles, and between mayor and federal Gente Nueva, Juarez police DTO deploys Aztecas   327   The relatively quick control of violence that occurred in Ciudad Juarez and Tijuana after the explosion of violence in 2008, compared for example with the endemic violence that took years to recede in Medellín, opens up a wide range of questions about the success of enforcement operations and the conditions under which drug related violence can be controlled. First of all, it suggests that in situations of protracted violence, relatively quick changes can be attainable if the state can recompose its forces and criminal organizations adapt to less violent strategies. Although the reduction of violence in the two cities is very positive, the fact that the recomposition of state forces can be an ad-hoc process (facilitated by military leadership in the case of Tijuana for example) suggests that any long-term transformation requires the creation of mechanisms and checks and balances that can institutionalize cooperation, and at the same time prevent the creation of protection rackets. Second, since the reduction of violence can be at least partially attributed to a recomposition, and even ad-hoc pacts between trafficking groups, it is crucial to recognize that these pacts can be unstable, and thus, the situation of these cities needs to be constantly monitored. The cases of disappearance in Ciudad Juárez and Tijuana since the 1990s reveal crucial aspects of the relation between visible and hidden forms of violence. In the 1990s disappearances proliferated but a mix between fear, corruption, stereotypes, and rumors about the involvement of victims in criminal activity opened the door for growing impunity and apathy in cases of disappearance. Families feared the retaliation or the re- victimization of their loved ones, as they knew that in many cases they were, in fact, related to drug traffickers. Rumors about the fate of victims abounded, for example that they had struck a deal with Mexican or US authorities, and were thus simply hiding. This   328   situation delegitimized the voice of the few relatives who dared to push for government’s answers. When violence exploded in 2008 disappearances multiplied, thus suggesting that in cases of extreme violence, all forms of violence, including less visible ones, mingle and intermix. But as violence became more visible and widespread among civilians considered to be innocent, i.e. not related to traffickers, the stigma on victims of violence decreased, and thus the efforts of families to find their loved ones became more legitimate and public, even if still doomed by impunity. People feared less the stigmatization, and thus the method of disappearing people has become, paradoxically, more visible. But as disappearances of the nineties show, disappearing can be a powerful and effective way to hide the perpetration of violence, especially when other forms of violence are relatively uncommon and violence is perceived to be predictable and limited to those “who deserve it”. Finally, Ciudad Juarez and Tijuana, as the other cities analyzed in this dissertation, illustrate that in many cases the lines between state and criminals can become really blurred. Cops and politicians sometimes not only allow or facilitate drug trafficking but become direct perpetrators of violence and criminality: for example, during the 1990s enforcement agents were often directly involved in disappearing people, and sometimes used the façade of conducting a legal detention. In this context, separating a state from a criminal actor may seem meaningless. Nonetheless, as the political economy framework for explaining drug violence highlights, even when the boundaries between state and criminals seem blurry, it is possible to understand systematically the varied interactions that can exist between state security apparatuses and criminal actors,   329   and thus disentangle the varied impacts that such interactions can have on political and social order, and more specifically, on the patterns of drug violence that are the focus of this dissertation. The involvement of state forces in violence in criminality also highlights the importance of zooming into the role of the state in perpetrating violence; for the most part this dissertation has focused on the violence perpetrated by criminals, but it is clear that specially in moments of extreme violence the state acts both as an enforcer and as a perpetrator of violence. State actors, just as criminals, can also strategically use visible and hidden forms of violence.   330   CONCLUSION Violence is not an inherent attribute of the drug trade, but rather the result of complex interactions between states and criminal actors in illegal markets. In this dissertation I have shown how varied types of drug violence result from interactions between the state security apparatus and the structure of competition in the illegal market, and are also affected by criminal’s strategies of armed coercion. In this sense, my argument challenges ideas that see violence in drug markets as an automatic result of the illegality of the market, or as solely dependant on the economic aspects of the trade such as the size and profitability of markets. It also challenges arguments that explain violence as resulting only from government enforcement actions. My argument is relational in nature, and explains violence as the joint result of dynamics of power within the state, and dynamics of power within the market. Thus, this argument also challenges the idea that crime and illegality simply reflect an absence of the state, and highlights the importance of unpacking power relations within the state in order to understand dynamics of illegal markets. One essential aspect in understanding drug violence is considering the wide range of variation in lethal forms of violence that can emerge within the same illegal market. In this study I have advanced a multidimensional understanding of violence that considers not only the frequency of violence, but also the visibility, that is, the instances where criminals choose to expose the evidence, or claim responsibility for the attacks they carry out. By including and assessing visibility in a systematic way, I have been able to   331   uncover patterns of violence that would not be apparent if one simply looked at the frequency of violence. Most importantly, by analyzing the visibility of violence, I have been able to uncover profound puzzles in criminal behavior, and to show that criminals not only decide whether they should, or should not, use violence. They also decide on whether they should, or should not, expose such violence beyond the direct targets of violent action. The multidimensional approach also provides a stronger theory about violence in drug markets because it allowed me to analyze not only extreme cases of violence, which tend to gain the most attention from scholars and media, but also the non-visible, and less frequent forms of violence. By assessing two interrelated but opposite puzzles –if violence is unavoidable to regulate transactions in illegal markets, how can criminals control its use, and if visible violence can attract attention that is detrimental for the business, why do criminals decide to expose violence- I found that the structure of the state security apparatus affects the visibility of violence. Criminals are more likely to hide violence when they receive credible protection from the state, or when they fear the retaliation of an effective state, and this occurs when the security apparatus is cohesive. By contrast, when criminals do not receive predictable protection, or do not fear an effective enforcement of the law, they lose incentives to avoid state attention and may decide to expose violence, or to abandon the extra steps necessary to hide it. This occurs when power within the state is fragmented. While the structure of the security apparatus affects the visibility of violence, competition within the illegal market determines the frequency of violence, and thus, competitive markets are more likely to generate frequent violence, than monopolistic   332   markets. While certain competitive markets can sometimes be peaceful, this is usually a result of pacts or different arrangements of collaboration between illegal actors, which tend to be unstable and not inherent to a competitive market. The framework I developed also shows that extreme spikes of violence occur when criminals outsource their use of violence to youth gangs. Drug violence has become an extreme concern over the last decade for countries around the world, especially in Latin America, and thus, the framework I develop can help understanding very complex situations such as the increase in criminal violence in Central America since the early 2000s. Many observers have been quick to attribute violence in Central America to an expansion of criminal actors from Mexico, and to an increase in drug trafficking. Yet, several elements suggest that the situation is more complex, precedes the wave of violence in Mexico, and than in fact is closely tied to domestic political conditions.361 First, homicide rates were on the rise in countries like Guatemala way before violence skyrocketed in Mexico, and drug trafficking networks have had a long tradition in the region. Thus, it is possible that the increasing perception of a more extensive violence problem may well be related to increases in the visibility of violence, and thus to changes in state power in the country. Second, increases in violence have not been homogenous within the region; while Honduras and Guatemala have seen their levels of violence rise, historically violent countries like El Salvador have seen a recent decline in violence, without any drastic change in the criminal market. A more                                                                                                                 361 For a good discussion of violent trends in Central America See: Cruz, J.M., Fernández de Castro, R and Santamaría, G. 2012 “Political transition, social violence, and gangs: Cases in Central America and Mexico,” in C.J. Arnson (Ed.) In the wake of war: Democratization and internal armed conflict in Latin America. Stanford, CA: Woodrow Wilson Center Press and Stanford University Press.   333   detailed analysis of the structure of power in the security apparatus and in the illegal market of these countries thus provides avenues to understand the nuances of these cases. For the most part I emphasize how criminals decide to employ visible violence or to hide it responding to the incentives generated within the state. But criminals may also manipulate their use of violence strategically to elicit responses from state officials. In fact, most of the times the two things may be occurring at the same time, criminals are responding to incentives in the state, and manipulating state behavior, as when they decide to become “traffickers in silence” to minimize state attention. For example in Medellin, “Don Berna” responded to the incentives generated within the state through the peace process, but also knew that manipulating the visibility of violence was necessary to continue extracting judicial benefits from the state. It is possible that this situation has been occurring in El Salvador since March 2012, when the two major gangs in the country, Mara Salvatrucha and Barrio 18, struck a truce that has led to a reduction in homicide rates.362 Although there seems to be a real decrease in the frequency of violence, it is likely that significant violence that still occurs is now hidden, because maintaining the appearance of peace is essential for the gangs to try to extract judicial concessions from the state.363 In the remainder of this conclusion I examine some theoretical implications of the argument beyond the realm of drug violence, then I explore some directions for future research, and finally conclude by briefly reflecting on the policy implications for policies                                                                                                                 362 Dudley, S. 2013. 5 Things the El Salvador gang truce has taught us. Insight Crime. March 12. Available at http://www.insightcrime.org/news-analysis/5-things-el-salvador-gang-truce-taught-us, Accessed [6 April 2013]. 363 I owe this observation to Jose Miguel Cruz.   334   aimed at addressing violence and drug trafficking, and for discussions about the need to reform the current drug prohibition regime. The centrality of the state in the analysis of violence One crucial finding of this study is that to understand drug violence it is necessary to unpack the power dynamics within the state. In other words, the state should not be conceptualized as one unitary actor that simply opposes or struggles against criminal actors, but also as an actor that sometimes collaborates with them. Furthermore, power struggles within the state shape the ways in which “the state” confronts or collaborates with criminal and non-state actors. While differentiating between cohesive and fragmented states, I showed that the actions of the state, either to confront or to protect criminals, vary depending on power struggles within the security apparatus. This emphasis on the centrality of power within the state not only contributes to understanding criminal violence, but can also expand the already vibrant and ever- growing research on political violence and civil wars. Of course many interesting debates derive from discussing how criminal violence is different, and sometimes even more lethal, than traditional forms of civil war and political violence (Lessing 2012a, Osorio 2013), and therefore, from discussing how distinct theoretical frameworks are required to understand criminal violence. However, there are also crucial overlaps and conceptual advantages in bridging the analysis of civil wars and the analysis of criminal violence. For the most part, research on civil wars has analyzed states as unitary actors that oppose and confront armed opponents. Understanding how different actors within the state   335   behave, collude, and bargain security policies, can also greatly expand the way in which we understand conflict dynamics. A greater focus on the state and its relations to criminals also reminds us that the limits of state power are often times blurred. For example, how should we conceptualize a policeman that not only protects, but also directly carries out illegal actions for a trafficking group? Is he a criminal or a state actor? Can we predict when and how the limits between the state and criminal actors can become more blurred? Here, the long time tradition of analysis about ideas of state autonomy and their impact on development outcomes (Evans 1995) can prove very useful if extended to the analysis of violence, law enforcement, and stability. Some ideas about the potential tradeoffs between state autonomy and state capacity were explored in Chapter 3, and can be further extended. Institutional change and violence In chapters 1 and 3, I discussed how processes of institutional change such as democratization, decentralization, and anticorruption reforms, can affect the cohesion in the state security apparatus. Consequently, these processes can affect the interactions between states and criminals, and shape the incentives for criminals to use visible violence. My argument thus connects to long-standing debates about how institutional change comes about. Traditional approaches to institutions emphasize the persistence of institutions and the difficulty of changing them in the short run (North 1990; Pierson 2001). More recent approaches have emphasized how institutional change can occur in the margins of state institutions, and can be cumulative (Streeck and Thelen 2006). Following the essence of this latter conception of institutional change, my analysis of   336   state security apparatuses shows that institutions can indeed change significantly in the short term. Some changes can be cumulative, as occurred in Mexico since the mid 1990s when increasing levels of political competition and reforms in security agencies started to transform power relations within the state security apparatus. However, there can also be abrupt changes, such as those institutional changes occurring during the Presidency of Felipe Calderón in Mexico, which accelerated the slow cumulative changes that had been occurring over the past two decades. My case studies show that institutional transformation can also be non-linear and regressive. In the case of Medellin we saw how despite positive evolutions in the state apparatus of the city, such as greater coordination between enforcement agencies, the emergence of independent political actors and their ability to advance progressive social reform agendas (Moncada 2011), the city returned to a situation of violence after a period of apparent pacification. Of course, the violent situation after 2008 is not nearly as bad as it was in the 1980s and suggests a positive path of evolution, but the point is that the outcomes derived from institutions –in this case those institutions associated with the provision of security- are not necessarily linear, cumulative, and progressive. An analysis of the state security apparatus at the local level, such as the one I carried out, also highlights the importance of thinking about institutions and regimes at the subnational level. There is an already established, but also increasingly expanding, literature analyzing regimes at the subnational scale (Moncada and Snyder 2012). My study contributes to this area of research by showing how varied institutional contexts at the local level can shape the behavior of criminal actors that in many cases transcend   337   local spaces, and connect to both national and trasnational arenas. My analysis does not isolate criminal actors from the national and trasnational forces that shape them, but rather shows how these forces are reshaped and vary at the local level. Directions for future research The links between gangs and criminal actors I explored how the type of armed coercion criminals employ, and more specifically, the insourcing or outsourcing of violence to youth gangs, can determine whether violence reaches extreme levels. As illustrated in the cases of Cali, and Ciudad Juarez before 2008, outsourcing is not an automatic result of the existence of youth gangs or of marginalized young populations. Thus, for outsourcing to emerge, there needs to be a concurrence between the supply of youth gangs and the demand for armed coercion. It is crucial to specify with more detail how such concurrence takes place. Do the preferences of youth gangs about whether they want to engage in criminal activities and violence matter? Or rather, is outsourcing mostly a top-down process, determined by the preferences of higher echelons of organized crime? Analyzing this is crucial not only to better understand the ways in which young populations engage in crime and violence, but also to think about policy interventions that can prevent these connections between criminals and gangs from occurring. The limits between cohesive and fragmented states, and the paradoxes of democratization I advanced a broad definition of state security apparatus that encompasses three complex dimensions: collaboration between levels of government, collaboration between   338   enforcement agencies, and time horizons of public officials, the latter crucially determined by processes of democratization and anticorruption reforms. As such, this is a broad definition that was necessary to advance the theory on an under researched topic, but that can be disaggregated into multiple and interesting research questions, of which I want to highlight two. First, I argued that these three dimensions do not necessarily vary together, and thus, a crucial question is when does electoral competition affect the other two dimensions of the security apparatus and translates into state fragmentation? Recent research shows that electoral competition is related to growing levels of criminal violence in Mexico (Osorio 2013) and there is also vibrant literature on the potential of political competition to increase violence in civil wars (Brancati and Snyder 2011, Brass 1999, Snyder 2000). Thus one may assume that, along the lines of my argument, for the most part higher levels of competition can fragment the state, and create more violence by affecting the other two dimensions of the security apparatus. Democratization processes led to crucial reforms within the Attorney General’s Office (PGR) during the Fox and Calderón’s governments; these reforms were aimed at reducing the long-standing power of the PRI in the security apparatus in Mexico. Yet, in other cases, electoral competition may not have a fragmenting effect on the state, or may not lead to further transformations in enforcement agencies. In Mexico, rotations in enforcement agencies in the 1990s created momentary increases in state fragmentation but occurred as a result of anticorruption efforts aimed at appeasing growing suspicions of corruption, rather than as a result of democratization process. Similarly, political competition may not necessarily fragment the state if some political elites remain ingrained as occurred in the first years of   339   electoral alternation in the northern states of Mexico, or as has occurred through years of electoral competition in the city of Cali. Along the lines of the discussion presented in Chapter 3, in order to analyze the conditions under which electoral competition affects relations between levels of government, and the organization and capacities of enforcement agencies, we need to consider the possible tradeoff between the state capacity to enforce the law and the state’s autonomy from criminal influences. This in turn, can shed light on the situations in which increased electoral competition may reduce the state capacity to enforce the law and situations in which, over time, the state can regain such capacity. This also requires introducing the analysis of law enforcement more directly into studies of violence and democratization. Analyzing how and when political competition leads to more state fragmentation, and thus more visible violence, relates to a second question that is essential for further research: can fragmented states become effective in enforcing the law? And, can cohesive states become accountable and independent from criminal influence? There are cases of fragmented security apparatuses, such as that of the United States, that are very effective in enforcing the law. This can be a result of incentives that generate cooperation, and thus cohesion, between multiple layers of government and between different enforcement agencies. This possibility, however, suggests that there is room to better conceptualize enforcement structures to understand when and how they can become both more efficient, and more accountable and independent from criminal influences.   340   The state as a performer of visible violence I have focused on analyzing when and how criminals change their violent strategies, but as became evident in many of the case studies, in its effort to confront criminal organizations, the state itself is a crucial perpetrator of both lawful and unlawful violence. As mentioned earlier, state actors can directly perpetrate violence for criminal organizations. Analyzing when and how a symbiotic relationship between states and criminals that supersedes simple collaboration or indirect tolerance of criminal activities, takes place, should be subject to further research. For example, why and how a criminal organization like La Linea that directly deployed police members, emerged in Ciudad Juárez ? Another interesting possibility for research is to apply the concept of visible violence to state actions, and analyze when, in the process of law enforcement, would state actors perform violence visible, and when would they hide it. States may need to hide violence in their interaction with criminal actors when they recur to dubious human rights practices in the process of identifying or incarcerating suspects: forced disappearance by state agents is the best illustration of state forces hiding violence. Yet, state actors can also choose to publicize their use of violence when it can bolster their popularity and legitimacy, and reinforce the idea that the state holds the monopoly of the use of force. Highly publicized military deployments to confront criminals in urban areas are perhaps the best illustration of a state carrying out visible violence to confront criminals. Such operations have not only emerged in Mexico and Colombia, but are also recurrent in places like Brazil and Central America.   341   The consequences of different forms of violence While analyzing the causes of drug violence, one inevitably becomes interested in the consequences of drug violence. In the cases I analyzed for this dissertation I found very different reactions towards criminal violence, especially in terms of organized responses through civil society organizations. Medellín and Ciudad Juárez display an extensive network of civil society organizations that advocate against violence and for justice, monitor violence trends, and implement creative solutions that attempt to address the root causes of violence and its consequences on local communities. By contrast, Cali and Culiacán display seemingly weaker civil society efforts to confront criminality, despite their long histories with drug violence. Thus, an interesting line of research derives from comparing these diverse responses to drug violence, and to criminal violence in general. Analyzing the consequences of drug violence requires bridging bodies of research that make opposing assessments about citizen’s reactions to violence. On the one hand, research on the institutional and social consequences of civil war has found that “victimized people have been found to be more politically active and to display greater trust toward citizens as compared to non-victimized people” (Balcells 2012, p. 312), and this effect can even extend to the perpetrators of violence (Blattman 2009) and to peacetime (Bateson 2012). On the other hand, research on the fear of crime shows how crime nurtures vigilantism, segregation, and privatization of security (Caldeira 2000, Malone 2010, McIllwaine and Moser 2004) and hinders collective responses to violence. The seeming variation in the ways citizens react to violence suggests that violence can   342   both hinder and promote civil society organization depending on the specific types of violence that prevail. Interesting research is already growing in terms of analyzing the diverse individual and communal responses to violence in crime-ridden areas, and points to the importance of political and social networks in shaping responses to violence. The framework I presented in this dissertation suggests that the type of violence, and especially its visibility, can also shape differential citizen responses. For instance, in Culiacán and Cali the weaker civil responses coincide with the prevalence of low visibility forms of violence. Policy implications Anticrime and antinarcotics policies Although the objective of this dissertation was not to evaluate the effectiveness of policies to prevent and confront violence and criminality, there are some clear implications that derive from the analysis. First, a decline in violence does not necessarily imply the disappearance of trafficking organizations or the elimination of the most powerful ones. Low levels of violence in criminal markets can be associated either with the ability of an organization to establish a monopoly, or with the ability of a few organizations to set up temporary non- violent arrangements. The best illustration of the first option is the large reduction of violence that occurred in Medellin between 2003 and 2007. An illustration of the second option is the reduction of violence that occurred in Tijuana between 2010 and 2011. A crucial implication of understanding that criminal groups can mitigate their use of violence is that situations of relative peace need to be monitored constantly because   343   monopolies and criminal arrangements that bring about peace are prone to instability, and thus to violence relapses. Second, the acknowledgement that violence can sometimes be non-visible reminds us that sometimes situations labeled as not so violent may actually be situations where criminals have recurred to less visible forms of violence, and thus violence oriented policies cannot stop when the most visible violence has decreased. For example, the increase in violence in 2008 that followed a historic reduction of violence in Medellin, does not cause the same level of national and international outcry that violence of the 1980s caused. Yet this violence is not less harmful for the local population. Thus, all programs aimed at tackling violence should adopt metrics that allow for a broader assessment of violence. Third, even though drug trafficking is a transnational phenomenon, the violent consequences of trafficking are most strongly felt at the local level. Furthermore, local conditions matter in determining violent outcomes, as illustrated by the fact that the same organization can exhibit quite different violent behavior in different places: the Sinaloa DTO uses a wide range of violent methods in different parts of Mexico. Thus, policies aimed at reducing violence and/or targeting certain organizations should consider the conditions under which local enforcement agencies operate. One implication of considering the structure of enforcement as consequential for violence is that some well- intended policies aimed at reducing corruption can generate violence while fragmenting the state security apparatus in the short term. This is especially the case with massive purges of security forces that not only can generate violent reactions by criminals but also can push many of the purged elements into criminal organizations. This is also the case   344   with massive military deployments, which can further fragment the security apparatus, and even generate human rights violations. This does not mean that anticorruption policies or enforcement actions should not be carried out, but rather that they should always be thoroughly planned, paying special attention to local enforcement conditions. Finally, the impact that outsourcing can have on drug violence highlights the importance of creating programs that prevent the engagement of youth, and of non- violent youth gangs, in criminal activities. It is crucial to understand that direct connections between gangs and organized criminals are not automatic, and thus, it is essential to avoid the stigmatization and criminalization of young marginalized kids, and of gangs, as violent actors. If the connection emerges, then the emphasis should be on breaking up the connection between gangs and criminals. This implies targeting the brokers that connect not necessarily violent youth gangs with criminal organizations, and controlling the mechanisms that allow the reproduction of connections between gangs and criminals, such as recruitment within prisons facilitated by lax prison controls and inappropriate prison conditions. It also implies providing marginalized youth with viable alternatives outside criminal activities. The future of drug policies and their possible impact on violence In Chapter 1, I emphasized how the complex variation in types of drug violence could not be simply explained by the illegality of the market. Local state power and power dynamics in the illegal market, social conditions that can determine the existence of gangs and the engagement of criminal actors with young populations, are essential to understand why, within the same prohibition regime, criminals can behave in so many   345   different ways. This does not mean, however, that prohibition does not contribute to create a favorable environment for the emergence of violence in a given market. The Global Drug Prohibition Regime that emerged at the beginning of the 20th century and that consolidated between the 1940s and 1970s has not succeeded in its intended outcomes of reducing the consumption and production of drugs, or eliminating criminal organizations. Drug markets keep growing and expanding geographically, and criminal organizations keep finding incentives to engage in activities that are highly profitable. In 2010, according to the UNODC, an estimated 230 million people, or 5% of the world’s population, had used illicit drugs at least once during that year (UNODC 2012b).364 The prohibition regime has also curtailed the exploration of alternative ways to deal with drug consumption because it criminalizes consumers. Criminalization has generated problems of prison overcrowding, and high costs in criminal justice systems that often deal with non-problematic consumers rather than with highly dangerous drug lords. In this context, and in the midst of a historic change in the terms of debate about drug regulation, two crucial questions emerge: is it possible to foresee a change in the current prohibition regime? if so, how can that change affect current levels of violence? While answering these questions clearly requires an effort that supersedes what I can say over the next paragraphs, and that is in fact being undertaken by some academics, and especially by many civil society groups, I just want to suggest a few thoughts. In his analysis of the Global Drug Prohibition Regime and its workings within the United Nations system, David Bewley-Taylor (2012) has described cleverly how, over                                                                                                                 364 According to UNODCCP (1999), the predecessor to the UNODC, in the 1990s an estimated 192.7 million people had used drugs within the last year, an annual prevalence of about 3.34% of the world’s population.   346   the past decade, the prohibition regime has experienced a substantial weakening derived from the “soft-defection” of countries that have implemented harm reduction approaches, and alternative regulations on cannabis. However, he also shows that paradoxically, this soft defection has contributed to maintaining the regime for example when the UNODC claims that everything done under the Drug Conventions fits the idea of harm reduction.365 In this context Bewley-Taylor argues that major overhauls of the regime do not seem plausible in the short run; considering that the regime is closely tied to the global distribution of power, change is unlikely to happen unless major powers, chiefly the United States, agree on it. Since 2012, however, a crucial change has indicated an alternative road to change, which may not occur only through powerful players in the international system. For the first time, seating Presidents, including those of Colombia, Guatemala, and Mexico, declared publicly their willingness for discussing the pertinence of current drug policies. The declarations of these Presidents, all of them from Latin America, and thus at the center of the “war on drugs”, marked a sea-level change in discussions about the need to redefine the war on drugs. For the past six decades any public figure expressing concerns about the regime was demonized, and no seating President even dared to question the need of “fighting” against drugs. The open involvement of Presidents in the discussion, followed by the legalization of marijuana in two states in the United States (Washington and Colorado in November 2012) suggests that some substantial change in the way in                                                                                                                 365 A clear example of this is the preface to the World Drug Report 2012 where the UNODC’s Director asserts: “we need to be equally clear about the importance of the international conventions on drugs, organized crime and corruption. Indeed, almost everything mentioned in this preface — focusing on drug demand, rehabilitation and reintegration, alternative development, shared responsibility, and fundamental human rights — are underscored in the conventions”.   347   which drugs are regulated is becoming politically feasible. That being said, the major consensus seems to be around the need to legalize marijuana and to depenalize the use of drugs, focusing more on harm reduction, and on a health approach to drug use. It is likely that legal changes, even if uneven, can occur over the next decade in these policy areas. However, regulatory changes on other drugs such as cocaine, heroine, and methamphetamines, seem unlikely in the near future. If legal changes regarding the possession of drugs and the status of marijuana occur, the question is then, what would be the impact on overall levels of drug related violence? The answer is not simple. First, there is considerable debate about how much income drug trafficking organizations derive from marijuana production, and thus about how much legalization could affect the income of trafficking groups.366 Regardless of the amount that criminal groups derive from marijuana, it is clear that legalization can reduce profits associated with an illegal commodity, and thus the incentives to protect turf violently. Thus, legalization could indeed reduce the income of trafficking groups, although there is uncertainly about how big the effect could be. Many of the potential effects of legalization or depenalization of drug use on the size of the market will depend on the taxing schemes in the newly legalized market. They will also depend on dynamics of policing and enforcement, because it is likely that even if the marijuana legalization trend expands more evenly across and within countries, there still will be different regulation schemes in place (for example from full legalization to                                                                                                                 366 Using very transparent methodologies, recent studies have estimated the size of income that Mexican organizations can derive from marijuana trade (Kilmer et al 2010) and the potential impact that legalization of marijuana in US states can have on the income of Mexican organizations (Hope and Clark 2012). Both studies highlight the uncertainty of the estimates, yet the first study highlights that the income may be significantly lower than assumed (between 15 and 26%), while the second suggests that with legalization some Mexican DTOs can lose up to 50% of their income.   348   depenalization only). The persistent illegality of other drugs could keep incentives for criminals to protect their turf and the high profits that can derive from it violently. And even if other drugs were legalized, it is possible that criminals can diversify portfolios into other criminal activities like extortion to compensate for their lost income. We still need a lot of research to understand when and how criminals diversify their portfolios and migrate into other illegal activities that can be violent. But even if diversification is possible, it is also plausible that many criminal organizations may not necessarily migrate into other illegal activities but into the legal sector and this could, in the long run, eliminate incentives for people to engage in criminal organizations. In any case, legalization or decriminalization of marijuana may not have an immediate impact on dismantling powerful criminal groups and on reducing violence, mainly because most of these groups combine income from different drugs. Yet, the depenalization of drug use, and the legalization of one market that has been demonized for decades, can have other consequences that indirectly can affect levels of violence in the short and medium run. First, it can reduce prisoner populations and thus reduce the pool of people that can get trapped into circles of recidivism and criminal behavior when entering prisons. Second, it can save expenses in criminal justice for non-violent offenses, and these savings can be directed into programs that prevent violence and drug use. Third, by simply changing the public discourse, legal changes on drug use and marijuana can open the space to think about alternate ways to deal with drug production and with harder and more problematic drugs. Fourth, and as the Global Commission on Drugs and Democracy has recognized, by breaking up the taboo that links drugs and drug consumers with crime, and by focusing on reducing the harms produced by the illegal   349   narcotics trade, depenalization of drug users can help in focusing enforcement on the most violent organizations that can generate more instability and insecurity. The last possibility brings me full circle to finish up by connecting these possibilities with my own analysis of violence. Changes in the global drug prohibition regime –at least those that seem plausible in the short to medium term- may not generate immediate results in dismantling criminal organizations and in eliminating violence. Yet, changing a regime that has clearly failed opens up an essential avenue for policy innovation, by treating drug trafficking, drug use, and violence as interrelated but different policy priorities. This can lead to better and more targeted enforcement, something that to some extent seems to have been successful in the UPP scheme in Rio de Janeiro, a scheme of highly targeted enforcement. The UPP, or pacification police units, implemented since 2010 in Rio de Janeiro, represents in very broad terms, a scheme of highly targeted and publicized military interventions in crime-ridden favelas (Lessing 2012b). It appears to have been effective in reducing homicide rates by combining both a highly selective enforcement approach and a combination of military and social policies. Yet, if targeted enforcement policies like the UPP are applied elsewhere, it is essential to recognize, as I have shown, that the link between drugs, criminal organizations, and violence is not automatic. Thus, relatively peaceful scenarios may sometimes be the result of highly powerful criminal organizations and of extended networks of state corruption and protection, and may hide situations of less visible violence. Also, the success of enforcement operations is highly dependent on the level of state cohesion and the efficiency of enforcement. In the case of the UPPs, state cohesion,   350   in the form of collaboration between all levels of government (city, state, and federation) seems to have been essential in the success. In other words, a selective enforcement approach is not efficient in a vacuum, but is dependent on enforcement institutions. In sum, all policy redefinitions that can become plausible in a more flexible regulatory framework for psychoactive drugs need to recognize the possible tradeoffs and unintended consequences that can emerge as enforcement is redirected towards reducing or targeting the most violent trafficking organizations. For example, such targeting can leave powerful organizations untouched, ignore certain forms of violence, or motivate criminals not to reduce violence altogether, but to make it less visible. No policy framework is free of possible tradeoffs, and this will certainly be the case with any new regulatory alternative on psychoactive drugs. It is necessary to be aware of all the possible scenarios and not to expect immediate results in the elimination of drugs, violence, or organized criminality. But it is clear that the current regime is exhausted, that the human costs are too high, and that there is an unprecedented window of political opportunity to finally rethink a failed regime.   351   BIBLIOGRAPHY Aguirre, A and Cobián, F. 1993. Quienes mataron al cardenal? Proceso No. 867, June 14. Alonso, M., and Valencia, G. 2008. Balance del proceso de desmovilización, desarme y reinserción (DDR) de los bloques Cacique Nutibara y Héroes de Granada en la ciudad de Medellín. Estudios Políticos, 33: pp. 11-34. Alonso, M., Giraldo, J and Sierra, D. 2007. Medellín: El complejo camino de la competencia armada In: Romero, M (Ed). La Parapolítica. La ruta de la expansión paramilitar y los acuerdos politicos. Bogotá: Corporación Nuevo Arcoiris, pp.109-145. Alvarado, A. 1997. Las elecciones para Gobernador de Sinaloa en 1992 In: R. Loyola Diaz (Ed.) La Disputa del Reino. Elecciones para Gobernador en Mexico en 1992. FLACSO, Mexico D.F., pp. 278-306. Amat, Y. 2012. “'Debe permitirse el consumo controlado de marihuana': general Naranjo”. El Tiempo. April 21. Interview with General Oscar Naranjo. Amnesty International. 2005. Mexico: Justice fails in Ciudad Juarez and the city of Chihuahua. February 2005. Available at: http://www.amnestyusa.org/node/55339?page=show [Accessed November 23, 2012]. Ampudia, L. 2009. Empleo y estructura económica en el contexto de la crisis en Ciudad Juárez: las amenazas de la pobreza y la violencia In: Diagnóstico sobre la realidad social, económica y cultural de los entornos locales para el diseño de intervenciones en materia de prevención y erradicación de la violencia en la región norte: el caso de Ciudad Juárez, Chihuahua, Comisión Nacional para Prevenir y Erradicar la Violencia contra las Mujeres (CONAVIM), México, Secretaria de Gobernación (SEGOB), Gobierno Federal. Andreas P. 2000. Border games: Policing the U.S.-Mexico divide. Cornell University Press. Andreas, P. 2011. Illicit globalization: Myths, misconceptions, and historical lessons. Political Science Quarterly, 126(3): pp. 1-23.   352   Andreas, P. 2013. Smuggler nation: How illicit trade made America. Oxford University Press. Andreas, P., and Greenhill, K. M (Eds.). 2010. Sex, drugs, and body counts: The politics of numbers in global crime and conflict. Ithaca: Cornell University Press. Andreas, P., and Wallman, J. 2009. Illicit markets and violence: What is the relationship? Crime, Law and Social Change, 52(3) September 2009: pp. 225-229. Angarita, P.E. et al., 2008. Dinámicas de guerra y construcción de paz: Estudio interdisciplinario del conflicto armado en la Comuna 13 de Medellín. Iner, Universidad de Medellín, Corporación Región, Instituto Popular de Capacitación.   Arenas Gomez, J.C., and Escobar, J.C., 2000. Discursos políticos y resultados electorales en Medellín durante los años noventas. Revista de Estudios Politicos 16. Arias, E.D. 2006. Drugs and democracy in Rio de Janeiro: Trafficking, social networks and public security. Chapel Hill: U. of North Carolina Press. Arjona, A. 2010. Social order in civil war. PhD Dissertation, Yale University. Arlacchi, P. 1986. Mafia business: the Mafia ethic and the spirit of capitalism. New York: Verso. Arocha, J; Cubides, F., Jimeno, M. 1998. Las Violencias: Inclusión Creciente. Bogotá: Facultad de Ciencias Humanas, Universidad Nacional, CES. Arrázola, M., 2011. “Todo empezó en Envigado” El Espectador February 12. Arzt, S. 2011. Democracia, Seguridad y Militares en México. Open Access Dissertations. Paper 544. Available at: http://scholarlyrepository.miami.edu/oa_dissertations/544 Astorga, L. 1995. Mitología del narcotraficante en México. Plaza y Valdes: Mexico, D.F. Astorga, L. 2003. Drogas sin fronteras. Grijalbo. Astorga, L. 2005. El siglo de las drogas. Mexico: Plaza & Janes. Astorga, L., and Shirk, D. 2010. Drug trafficking organizations and counter-drug strategies in the U.S.-Mexican Context. Working Paper Series on U.S.-Mexico Security Cooperation. Woodrow Wilson International Center for Scholars, Mexico Project and Transborder Institute, University of San Diego.   353   Avila, A., y Velasco, J.D. 2012. Democracias en venta: partidos, corrupción electoral, violencia y crimen en Colombia (2007-2011). Bogotá: Corporación Nuevo Arco Iris. Bagley, B. 2012. Drug trafficking and organized crime in the Americas: Major trends in the twenty first century. Woodrow Wilson International Center for Scholars, Latin American Program. Bahamon, A. 1991. Mi guerra en Medellín. Intermedio Editores. Bailey, J. 2011. “What do the Zetas and Mc Donalds have in common?” Insight Crime December 5. Available at: http://insightcrime.org/insight-latest-news/item/1933-what-do-the- zetas-and-mcdonalds-have-in-common [Accessed 1 February 2011 ] Baird, A. 2012. ¿Héroes olvidados? activismo desde la sociedad civil & políticas de juventud en Medellín In: A. Baird, y J.F. Serrano, (Eds.) Paz Paso a Paso: Una mirada desde los estudios de paz a los conflictos colombianos, Pontificia Universidad Javeriana, Bogotá. Bakke, K., Gallagher Cunningham, K., and Seymour, L. 2012. A plague of initials: fragmentation, cohesion, and infighting in civil wars. Perspectives on Politics June 2012, 10(2): pp. 265-283. Balbín, B. A., and Harris, B. 2004. Violencias y conflictos urbanos: un reto para las políticas públicas. Medellín: Instituto Popular de Capacitación IPC. Balcells, L. 2011. Continuation of politics by two means: Direct and indirect violence in civil wars. Journal of Conflict Resolution 55(3): pp. 397-422. Balcells, L. 2012. The consequences of victimization on political identities: Evidence from Spain. Politics and Society, 40(3): pp.311-347. Barbary, O. 1999. Afrocolombianos en Cali: Cuántos son, dónde viven, de dónde vienen? In: O. Barbary et al. Afrocolombianos en el area metropolitana de Cali: Estudios sociodemográficos. Documento de trabajo No. 38. Cali: Universidad del Valle, Facultad de Ciencias Sociales y Económicas. Barraza, L (Coord). 2009. Diagnóstico sobre la realidad social, económica y cultural de los entornos locales para el diseño de intervenciones en materia de prevención y erradicación de la violencia en la región norte: el caso de Ciudad Juárez, Chihuahua. Comisión Nacional para Prevenir y Erradicar la Violencia contra las   354   Mujeres (CONAVIM), México, Secretaria de Gobernación (SEGOB), Gobierno Federal. Barreto Nieto, L., and Rivera, S. Una mirada a la impunidad en el marco del sistema penal acusatorio en Colombia. Bogotá: Ministerio del Interior y Justicia. Bateson, R. 2012. Crime victimization and political participation. American Political Science Review, 106(3): pp 570-587 Bayley, J., and Chabat, J. 2002. Transnational crime and public security: Challenges to Mexico and the United States. San Diego: Center for U.S.-Mexican Studies University of California San Diego. Bayley, J., and Dammert, L (Eds.). 2005. Public security and police reform in the Americas. University of Pittsburgh Press. BBC News. 2011. “Mexico’s Drug War: Lessons and Challenges”. BBC News. 31 December. Available at http://www.bbc.co.uk/news/world-latin-america-16337488 Beissinger, M. 1998. Nationalist violence and the state: Political authority and contentious repertoires in the former USSR. Comparative Politics 30(4): pp-401-422 Bejarano, A.M. and Pizarro, E. 2005. From ‘Restricted’ to ‘Besieged’: The changing nature of the limits to democracy in Colombia In: F. Hagopian and S. Mainwaring (Eds.) The third wave of democratization in Latin America: Advances and setbacks. Cambridge, Cambridge University Press, pp. 235-260.   Benitez-Manaut, R. 2008. Seguridad nacional y gobernabilidad en México: Criminalidad y Fronteras. Revista Criminologia y Sociedad. Numero 1 Año 1. Benitez-Manaut, R. 2010. México 2010: Crimen organizado, seguridad nacional y geopolítica In: R. Benitez Manaut (Ed.) Crimen organizado e iniciativa Mérida en las relaciones México-Estados Unidos. Mexico D.F: Colectivo de Análisis de la Seguridad con Democracia, Pp. 9-30. Betancourt, D. 1998. Mediadores, rebuscadores, traquetos y narcos. Las organizaciones mafiosas del Valle del Cauca entre la historia, la memoria, y el relato 1890-1997. Bogotá: Ediciones Anthropos.   Betancourt, D., Garcia, M.L. 1994. Contrabandistas, marimberos y mafiosos: historia social de la mafia colombiana, 1965-1992. Santafe de Bogotá: Tercer Mundo Editores.   355   Bewley-Taylor, D. 2012. International drug control: Consensus fractured. Cambridge: Cambridge University Press. Blancornelas, J. 1997. “Descobijaron a los narcojuniors” Zeta, November 14-20. Blancornelas, J. 2002. El Cartel: Los Arellano Felix, la mafia más poderosa en la historia de América Latina. Grijalbo Editores. Blancornelas, J., and Gonzalez, H. 1996. “Juniors mataron mas que narcos” Zeta October 4-10, p. 26A. and Blancornelas, J., Blancornelas, J., Gonzalez, H. 1997. “Dicho por el propio procurador Anaya: policias corruptos, no tontos” Zeta, August 15-21. Blancornelas, J., Ortiz, F., Gonzalez, H. 1996. “La Mafia sigue adentro de la PGR.” Zeta, November 20-26 . P. 26ª. Blatmann, C. 2008. From violence to voting: War and political participation in Uganda. American Political Science Review, 103(2): pp. 231-247. Blok, A. 2002. Mafia and blood symbolism In: Salter, F., (Ed.) Risky transactions: trust, kinship and ethnicity. New York: Berghahn Books, Pp. 109-129. Borunda, D. 2011. “Juárez mayor confronts feds: 'You don't rule city.'” El Paso Times, May 5. Bourdieu, P., and Wacquant, L. 2004. Symbolic violence In: Scheper-Hughes, N., & Bourgois, P. I. (Eds.) Violence in war and peace. Blackwell Pub, Pp. 272-274. Bowden, C. 2004. Down by the river: Drugs, money, murder, and family. New York: Simon and Schuster. Bowden, C. 2010. Murder City: Ciudad Juarez and the global economy’s new killing fields. Nation Books: New York. Bowden, M. 2002. Killing Pablo: The hunt for the world’s greatest outlaw. Atlantic Books. Brancati, D., and Snyder, J. 2011. Rushing to the polls: The causes of premature postconflict elections. Journal of Conflict Resolution 55: pp. 469-492. Brass, P. 1997. Theft of an idol. Princeton: Princeton University Press. Brinks, D. 2012. A tale of two cities: The judiciary and the rule of law in Latin America In: P. Kingstone and D. Yashar, (Eds.) Handbook of Latin American Politics. New York: Routledge, pp. 61-75.   356   Britto, L. 2010. A trafficker’s paradise: The “war on drugs” and the new Cold War in Colombia. Contemporánea 1(1): pp. 159-177. Britto, L. 2012. The marihuana axis: A regional history of Colombia's first narcotics boom, 1935-1985. Ph.D. Dissertation, New York University. Brubaker, R. 1996. Nationalism reframed. Nationhood and the national question in the new Europe. Cambridge: Cambridge University Press. Bruinsma, G., and Bernasco, W. 2004. Criminal groups and transnational illegal markets: A more detailed examination on the basis of Social Network Theory. Crime Law and Social Change 41(1): pp. 79-94. Bueno de Mesquita, E. 2008. The political economy of terrorism: an overview of recent work. The Political Economist. Bunker, R, and Sullivan, J. 2011. Extreme barbarism, a death cult, and holy warriors in Mexico: Societal warfare south of the border? Small Wars Journal 22 May 2011. Caldeira, T. 2000. City of walls: Crime, segregation, and citizenship in Sao Paulo. University of California Press. Camacho, A. 1994. La reforma de la policía Colombiana. ¿Esperanzas o frustraciones? Nueva Sociedad NRO. 129 Enero-Febrero 1994 : pp. 27-40. Camacho, A., and A. Lopez. 2001. From smugglers to drug lords, to “traquetos”: changes in the Colombian illicit drugs organizations In: C. Welna and G. Gallon (Eds.) Peace, democracy and human rights in Colombia. Notre Dame: University of Notre Dame Press, pp. 60-89. Campbell, H. 2009. Drug war zone: frontline dispatches from the streets of El Paso and Juárez. Texas, University of Texas Press. Casas Dupuy, P. 2005. Reformas y contrarreformas en la policía colombiana. Fundación Seguridad y Democracia. Castillo, N. 2009. Capital social y nivel de cohesión en Ciudad Juárez In: Diagnóstico sobre la realidad social, económica y cultural de los entornos locales para el diseño de intervenciones en materia de prevención y erradicación de la violencia en la región norte: el caso de Ciudad Juárez, Chihuahua, Comisión Nacional para Prevenir y   357   Erradicar la Violencia contra las Mujeres (CONAVIM), México, Secretaria de Gobernación (SEGOB), Gobierno Federal. Caulkins, J., and, Reuter, P. 1998. What price data tell us about drug markets. Journal of Drug Issues, 28(3): pp. 593-613. Caulkins, J., et al. 2006. Can supply restrictions lower price? Violence, drug dealing and positional advantage. Contributions in Economic Analysis & Policy, 5(1): pp. 1538-0645. Cave, D. 2011. A crime fighter draws plaudits, and scrutiny. The New York Times. December 23. Ceballos, R., and Cronshaw, F. 2001. The evolution of armed conflict in Medellín: An analysis of the major actors. Latin American Perspectives 28(1) Colombia: The forgotten war (Jan 2001): pp. 110-131. Chaparro, C. 2005. Historia del cartel de Cali. Bogotá: Intermedio. Chepesiuk, R. 2003. The bullet or the bribe: Taking down Colombia’s Cali drug cartel. Westport, Conn: Praeger. Clawson, P., and Lee, R. W. 1998. The Andean cocaine industry. Palgrave Macmillan. Comisión de Estudios sobre la Violencia. 1987. Colombia: Violencia y democracia: informe presentado al Ministerio de Gobierno. Bogotá: Universidad Nacional de Colombia. Comisión Nacional de Reparación y Reconciliación CNRR, Colombia. 2011. La huella invisible de la Guerra: desplazamiento forzado en la Comuna 13. Informe del grupo de memoria histórica de la Comisión Nacional de Reparación y Reconciliación. Comisión Nacional de Reparación y Reconciliación CNRR, Colombia. 2007. Disidentes, rearmados y emergentes: Bandas criminales o tercera generación paramilitar? Informe 1 de la Comisión Nacional de Reparación y Reconciliación. Comunicado Conjunto Sedena, PGR. 2011. “El Ejército Mexicano detiene a Noel Salgueiro Nevarez (a) “El Flaco Salgueiro”, fundador y líder de “Gente Nueva”, grupo delictivo de la organización criminal “Guzmán Loera.”” October 5.   358   Concha-Eastman, A., Espitia, V. E., Espinosa, R., and Guerrero, R. 2002. La epidemiología de los homicidios en Cali, 1993-1998: seis años de un modelo poblacional. Revista Panamericana Salud Publica, 12(4): pp. 230-239. Corporación Nuevo Arco Iris. 2011. Mafias y agentes ilegales buscan rentas y el poder local. Segundo Informe de la Corporación Nuevo Arco Iris, riesgos electorales 2011. October 2011. Bogotá: Nuevo Arco Iris. Craig, R. 1980. Operation Condor: Mexico's antidrug campaign enters a new era. Journal of Interamerican Studies and World Affairs, 22(3) (Aug. 1980): pp.345- 363. Crenshaw, M. 2007. Explaining suicide terrorism: A review essay. Security Studies 16(1): pp. 133-162. Cruz, J.M., Fernández de Castro, R and Santamaría, G. 2012. Political transition, social violence, and gangs: Cases in Central America and Mexico In: C.J. Arnson (Ed.) In the wake of war: Democratization and internal armed conflict in Latin America. Stanford, CA: Woodrow Wilson Center Press and Stanford University Press. Dammert, L and Malone, M. 2006. Does it take a village? Policing strategies and fear of crime in Latin America. Latin American Politics and Society, 48(4) Winter 2006: pp. 27-51 Davis, D. 2006. Undermining the rule of law: Democratization and the dark side of police reform in Mexico. Latin American Politics and Society, 48 (1): pp. 55–86. Davis, D. 2010. Irregular armed forces, shifting patterns of commitment, and fragmented sovereignty in the developing world. Theory and Society, 39: pp. 397-413. Davis, D. 2011. Theoretical and empirical reflections on cities, sovereignty, identity, and conflict In: D.E. Davis and N. Libertun de Duren (Eds.) Cities and sovereignty: Identity politics in urban spaces. Bloomington: Indiana University Press, pp. 226-256. De Francisco, G. 2005. El doble reto del conflicto armado y la seguridad pública: la evolución de la policía nacional de Colombia In: L. Dammert and J. Bailey (Eds.) Seguridad y reforma policial en las Americas: Tendencias y desafios. Siglo XXI Eds: Mexico.   359   Dillon, S. 1997. “A toll of disappearances in Mexico’s war on drugs.” The New York Times. October 7. Dillon, S. 1998. “Mexican drug lord found slain.” The New York Times, September 12. Dorn, N., et al. 1992. Traffickers, drug markets, and law enforcement. Routledge. Duncan, G. 2005. Del campo a la ciudad en Colombia: la infiltración urbana de los señores de la Guerra. Documento Cede 2005-2. Bogotá: Universidad de los Andes. Dudley, S. 2013. 5 Things the El Salvador gang truce has taught us. Insight Crime. March 12. Available at http://www.insightcrime.org/news-analysis/5-things-el- salvador-gang-truce-taught-us, Accessed [6 April 2013]. Duncan, G. 2005. Narcotraficantes, mafiosos y guerreros. Historia de una subordinación In: Narcotráfico en Colombia: Economia y violencia. Fundación Seguridad y Democracia. Dunning, T. 2011. Fighting and voting: Violent conflict and electoral politics Journal of Conflict Resolution 55: pp. 327-339. Duran-Martinez, A., Hazard, G., and Rios, V. 2010. 2010 Mid-year report on drug violence in Mexico. Transborder Institute, University of San Diego. Eilstrup-Sangiovanni, M., and Jones, C. 2008. Assessing the dangers of illicit networks: Why al-Qaida may be less threatening than many think. International Security, 33(2): pp. 7–44 El Colombiano. 2003. “Desarticulada en Medellin red logística de las FARC: Trabajo conjunto ha sido clave” Entrevista con General Leonardo Gallego, 2003. El Colombiano, September 13. El Colombiano. 2006. “Don Berna reconoció las desapariciones en la Comuna 13”. El Colombiano, October 6. El Diario. 2010. “Dejan 2 decapitados en Hacienda de las Torres.” El Diario, July 28. El Diario. 2010. “Qué quieren de nosotros.” El Diario, September 19. El Diario. 2011. “Pos sí, pero no tiré’: federal que encañonó a los escoltas de 'Teto.” El Diario de Juarez, May 5.   360   El Espectador. 2008.“Cerrojo a la Oficina de Envigado.” El Espectador, July 27. El Espectador. 2011. “El Rio la fosa más grande de Medellín.” El Espectador, February 8. El Espectador. 2011. “El Rio, la fosa más grande de Medellín” El Espectador, February 8. El Mundo, 2010. “Confesiones de un alcalde de Ciudad Juárez: no hay colaboración entre gobiernos.” El Mundo. October 13. El Pais. 2007. “El Ocaso de los narcos” El País, January 15 El Paso Times. 2010. “Expert: Azteca leader's arrest won't end violence”. El Paso Times, December 1. El Sol de Sinaloa. 2008. “Operativo Culiacán-Navolato; 2 mil 723 elementos federales.” El Sol de Sinaloa, May 14. El Tiempo. 1990. “Envigado: detenidos doce miembros del departamento de seguridad y control”. El Tiempo March 14. El Tiempo. 1991. “El Fin de Los Extraditables.” El Tiempo. July 4. El Tiempo. 1991. “El Fin de Los Extraditables”. El Tiempo. July 4. El Tiempo. 1995. “Bajo la presión de E. U. Asume comandante de policía de Cali”. El Tiempo March 9. El Tiempo. 1996. “Detenido el senador Gustavo Espinosa.” El Tiempo, February 15. El Tiempo. 2010. “Ante violencia urbana, Cali promueve el desarme” El Tiempo. August 31. El Tiempo. 2012. “Quien era Lorena Henao Montoya la viuda de la mafia?” El Tiempo, December 27. El Tiempo. 2013. “Son 903 los desplazados por bandas en el litoral del San Juan, Chocó.” El Tiempo, January 17. El Universal. 2004. “Desertan 1382 militares de elite.” El Universal. March 28. El Universal. 2011. “Balean a activista en Ciudad Juárez.” El Universal, December 2. Escalante, F. 2009. El homicidio en México entre 1990 y 2007. El Colegio de Mexico, Mexico D.F.   361   Escalante, F. 2011. La muerte tiene permiso. Nexos January 3, 2011.Available at: http://www.nexos.com.mx/?P=leerarticulo&Article=1943189. Evans, P. 1995. Embedded autonomy: States and industrial transformation. Princeton University Press. Faiola, A., 2011. “Sustaining the Medellin miracle: Colombia struggles to hold on to gains from globalization.” The Washington Post, July 11. Faletti, T. 2005. A sequential theory of decentralization: Latin American cases in comparative perspective, American Political Science Review 99(3): pp. 327-346. Fearon, J. 2004. Why do some civil wars last so much longer than others? Journal of Peace Research, 41(3): pp. 275-301. Fearon, J. 2005. Primary commodity exports and civil war. The Journal of Conflict Resolution, 49(4) Paradigm in distress? Primary Commodities and Civil War (Aug., 2005): pp. 483-507. Fearon, J.D. 2004. Why do some civil wars last so much longer than others? Journal of Peace Research 41 (3): pp. 275-301. Fearon, J.D., and Laitin, D. 2003. Ethnicity, insurgency and civil war. American Political Science Review 97 (1): pp. 75–90. Felbab-Brown, V. 2009. The violent drug market in Mexico and lessons from Colombia. Brookings Institution. Policy Paper Number 12. Felbab-Brown, V. 2010. Shooting up: Counterinsurgency and the war on drugs. Washington D.C: The Brookings Institution. Felbab-Brown, V. 2011. Human security and crime in Latin America: The political capital and political impact of criminal groups and belligerents involved in illicit economies. Applied Research Center, Florida International University. Finnegan, W. 2010. “Silver or Lead.” The New Yorker, May 31. Finnegan, W. 2012. “The Kingpins: the Fight for Guadalajara". The New Yorker, July 2. Flores, C. 2005. El estado en crisis: crimen organizado y política: desafíos para la consolidación democrática. Tesis para obtener el título de doctor en Ciencias Políticas y Sociales. Universidad Nacional Autónoma de México.   362   Fox, E, 2012, “US report on Colombia cocaine raises more questions than answers” Insight Crime , August 1. Available at http://www.insightcrime.org/news-analysis/us- report-on-colombia-cocaine-production-raises-more-questions-than-answers Accessed [16 February 2013] Franco, V. 2005. Prácticas hegemónicas de la coalición políticamente dominante en Medellín y su entorno urbano-regional. Estudios Políticos, No. 26. Instituto de Estudios Políticos: Colombia. Enero - Junio. 2005. Friman, R. 2004. Forging the vacancy chain: Law enforcement efforts and mobility in criminal economies. Crime, Law and Social Change 41: pp. 53-77. Friman, R. 2009. Drug markets and the selective use of violence. Crime, Law and Social Change 52(3): pp. 285-295. Fruhling, H. 2003. Policia comunitaria y reforma policial en America Latina. Cuál es el impacto? Serie Documentos. Centro de Estudios en Seguridad Ciudadana, Santiago. Fruhling, H., Tulchin, J., and Golding, H. 2003. Crime and violence in Latin America: Citizen security, democracy, and the state. Baltimore: Johns Hopkins University Press. Fukuyama F., and Colby, S., 2011. “Half a miracle”. Foreign Policy, May-June. Gambetta, D. (Ed.). 2005. Making sense of suicide missions. New York: Oxford University Press. Gambetta, D. 1996. The Sicilian Mafia: the business of private protection. Cambridge: Harvard University Press. Gambetta, D. 2008. Codes of the underworld: how criminals communicate. Princeton University Press. Ganor, B. 2008. Terrorist organization typologies and the probability of a boomerang effect. Studies in Conflict and Terrorism 31(4): pp. 269-283. Garzón, J.C. 2008. Mafia & Co.: las redes criminales en Mexico, Brasil, y Colombia. Bogotá: Planeta. Garzon, J.C. 2013. “La violencia que seremos” Razón Pública, February 10. Available at http://www.razonpublica.com/index.php/conflicto-drogas-y-paz-temas-30/3556-la- violencia-que-seremos.html, Accessed [10 February 2013]   363   Gibson, E. 2005. Boundary control: Subnational authoritarianism in democratic countries. World Politics 58(1): pp. 101-132. Golden, T. 1997 “Misreading Mexico: How Washington stumbled” The New York Times, July 11. Goldstein, P, et. Al. 1989. Crack and homicide in New York City, 1988: A conceptually based event analysis. Contemporary Drug Problems 16: pp. 651. Goldstein, P. 1985. The drugs violence nexus: A tripartite conceptual framework. Journal of Drug Issues, 15 (4): pp. 493-506. Goldstein, P. 1997. The relationship between drugs and violence in the United States of America In: UNODC. World Drug Report 1997. Gonzalez, G., and Tienda, M. 1989. The drug connection in U.S.- Mexican relations. Papers presented for the Bilateral Commission on the future of United States- Mexican relations. Dimensions of U.S.-Mexican relations Volume 4. San Diego: Center for U.S-Mexican Studies, University of California, San Diego. Gonzalez, H. 1997. “ Amigo de Xico Jr possible víctima del “Lobo” Hodoyán y Fabián “El Tiburón”. Zeta October 11-17, P. 26A. Gonzalez, H., and Ortiz, F. 1996. “Tijuana: las ejecuciones alrededor de la PGR” Zeta May 24-Apr. 3. P. 28A. Gootenberg, P. 2009. Andean cocaine: The making of a global drug. The University of North Carolina Press. Gootenberg, P. 2012. Cocaine’s long march North 1900-2010. Latin American Politics and Society 54(1): pp. 159-179. Granada, S., Restrepo, J., and Vargas, A. 2009. El agotamiento de la política de seguridad: evolución y transformaciones recientes en el conflicto armado colombiano In: Restrepo, J., and Aponte, D (Eds). Guerra y violencia en Colombia: herramientas e interpretaciones. Bogotá: Universidad Javeriana, CERAC, GTZ. Ch. 1. Grayson, G. 2010. Mexico: Narco violence and a failed state? New Brunswick, New Jersey: Transaction Publishers. Grzymala- Busse, A. 2008. Beyond clientelism: Incumbent capture and state building. Comparative Political Studies, 41(4-5): pp. 638-673.   364   Guerrero, E. 2010. Security, drugs and violence in Mexico: A survey. Prepared for the sixth North American Forum. Guerrero, E. 2011. La raíz de la violencia. Nexos June 1, 2011. Available at: http://www.nexos.com.mx/?P=leerarticulo&Article=2099328 Guerrero, E. 2012. La dispersion de la violencia. Nexos February 1, 2012. Available at: http://www.nexos.com.mx/?P=leerarticulo&Article=2102543 Gutierrez, A and Ramirez, I. 1997. “El Escurridizo.” Proceso, No. 1054, January 13. Gutierrez, A. 1996. “Hombre inteligente, respetado y querido.” Proceso, No. 1003, January 22. Gutierrez, F. 2004. Criminal rebels: a discussion of civil war and criminality from the Colombian experience. Politics and Society, 32(2): pp. 257-285. Gutierrez, F. 2007. Lo que el viento se llevó? Los partidos politicos y la democracia en Colombia (1958-2002). Bogotá: Editorial Norma. Gutierrez, F., and Baron, M. 2006. Estado, control territorial paramilitar y orden politico en Colombia: Notas para una economía política del paramilitarismo 1978-2004 In: G. Sanchez and F. Gutierrez (Eds.) Nuestra Guerra sin nombre: Transformaciones del conflicto en Colombia. Bogotá: IEPRI, Editorial Norma, Pp. 267-311. Gutierrez, F., and Jaramillo, A.M. 2004. Crime, (counter-) insurgency and the privatization of security: the case of Medellin (Colombia). Environment and Urbanization 16(2): pp. 17-30. Gutierrez, F., and Jaramillo, A.M. 2005. Paradoxical pacts In: De Sousa Santos, B., (Ed.) Democratizing democracy: beyond the liberal democratic canon. New York: Verso, Pp. 193-219. Gutierrez, F., and Sanchez, G. 2006. Nuestra Guerra sin nombre: Transformaciones del conflicto en Colombia. Bogotá: IEPRI, Editorial Norma. Introduction. Gutiérrez, F., Pinto, MT., Arenas, J.C., Guzman, T., Gutierrez, M.T. 2009. Politics and security in three Colombian cities. Crisis States Research Centre, Working paper series No. 2, Working Paper 44. London School of Economics, Development Studies Institute.   365   Guzman, A. 1993. Diagnóstico sobre la violencia homicida en Cali, 1993. Desepaz, Cidse. Guzman, A., and Camacho, A. 1990. Colombia, ciudad y violencia. Ediciones Foro Nacional. Guzman, A., and Moreno, R. 2007. Autodefensas, narcotráfico y comportamiento estatal en el Valle del Cauca, 1997-2005 In: Romero, M (Ed) La Parapolítica. La ruta de la expansión paramilitar y los acuerdos politicos. Bogotá: Corporación Nuevo Arcoiris. Pp. 165 Hansen, T., and Stepputat, F. 2001. States of imagination: Ethnographic explorations of the postcolonial state. Duke University Press, Pp. 3-37. Hernandez, T., and Negrete, J (Eds.). 2001. La Experiencia del PAN: Diez años de gobierno en Baja California. Tijuana: El Colegio de la Frontera Norte, Plaza y Valdés Editores. Herrera, L. 2008. El desgobierno de la ciudad y la política de abandono: Miradas desde la frontera de México. Universidad Autónoma de Ciudad Juárez. Universidad Complutense de Madrid. Hervella, J. 2006. Ciudad Juárez: Residencial narcofosas. El Paso, TX: Jesus A Blanco. Hess, K. 2009. Introduction to law enforcement and criminal issues. Ninth Edition. Wadsworth Cengage Learning. Hoffman, B, and McCormick, G. 2004. Terrorism, signaling, and suicide attack. Studies in Conflict & Terrorism, 27(4): pp. 243–281. Hope, A., and Clark, E. 2012. “Si los vecinos legalizan” Instituto Mexicano para la competitividad IMCO. Available at: http://imco.org.mx/images/pdf/presentacion- legalizacion-marihuana.pdf, Accessed [6 April 2013] Human Rights Watch. 2010. Paramilitaries heirs: The new face of violence in Colombia. Available at: www.hrw.org Accessed [February 4 2013] Human Rights Watch. 2013. Los desaparecidos de México: El persistente costo de una crisis ignorada. Idler, A. 2012. Exploring agreements of convenience made among violent non- state actors. Perspectives on Terrorism Volume 6, Issues 4-5. October.   366   Indepaz. 2012. Seventh report on presence of paramilitary groups 2011. Instituto Popular de Capacitación IPC, 2003. En deuda con los derechos humanos y el DIH – Diversas miradas críticas en Medellín. Relecturas 27. Medellín: IPC. Instituto Popular de Capacitación IPC, 2004. Un poco de luz en la oscuridad: Derechos Humanos Integrales. Relecturas 28. Medellín: IPC. Instituto Popular de Capacitación IPC, 2006. Entre la adversidad y la persistencia: Derechos Humanos en Medellín. Relecturas 30. Medellín: IPC. Instituto Popular de Capacitación IPC, 2007. Desponjándose de los miedos…. Emergen las víctimas. Informe de Derechos Humanos 2007. Relecturas 31. Medellín: IPC. Instituto Popular de Capacitación IPC, 2008. Selectividad y guerra: avanza el círculo vicioso. Informe de derechos humanos 2008. Relecturas 32. Medellín: IPC. Instituto Popular de Capacitación IPC, 2009. El péndulo de la violencia: Rupturas y continuidades del conflicto en Antioquia. Relecturas 33. Informe de Derechos Humanos 2009. Medellín: IPC. Instituto Popular de Capacitación IPC, 2011. Transporte público: Un botín en contextos de conflicto. Boletín Informativo 1 / 2011 Febrero 10. International Crisis Group. 2012. Desmantelar los nuevos grupos armados ilegales en Colombia: Lecciones de un sometimiento. Informe sobre América Latina No. 48, June 8 2012. Jaramillo, A. M., Ceballos, R., and Villa, M. I. 1998. En la encrucijada. Conflicto y cultura política en Medellín en los noventa. Medellín, Corporación Región. Jaramillo, A.M., 1994. Milicias populares en Medellín: entre la guerra y la paz. Medellín: Corporación Región. Juris, J. 2005. Violence performed and imagined: Militant action, the black bloc, and the Mass Media in Genoa. Critique of Anthropology, 25(4): pp. 413-432. Jusidman, C. (Coord.). 2008. La realidad social de Ciudad Juárez, Tomo 1,   367   Kaldor, M. 2001. New and old wars: Organized violence in a global era. Stanford University Press. Kalyvas, S. 1999. Wanton and Senseless: The Logic of massacres in Algeria. Rationality and Society, 11(3): pp. 243-285. Kalyvas, S. 2006. The logic of violence in civil war. Cambridge University Press. Karlin, M. 2012. “Fueled by war on drugs, Mexican death toll could exceed 120,000 as Calderon ends six-year reign” Truthout November 28, Available at http://truth-out.org/news/item/13001-calderon-reign-ends-with-six-year-mexican-death- toll-near-120000 [Accessed 28 November 2012] Kenney, M. 2007. From Pablo to Osama: Trafficking and terrorist networks, government bureaucracies and competitive adaptation. The Pennsylvania State University. Kilmer, B., Caulkins, J., Bond B., and Reuter, P. 2010. Reducing drug trafficking revenues and violence in Mexico: Would legalizing marijuana in California help? Santa Monica, CA: RAND Corporation, 2010. Available at: http://www.rand.org/pubs/occasional_papers/OP325 Kleiman, M. 2011. Surgical strikes in the Drug Wars: Smarter policies for both sides of the border. Foreign Policy Volume 90, No. 5 (September-December 2011). Kruijt, D., and Koonings. 1999. Societies of fear: the legacy of civil war, violence and terror in Latin America. London: Zed Books. Kydd, A., and Walter, B. 2006. The strategies of terrorism. International Security, 31(1): pp. 49-80. La Crónica de Hoy. 2005.“Los Zetas fueron capacitados por los kaibiles antes de desertar del Ejército mexicano.” La Crónica de Hoy. October 3. La Jornada. 2008. “Hallan en Tijuana tres cadáveres en tambos con ácido” La Jornada. October 1. La Jornada. 2005. “Cuatro ejecutados en Chihuahua y Sinaloa; uno era jefe policiaco de la primera entidad.” La Jornada. December 25. La Jornada. 2010. “En Juarez ejército se dedico a ver pugna entre carteles: consulado EU.” La Jornada, March.   368   La Jornada. 2011. “Violencia en Juarez por falta de coordinación: Comisión Nacional de Derechos Humanos.” La Jornada, February 21. Lafranchi, H. 2001. “Mexico declares war against drug traffickers”. The Christian Science Monitor. January 30. Lamb, R. 2010. Microdynamics of illegitimacy and complex urban violence in Medellin, Colombia. Ph.D. Dissertation, University of Maryland, College Park. Lankina, T. 2012. Sisyphean endeavor or worthwhile undertaking?: transcending within nation, within region sub-national democracy analysis. Comparative Democratization Newsletter, America Political Science Association. Volume 10, No. 1. Latin American Commission on Drugs and Democracy. 2012. Drugs and democracy: Toward a paradigm shift. Report by the Latin American Commission on Drugs and Democracy. Lazcano y Ochoa, M. 1992. Una vida en la vida sinaloense: Memorias de Manuel Lazcano y Ochoa. Universidad de Occidente. Lee, R., and Thoumi, F. 1999. The political-criminal nexus in Colombia. Trends in Organized Crime, 5(2): pp. 59-84. Lessing, B. 2012a. The logic of violence in criminal war: Cartel-State conflict in Mexico, Colombia, and Brazil. Ph. D. Dissertation. University of California, Berkeley. Lessing, B. 2012b. When business gets bloody: State policy and drug violence In: Small Arms Survey 2012: Moving targets. Cambridge University Press. Pp 41-72. Levitsky, S., and Way, L. 1997. Democracy with adjectives: Conceptual innovation in comparative research. World Politics (49) 3: pp 430-451. Levitsky, S., and Way, L. 2002. The rise of competitive authoritarianism. Journal of Democracy ,13(2): pp. 51-65. Ley, S. 2012. “El desafío de contra a nuestros muertos” Letras Libres, September 12, Available at: http://www.letraslibres.com/blogs/polifonia/el-desafio-de-contar- nuestros-muertos, [Accessed September 12 2012] Lieberman, E. S. 2005. Nested analysis as a mixed-method strategy for comparative research. American Political Science Review 99 (3): pp. 435-452.   369   Llorente, M.V. 2005. ¿Desmilitarización en tiempos de Guerra? La reforma policial en Colombia In: L. Dammert and J. Bailey (Eds.) Seguridad y reforma policial en las Americas: Tendencias y desafios. México: Siglo XXI Editores, Pp. 192-218. Lomnitz, C. 2005. Death and the idea of Mexico. Zone Books: New York. Lopez, A. 2006. Narcotráfico, ilegalidad, y conflicto en Colombia In: G. Sanchez and F. Gutierrez (Eds.) Nuestra Guerra sin nombre: Transformaciones del conflicto en Colombia. Bogotá: IEPRI, Editorial Norma, Pp. 405-440. Lopez, C (Ed.). 2010. Y refundaron la patria: De cómo mafiosos y politicos reconfiguraron el estado colombiano. Barcelona: Random House Mondadori Debate. Lozano, J and Morris, H. 2011. Documentary Impunity. Lupsha, P. A. 1991. Drug lords and narco-corruption: The players change but the game continues. Crime, Law and Social Change, 16: pp. 41–58. Lupsha, P.A, and Pimentel, S. 1997. The nexus between crime and politics: Mexico. Trends in Organized Crime, 3(1): pp 65-67. Magaloni, B. 2008. Voting for autocracy: Hegemonic party survival and its demise in Mexico. Studies in Comparative Politics. Cambridge University Press. Malm, A., and and Bichler, G. 2011. Networks of collaborating criminals: assessing the structural vulnerability of drug markets. Journal of Crime and Delinquency, 48 (2): pp. 271-297. Malone, M. 2010. The verdict is in: The impact of crime on public trust in Central American justice systems. Journal of Politics in Latin America, Vol.2 Issue 3. Marenin, O. 1985. Police performance and state rule: Control and autonomy in the exercise of coercion. Comparative Politics, 18 (1): pp. 101-122. Marosi, R. 2008. “Antidrug general ousted” LA Times, August 9. Martinez Canizales, M., and Howard, C. 2006. Mortalidad por homicidio, una revision comparativa en los municipios de Tijuana y Juárez , 1985-1997 In: P. Ravelo, and H. Dominguez (Eds.) Entre las duras aristas de las armas: violencia y victimización en Ciudad Juárez. Publicaciones de la Casa Chata, pp. 85-115.   370   McCormick, G., Giordano, F. 2007. Things come together: symbolic violence and guerrilla mobilisation. Third World Quarterly, 28(2) Special Issue: The long war— insurgency, counterinsurgency and collapsing States: pp. 295-320. Medina Franco, G., 2006. Historia sin fin…. Las milicias de Medellín en la década del noventa. Medellín: Instituto Popular de Capacitación. Mejia, D. 2010. Evaluating Plan Colombia In: P. Keefer and N. Loaiza (Eds.) Innocent bystanders: Developing countries and the war on drugs. The World Bank and Palgrave Macmillan, Pp. 135-164. Mendoza, E, and Mossa. R. 2012. “El Presidente de las 83 mil ejecuciones” Zeta, November 26. Mexico Evalua. 2012. Indice de víctimas visibles e invisibles de delitos graves. Mexico D.F. Misión de Apoyo al Proceso de Paz en Colombia, Organización de Estados Americanos MAPP-OEA. 2006. Séptimo Informe Trimestral del Secretario General al Consejo Permanente sobre la Misión de Apoyo al Proceso de Paz en Colombia MAPP- OEA. August 30, 2006. Molano, A. 2013. “Buenaventura: entre la pobreza y la violencia”. El Espectador, February 23. Molzahn, C., Rios, V., and Shirk, D. 2012. Drug violence in Mexico: Data and analysis through 2011. Transborder Institute, University of San Diego. Molzahn, C., Rodriguez, O., and Shirk D. 2013. Drug violence in Mexico: Data and analysis through 2012. Transborder Institute, University of San Diego. Monarrez Fragoso, J. 2009. Trama de una injusticia: feminicidio sexual sistémico en Ciudad Juárez. Tijuana: El Colegio de la Frontera Norte, Porrúa. Moncada, E. 2009. Toward democratic policing in Colombia? Institutional accountability through lateral reform. Comparative Politics, 41 (4): pp. 431-449. Moncada, E. 2010. Counting bodies: crime mapping, policing and race in Colombia. Ethnic and Racial Studies, 33(4): pp. 696-716. Moncada, E. 2011. Politics, business and violence: urban security in Colombia (1988-2008). Ph.D. Dissertation, Department of Political Science, Brown University. Moncada, E., and Snyder, R. 2012. Subnational comparative research on   371   democracy: taking stock and looking forward. Comparative Democratization Newsletter, America Political Science Association. Volume 10, No. 1. Moodie, E. 2010. El Salvador in the aftermath of peace: crime, uncertainty and the transition to democracy. Philadelphia: University of Pennsylvania Press. Moore, G. 2011. “The legacy of Sinaloa Cartel lieutenant El Flaco”. Insight Crime, October 17. Available at: http://www.insightcrime.org/investigations/the-legacy-of-sinaloa- cartel-lieutenant-el-flaco Moore, M. 1997. “Drug traffickers thrive amid Tijuana terror.” Washington Post, 16 February, sec. A, p. 1. Moser, C. 2006. Reducing urban violence in developing countries. The Brookings Institution: Washington D.C. Moser, C., and McIlwaine, C. 2004. Encounters with violence in Latin America. Urban poor perceptions from Colombia and Guatemala. Routledge Moser, C., and Rodgers, D. 2005. Change, violence and insecurity in non- Conflict situations. Overseas Development Institute. Murillo, M. V. and C. Martinez-Gallardo. 2007. Political competition and policy adoption. American Journal of Political Science, 51(1): pp. 120-139 N.A. 2012. Video Source. Complete Interview with El Diego available at http://www.youtube.com/watch?v=Wd_3cpZgCDk [Accessed November 15, 2012] Naim, M. 2006. Illicit: How smugglers, traffickers, and copycats are hijacking the Global economy. Anchor. Naylor, R.T. 1997. Mafias, myths and markets: On the theory and practice of enterprise crime. Transnational Organized Crime 3(3). Noroeste. 2008. “Son doce los muertos por balacera en la Guadalupe.” Noroeste, May 1. Noroeste. 2008. “Arroja hora violenta cuatro muertos en Culiacán.” Noroeste, May 1. Noroeste. 2008. “Más Narcomantas.” Noroeste, May 8. Noroeste. 2011. “Vuelve el pasado: Regresa Alfredo Mejía Pérez a la PEP.” Noroeste, March 30. North, D. 1990. Institutions, institutional change and economic performance.   372   Cambridge: Cambridge University Press. North, D., Wallis, J.J., and Weingast, B. 2009. Violence and social orders: A conceptual framework for interpreting recorded human history. Cambridge University Press. O’Donell, G. 1994. Delegative democracy. Journal of Democracy, 5(1): pp. 55- 69. Observatorio Social Alcaldía de Cali. 2010. Homicidios y pandillas juveniles en Santiago de Cali 2005-2009. Cali: Alcaldía de Santiago de Cali. Ochoa Reza, E. 2004. Multiple arenas of struggle: Federalism and Mexico’s transition to democracy In: Gibson, E (Ed.) Federalism and democracy in Latin America. The Johns Hopkins University Press, pp. 255-297. Office of the High Commissioner for Human Rights OHCHR. 2006. Informe de la Alta Comisionada de las Naciones Unidas para los Derechos Humanos sobre la situación de los derechos humanos en Colombia 2005. Office of the High Commissioner for Human Rights OHCHR. 2006. La Oficina en Colombia del Alto Comisionado de las Naciones Unidas para los Derechos Humanos reitera su posición sobre la “Ley de justicia y paz” Bogotá, October 13. Office of the High Commissioner for Human Rights OHCHR. 2011. Report of the United Nations High Commissioner for Human Rights on the situation of human rights in Colombia. February 3. Olson, M. 1993. Dictatorship, democracy and development. The American Political Science Review, 87(3): pp. 567-576. Open Society Foundations. 2012. The alternative world drug report: counting the costs of the war on drugs. Organization of American States. 2010. Pandillas delictivas. Available at: http://scm.oas.org/pdfs/2010/CP23778S.pdf Accessed [8 December 2012]. Ortiz Pinchetti, F. 1989. “El hombre más buscado del mundo.” Proceso No 650, April 17 1989, reprinted Edición Especial No, 32 February 2011. Osorio, J. 2012. Democratization and drug violence. Manuscript.   373   Oude Breuil, B.C., and Rozema, R., 2009. Fatal imaginations: death squads in Davao City and Medellín compared. Crime Law and Social Change, 52(4): pp. 405- 424. Palacios, M., and Safford, M., 2002. Colombia: País fragmentado, sociedad dividida. Bogotá: Editorial Norma. Palacios, M., and Serrano, M. 2010. Colombia y Mexico: las violencias del narcotráfico In: Alvarado, A. and Serrano, M (Eds.) Seguridad nacional y seguridad interior. México, D.F: El Colegio de México, pp. 105-155. Paoli, L. 2002. The paradoxes of organized crime. Crime, Law & Social Change, 37: pp. 51–97. Pardo, R. 1996. De primera mano: Colombia 1986-1994 entre conflictos y esperanzas. Bogotá: Grupo Editorial Norma. Payan, T. 2006. The three US- Mexico border wars: drugs, immigration and Homeland Security. Greenwood Publishing Group. Perez-Espino, J, and Paez, A. 2009. “Los Carrillo llegan a Juárez.” El Universal April 3. Personeria de Medellín. 2005. Informe sobre la situación de derechos humanos. Medellín 2005. Personeria de Medellín. 2011. Informe sobre la situación de derechos humanos. Medellín. Petrich, B. 2012. “Librarse de la sombra de Hank Rhon, el mayor reto para el PRI, según EU.” La Jornada. June 6. Pierson, P. 2000. Increasing returns, path dependence and the study of politics. American Political Science Review, 94(2): pp. 251-268. Pineda Jaimes, S and Herrera Robles, L.A. 2007. Alternancia y violencia: La seguridad pública en Ciudad Juárez 1992-2004 In: Zavaleta Betancourt, J.A (Coord). 2007. La seguridad pública local: Inseguridad, delincuencia y participación ciudadana en Ciudad Juárez. Universidad Autónoma de Ciudad Juárez, Consejo Nacional de Ciencia y Tecnología, Presidencia Municipal de Juárez, Fondo Mixto de Fomento a la Investigación Científica y Tecnológica, chapter II. Pinheiro Gawryszewski, V et at. 2012. Homicídios na região das Américas:   374   magnitude, distribuição e tendências, 1999-2009. Ciência & Saúde Coletiva, 17(12): pp. 3171-3182. Pinto, M.T. 2008. La disputa política en torno a la alcaldía de Santiago de Cali de 2007. Revista Sociedad y Economía, 14, p. 201-226. Poire, A. 2010 “Los homicidios y la violencia del crimen organizado.” Nexos, February 2010. Poole, M. A. 1995. The spatial distribution of political violence in Northern Ireland In: O’Day, A., (Ed.) Terrorism’s laboratory: The case of Northern Ireland. Aldershot: Darmouth, , pp. 27-45. Poppa, T. 2010. Drug lord: the life and death of a Mexican kingpin. Cinco puntos Press. Presidencia de la República de Mexico, Boletín de la puesta en marcha del operativo México Seguro. June 13 2005. Procuraduría General de la República, Mexico. Declaración ministerial MESA III AP/PGR/BC/TIJ/217/09-M-III. Programa de Naciones Unidas para el Desarrollo PNUD. 2008. Hacia un Valle del Cauca incluyente y pacífico. Informe regional de desarrollo humano 2008. Cali: PNUD. Przeworski A and Teune. H. 1970. The logic of comparative social inquiry. New York. Wiley and Sons. Przeworski, A. 1991. Democracy and the market: political and economic reforms in Eastern Europe and Latin America. Studies in Rationality and Social Change. Cambridge University Press. Quinn, T. 1991. “A Day with the Chessplayer.” Time. July 1. Quintana, V. 2008. Informe que presenta el diputado Victor Manuel Quintana correspondiente al periodo de receso de los meses de enero-febrero de 2008. Congreso del Estado de Chihuahua. Ragin, C. 1987. The comparative method: Moving beyond qualitative and quantitative Strategies. Berkeley. University of California Press. Ramirez Sanchez, M.A. 2009. Inseguridad pública en Tijuana, Tecate, y Rosarito: La paradoja del miedo y los delitos violentos In: López Estrada, S (Coord).   375   Diagnóstico sobre la realidad social, económica y cultural de los entornos locales para el diseño de intervenciones en Materia de prevención y erradicación de la violencia en la Región norte: el caso de Tijuana, Baja California Norte. SEGOB, Comisión Nacional para prevenir y erradicar la violencia contra las mujeres. Pp. 365-397. Ramirez, I.V., 2005. Medellín: Los niños invisibles del conflicto social y armado In: L. Dowdney (Ed.) Ni la guerra ni paz: comparaciones internacionales de niños y jóvenes en violencia armada organizada. Rio de Janeiro: Editorial Viveiros de Castro. Ramirez, M.C. 2010. Maintaining democracy in Colombia through political exclusion, states of exception, counterinsurgency, and dirty war In: E. D. Arias D. and Goldstein (Eds.) Violent democracies in Latin America. Duke University Press, pp. 84- 107. Ravelo, R. 2004. “La hora del desquite.” Proceso No. 1455, September. Ravelo, R. 2008. “Historias de Familia.” Proceso No. 1630, January 27, reprinted on Special Edition 32, May 2011. Ravelo, R. 2010 “Los Carrillo, bajo protección militar.” Proceso No. 1732, January 10. Ravelo, R., 2002. “Estructura intocada” Proceso No. 1324, March 17. Reed, S. 1997. “Certifiable: Mexico’s corruption, Washington’s indifference” The New Republic. Available at: http://www.pbs.org/wgbh/pagesAfrontline/shows/mexico/readings/newrepublic.html Accessed [9 November 2012] Restrepo, J., Spagat, M., and Vargas, J.F. 2006. El conflicto en Colombia: ¿ quién hizo qué a quién? Un enfoque cuantitativo” In: F. Gutiérrez and G. Sánchez (Eds) Nuestra Guerra sin nombre: transformaciones del conflicto en Colombia. Bogotá: Norma, pp. 5050-542. Reuter, P. 2009. Systemic violence in drug markets. Crime, Law and Social Change, 52(3): pp. 275-284. Reuter, P. 1983. Disorganized crime. The economics of the visible hand. Cambridge: The MIT Press. Revista Cambio. 2008. “Senador Juan Carlos Martínez se enreda por presuntos   376   vínculos con narcotraficantes.” Revista Cambio. November 2. Revista Cambio. 2011. “El hombre que maneja medio país desde la cárcel.” Revista Cambio. September 17. Revista Semana. 2005. “La Entrega de Don Berna.” Revista Semana Edicion 1204, May 30. Revista Semana. 2005. “La Jugada de Macaco.” Revista Semana, November 22. Revista Semana. 2006 “El Fin del Terror de Don Berna.” Revista Semana, Edicion 1206. June 13. Revista Semana. 2007. “Corrupción hasta el tuétano.” Revista Semana. August 11. Revista Semana. 2007. “Don Diego permanecerá en el bunker de la Fiscalía antes de ir a Combita.” Revista Semana. September 10. Revista Semana. 2007. “Los Secretos de Don Berna.” Revista Semana, Edición 1315. July 16. Revista Semana. 2007. “Un año de massacre de Jamundí en la batallón del ejército aniquila la major unidad antidrogas de la policía.” Revista Semana, May 21. Revista Semana. 2009. “La vida útil de los narcos.” Revista Semana, April 21. Revista Semana. 2012. “Fracaso la extradición de narcotraficantes?” Revista Semana April 14. Riaño, P. 2006. Dwellers of memory: Youth and violence in Medellín. Transaction Publishers. Rio Doce. 2007. “El Otro frente: copan al General Hidalgo Eddy y se le vence el plazo.” Rio Doce. June 11. Rios, V. 2012. How government structure encourages criminal violence: The causes of Mexico's Drug War. Ph.D. Dissertation, Department of Government, Harvard University. Rios, V. 2013. Why did Mexico become so violent? A self-reinforcing violent equilibrium caused by competition and enforcement. Trends in Organized Crime. Rios, V., and Shirk, D. 2011. Drug violence in Mexico: Data and analysis through 2010. Transborder Institute, University of San Diego.   377   Rodgers D, and Jones, G. 2009. Youth violence in Latin America: Gangs and juvenile justice in perspective. Palgrave Macmillan. Rodriguez Orejuela, G., and Rodriguez Orejuela, M. 2005. Las confesiones secretas de Miguel y Gilberto Rodriguez Orejuela: Documentos desclasificados. Bogotá: Quintero. Rodriguez, A. 2010. La iniciativa Mérida y la Guerra contra las drogas: Pasado y presente In: R. Benitez Manaut (Ed.) Crimen organizado e iniciativa Mérida en las relaciones México-Estados Unidos. Mexico D.F: Colectivo de Análisis de la Seguridad con Democracia, pp. 31-68. Rodriguez, S. 2011. “Truncaron sus estudios de secundaria… y se volvieron el ‘terror’ de 3 Colonias.” El Diario, December 26. Rodriguez, V., Ward, P.M. 1992. Policymaking, politics, and urban governance in Chihuahua: The experience of recent panista governments. U.S. Mexican Policy Report No. 3. Lyndon B. Johnson School of Public Affairs, The University of Texas at Austin. Roldan, M. 2003. Wounded Medellin: Narcotics traffic against a background of industrial decline In: J. Schneider and I. Susser. Wounded cities: destruction and reconstruction in a globalized world. Oxford: Berg, pp. 129-148. Romero, M (Comp.). 2007. Parapolitica: La ruta de la expansion paramilitar y los acuerdos politicos. Bogotá: Corporación Nuevo Arco Iris. Romero, M. 2003. Paramilitares y autodefensas 1982-2003. Bogotá: Instituto de Estudios Políticos y Relaciones Internacionales Universidad Nacional de Colombia. Romero, M. 2007. Nuevas guerras, paramilitares e ilegalidad: una trampa difícil de superar In: M. Romero (Ed.) La Parapolítica. La ruta de la expansión paramilitar y los acuerdos politicos. Bogotá: Corporación Nuevo Arcoiris, pp. 363-396. Rosi, G. 2011. Documentary “El Sicario, room 164” Film Forum, 2011. Ross, M. 2004. What do we know about natural resources and civil war? Journal of Peace Research, 41(3): pp. 337-356. Saenz Rovner, E. 2005. La conexión cubana: Narcotráfico, contrabando y juego en Cuba entre los años 20 y comienzos de la Revolución. Bogotá: Centro de Estudios Sociales (CES), Universidad Nacional de Colombia.   378   Saenz Rovner, E. 2008. Las redes de cubanos, norteamericanos y colombianos en el narcotráfico en Miami durante los años sesenta. Innovar, 18(32): pp. 111-126. Safford, F. 1965. Foreign and national enterprise in nineteenth-century Colombia. Business History Review, 39 (4) : pp 503-526. Salazar, A. 1990. No Nacimos Pa’semilla: la cultura de las bandas juveniles en Medellín. Cinep. Salazar, A. 2001. La parábola de Pablo: auge y caída de un gran capo del narcotráfico. Editorial Planeta. Salazar, A., and Jaramillo, A.M. 1992. Medellin: Las subculturas del narcotráfico. Santafe de Bogotá: Cinep. Sartori, G. 1970. Concept misformation in comparative politics. The American Political Science Review, 64 (4): pp. 1033-1053. Serrano, A. 2010. La multinacional del crimen: la tenebrosa oficina de Envigado. Debate. Serrano, M. 2007. Narcotráfico y gobernabilidad en México. Pensamiento Iberoamericano, 1: pp. 251-278. Shirk, D. 2010. Drug violence in Mexico: Data and analysis from 2001-2009. Transborder Institute, University of San Diego. Shirk, D. 2011. The Drug War in Mexico: Confronting a threat. Council Special Report No. 60, New York: Council on Foreign Relations. Skaperdas, S. 2001. The political economy of organized crime: providing protection when the state does not. Economics of Governance, 2: pp. 173-202. Skaperdas, S., and Konrad, K. 2005. The market for protection and the origin of the state. Cesifo Working Paper No. 1578. Skocpol, T. 1985. Bringing the state back in: strategies of analysis in current research In: Evans, P., et al (Eds.) Bringing the state back in. Cambridge University Press, pp. 3-38. Slone, M. 2000. Responses to media coverage of terrorism. Journal of Conflict Resolution, 44 (4): pp. 508-522. Smith, P.H. 1997. Drug trafficking in Mexico In: B. Bosworth, B., S. Collins, and N.C. Lustig (Eds.) Coming together: Mexico-U.S relations. Washington D.C.: The   379   Brookings Institution, pp. 125-147. Snyder, J. 2000. From voting to violence: Democratization and nationalist conflict. WW Norton. Snyder, R. 2001. Scaling down: The subnational comparative method. Studies in Comparative International Development, 36(1): pp. 93-110. Snyder, R. and A. Duran-Martinez. 2009a. Does illegality breed violence: Drug trafficking and state sponsored protection rackets. Crime, Law and Social Change, 52 (3): pp. 253-273. Snyder, R. and A. Duran-Martinez. 2009b. Drugs, violence and state-sponsored protection rackets in Mexico and Colombia. Colombia Internacional, 70 (Jul-Dic 2009): pp. 61-91. Soares, R., and Naritomi, J. 2010. Understanding high crime rates in Latin America: The role of social and policy factors In: R Di Tella, S. Edwards, and E. Schargrodsky (Eds.) The Economics of Crime: Lessons for and from Latin America. University of Chicago Press. Available at http://www.nber.org/chapters/c11831. Staniland, P. 2012. Cities on fire: Social mobilization, state policy, and urban insurgency. Comparative Political Studies, 43(12): pp. 1623-1649. Stannow, L. 1996. Social cleansing in Colombia. Thesis submitted in partial fulfillment for the requirement for the Degree of Master of Arts in the Department of Spanish and Latin American Studies, Simon Fraser University. Steele, A. 2011. Electing displacement: political cleansing in Apartado, Colombia. Journal of Conflict Resolution, 55(3): pp.423–445. Steller, T. 2012. “Years of killing hard to add up in Mexico.” Arizona Daily Star, November 27. Stone, H. 2012 “Under Calderón 80% of organized crime detainees went free.” Insight Crime. November 4. Available at http://www.insightcrime.org/news-briefs/calderon-80- organized-crime-detainees-free Accessed [28 January 2013] Stratfor. 2012. “Polarization and sustained violence in Mexico’s Cartel war” January 24. Available at: http://www.stratfor.com/analysis/polarization-and-sustained-violence- mexicos-cartel-war?utm_source=freelist-   380   f&utm_medium=email&utm_campaign=20120125&utm_term=engage&utm_content=link4&elq=632fd75 cce59438cb88a7ebe4b49cee4 [Accessed 1 February 2012 ]. Thelen, K and Streeck, W. 2006. Beyond continuity: institutional change in advanced political economies. Oxford: Oxford University Press. Sullivan, J, and Logan, S. 2012. La Línea: network, gang, and mercenary army. Free Republic, August 12, Available at: http://www.freerepublic.com/focus/f-news/2762765/posts Accessed [26 November 2012] Swanger, J. 2007. Feminist community building in Ciudad Juarez: A local cultural alternative to the estructural violence of globalization. Latin American Perspectives, 34(2) Globalizing Resistance: The new politics of social movements in Latin America: pp. 108-123. Tanner, M.S. 2000. Will the state bring you back in? Policing and Democratization. Comparative Politics, 33(1): pp. 101-124. Tarrow, S. 1994. Power in movement: Social movements, collective action, and politics. Cambridge: Cambridge University Press. The United States District Court For The Western District Of Texas El Paso Division. Indictment of Joaquín Guzmán Loera et al, Criminal No. EP-12-CR. April 11, 2012. The White House Office of the National Drug Control Policy. 2012. “Survey shows significant drop in cocaine production in Colombia”. Press release July 30 2012. Thoumi, F. 1995. Political economy and illegal drugs in Colombia. Boulder, Colorado: Lynne Rienner Publishers. Thoumi, F. 2002. Illegal drugs in Colombia: From illegal economic boom to social crisis. Annals of the American Academy of Political and Social Science, 582(1): pp. 102-116. Tickner, A. 2003. Colombia and the United States: from counternarcotics to counterterrorism. Current History (February 2003). Tickner, A., García, D., and Arreaza, C. 2011. Actores violentos no estatales y narcotráfico en Colombia. Politica Antidrogas en Colombia: Exitos, fracasos y extravios. Bogotá: Ediciones Uniandes, pp. 413-445.   381   Tilly, C. 1985. War making and state making as organized crime In: P. Evans et al. (Eds.) Bringing the state back in. Cambridge: Cambridge University Press, Pp. 169- 186. U.S. Department of Justice, National Drug Intelligence Center. 2011. National drug threat assessment 2011. United Nations Committee on the Elimination of Discrimination against Women CEDAW. 2005. Report on Mexico produced by the Committee on the Elimination of Discrimination against Women under article 8 of the Optional Protocol to the Convention, and reply from the Government of Mexico. 10-28 January 2005 Thirty Second Session. United Nations Office for Drug Control and Crime Prevention UNODCCP. 1999. Global illicit drug trends 1999. United Nations Office on Drugs and Crime UNODC. 2008. World Drug Report 2008. United Nations Office on Drugs and Crime UNODC. 2012a. Colombia coca cultivation survey 2011. United Nations Office on Drugs and Crime UNODC. 2012b. World Drug Report. United Nations Office for Drug Control and Crime Prevention UNODCCP. 1999. Universidad del Valle, Instituto de Altos Estudios Juridicos Politicos y Relaciones Internacionales. 1998. Construir la paz: Elementos de diagnóstico para el diseño de un modelo estratégico hacia el desarrollo, la seguridad y la paz en Cali. Cali: Universidad del Valle. Uprimny R., and Saffon, M.P. 2006. La ley de “justicia y paz”: ¿una garantía de justicia y paz y de no repetición de las atrocidades In: R. Uprimny et al (Eds) Justicia transicional sin transición: Verdad, justicia y reparación para Colombia. Bogotá: Dejusticia, Anthropos, Chapter 5. Uribe, M.V. 1990. Matar, rematar y contramatar: las masacres de la violencia en el Tolima, 1948-1964. Cinep. Controversia pp. 159-160.   382   Uribe, M.V. 2004. Dismembering and expelling: Semantics of political terror in Colombia. Public Culture, 16(1): pp. 79-95. Uribe, M.V., and Medina, T. 1995. Enterrar y callar: Las masacres en Colombia 1980-1993. Bogotá: Cinep. Valencia, L. 2007. Los caminos de la alianza entre paramilitares y politicos In: M. Romero (Ed.) La Parapolítica. La ruta de la expansión paramilitar y los acuerdos politicos. Bogotá: Corporación Nuevo Arcoiris, pp. 11-58. Valenzuela, J.M. 2009. El futuro ya fue. Socioantropología de los jóvenes en la modernidad. México, El Colegio de la Frontera Norte, Juan Pablo Editores. Valenzuela, J.M. 2009. Impecable y diamantina: P.S. Democracia adulterada y Proyecto Nacional. El Colegio de la Frontera Norte, Juan Pablos Editor: Mexico. Vanegas, G. 1998. Cali tras el rostro oculto de las violencias: Estudios etnográficos sobre la cotidianidad, los conflictos y las violencias en las barriadas Populares. Cali: Instituto Cisalva. Vargas, M., Lesmes, J., Tellez, E. 1996. El Presidente que se iba a caer: Diario secreto de tres periodistas sobre el 8000. Bogotá: Planeta. Velez, J.C. 2001. Conflicto y guerra: la lucha por el orden en Medellín. Revista de Estudios Politicos 18. Available at: http://quimbaya.udea.edu.co/estudiospoliticos/Rese%F1as/Abstracts/rev18c.htm Venkatesh, C., and Coughlin, B. 2003. The urban street gang after 1970. Annual Review of Sociology, 29: pp. 41-64. Venkatesh, S.A. 1997. The social organization of street gang activity in an urban ghetto. American Journal of Sociology, 103(1): pp. 82-111. Verdad Abierta 2012. “Cinco años de parapolítica.” Available at: http://www.verdadabierta.com/Especiales/cinco-anios-parapol/parapolitica-2012.swf Accesed [3 October 2012]. Vicepresidencia de la República de Colombia, Observatorio del Programa de Derechos Humanos y Derecho Internacional Humanitario. 2006. Dinámica reciente de la violencia en Cali. Bogotá, Vicepresidencia de la República, Teck Color Editores. Vicepresidencia de la República de Colombia. 2009. Caracterización del homicidio en Colombia 1995-2006. Bogotá: Observatorio del programa presidencial de   383   Derechos Humanos y DIH, Vicepresidencia de la República de Colombia, Editorial Scripto. Vivanco, J.M. 2004. Discurso de José Miguel Vivanco ante el Consejo Permanente de la OEA. Washington D.C. February 6 2004. Walton, J. 1977. Elites and economic development: Comparative studies on the political economy of Latin American cities. Institute of Latin American Studies, University of Texas at Austin. Washington Valdez, D. 2006. The killing fields: Harvest of women. The Truth about Mexico’s bloody border legacy. Burbank, California: Peace at the border. Weinstein, L. 2008. Mumbai’s development mafias: Organized crime, land development, and globalization. International Journal of Urban and Regional Research, 32(1): pp. 22-39. Weisel, D.L. 2002. The evolution of street gangs: an examination of form and variation In: Reed, W., and S. Decker (Eds.) Responding to gangs evaluation and research. National Institute of Justice, pp. 25-66. Williams, P. Cooperation Among Criminal Organizations In: M. Berdal and M. Serrano, (Eds.) Transnational Organized Crime and International Security: Business as Usual?. Boulder: Lynne Rienner, pp. 67-80. WOLA, 2012. “UN and US cocaine estimates contradict each other”. July 31. Available at: http://www.wola.org/commentary/un_and_us_estimates_for_cocaine_production_contradict_each_other Accessed [16 February 2013] Wood, E. 2006. Variation in sexual violence during war. Politics and society, 34(3): pp. 307-342. Youngers, C., and Rosen, E. Eds. 2005. Drugs and democracy in Latin America: The impact of U.S. policy. Boulder: Lynne Rienner Publishers. Zavaleta Betancourt, J.A (Coord). 2007. La seguridad pública local: Inseguridad, delincuencia y participación ciudadana en Ciudad Juárez. Universidad Autónoma de Ciudad Juárez, Consejo Nacional de Ciencia y Tecnología, Presidencia Municipal de Juárez, Fondo Mixto de Fomento a la Investigación Científica y Tecnológica.   384   Zavaleta Betancourt, J.A; Chavira Chavira R and Sánchez Carlos, J. 2007. El consenso de la inseguridad y las tasas delictivas In: Zavaleta Betancourt, J.A (Coord). La seguridad pública local: Inseguridad, delincuencia y participación ciudadana en Ciudad Juárez. Universidad Autónoma de Ciudad Juárez, Consejo Nacional de Ciencia y Tecnología, Presidencia Municipal de Juárez, Fondo Mixto de Fomento a la Investigación Científica y Tecnológica, Chapter I. Zepeda Bustos, S. 2010. Violencia y política electoral en Chihuahua. El Cotidiano, 164: pp. 11-18. Universidad Autónoma Metropolitana – Azcapotzalco. Distrito Federal, México. Zeta. 2007. “Pandillas: una amenaza en Tijuana.” Zeta 1709. December 29 2006 to January 4 2007.   385   APPENDIX 1 CODING PROTOCOL FOR DRUG VIOLENCE EVENTS367 This protocol explains the criteria used to classify information about drug violence events in Cali, Ciudad Juarez, Culiacan, Medellín and Tijuana using the main local newspapers in each city (in terms of circulation). These newspapers were El Pais (Cali), El Diario and El Mexicano (Ciudad Juarez), el Noroeste (Culiacán), El Colombiano (Medellín), and El Sol (Tijuana). The unit of analysis is the city, but the database also includes information on metropolitan areas and surrounding municipalities for each city. This document describes all the information coded for each event identified as drug related violence. An event of violence was considered drug related if a) the police, observers, or journalists attributed the action to drug trafficking b) the perpetrators accepted their responsibility c) the method suggested that the event was drug related (type of arms, torture, see Chapter 2 for an extended discussion on the definition of drug related violence). For each event the following information was coded: ID: consecutive identification number for each event preceded by a letter that corresponds to the newspaper source (thus, C1 is the first event coded in El Colombiano) 1. BASIC INFORMATION: Date and place of occurrence a. Date: Date in which the event took place. The date was inferred if there was no clear information (i.e. an event occurred 10 days ago). If there was no date or reference to time frame, it was assumed that the event occurred the day before the publication of the newspaper. b. Location: • Municipality and neighborhood: If the event occurred on an intermunicipal highway, the closest municipality was coded. • Place of occurrence: Name of the place (if available) and type of place according to the following categories: bar, restaurant, highway (connects municipalities), street (urban roads), mall, rehabilitation center, field, commercial building, public building (including prisons), creamery, park, private residence (including countryhouses), river (including ocean), cemetery, abandoned lot. If the victims died at a clinic after the                                                                                                                 367 This protocol draws on the protocol created by Benjamin Lessing for his project “Observatory of Drug Related Violence” OBIVAN. It was adapted for the specific needs of this dataset, identifying categories that indicate visible and low visible violence.   386   attack, the original location of the attack was coded. If an attack occurred in front of a residence or restaurant, on the street, the location was coded as street. • Urban: Determines if the event occurred in places of low population density and where agricultural activities prevail (rural) or in areas of high population density and with primacy of commercial activities (urban). The cell contains a 1 if the area was urban and 0 if rural. • Public: Public space was defined as any place where the land is public and access is unrestricted, such as parks, streets, highways, public buildings, and other places that become public because of the service they provide such as bus stations, libraries, schools, hospitals, malls. Private spaces are private residences, commercial buildings. Bars and restaurants were classified as private because access can be restricted. If an attack occurred in front of a private residence in the street, the location was classified as public. The cell contains a 1 if the place was public, and 0 if private. 2. TYPE OF EVENT: a. Unilateral action: An event that did not imply a reaction of the victims, or did not lead to combat. If an attack caused a retaliation that shortly followed the attack but was not immediate, the two events were coded separately. If a combat followed the arrival of police to a crime scene, the two events were also coded separately. The cell contains a 1 for a unilateral action, and 0 otherwise. b. Combat: An armed confrontation or exchange of fire, implies a reaction from the target of the attack. If an attack generated a reaction by bodyguards it was coded as combat. The cell contains a 1 for combat, and 0 otherwise. c. Common grave finding: In this cell 1 refers to the finding of three or more bodies in the same place. A common grave is created with the purpose of hiding bodies, therefore the finding of bodies in the middle of a street was not classified as common grave. d. Corpse finding: In this cell 1 refers to the finding of one or two bodies, when the bodies were found after the violent attack took place and there were no witnesses to determine the circumstances of the event. For example if it could be determined that there were paid killers on the scene, a 0 appears in this cell. 3. SPECIFIC TYPE OF ACTION a. Threat b. Publicity action c. Murder d. Attempt (when there was an attempt on the life of a person but the person did not die, or when an explosive was deactivated). When the objective was achieved, the action was not classified as an attempt.   387   e. Explosion f. Forced dissappearance (there was no information on the victims’ whereabouts) g. Kidnapping: When there was a ransom or the captors were known. If the victim returned alive or paid the ransom, no method was classified. If captors were not known and the victim was murdered, specific type of action was classified murder, and the method as dissappearance. h. Massacre i. Combat j. Enforcement operation: if there were no casualties in an enforcement operation, no method was classified. If there was a response from delinquents then the action was enforcement operation and the method was confrontation. If the operation was in response to a criminal action then the type of action was combat (if the reaction was immediate) k. Social cleansing (murder of unprotected sectors of the population such as drug abusers, prostitutes, petty criminals or homeless people). 4. GROUPS THAT PARTICIPATED IN THE ACTION (proper names): Up to three groups were coded. If no specific group was mentioned, the following categories were used: a. Armed command: (three or more people, armed). If the information did not clarify the number of people participating, but a group action could be inferred, it was classified as armed command. b. Common delinquent: if the perpetrators were identified as common criminals. c. Armed group: Armed collective with a political agenda. d. Organized crime: (if the note clearly identified the responsibility). Drug distributors were included in this category. e. Gang: Included terms like cholos, bandas, barrios. If a group of young people was armed with knives it was classified as gang rather than armed command. If there were only one or two individual perpetrators without clear membership in a group, this category was left blank. In a few cases, if the clipping provided information that allowed the coder to infer the group, then the group was inferred. For example a note from the Noroeste referred to a group of former military operating in Tamaulipas, and the coder inferred the group to be the Zetas, a criminal organization formed by former military and originally from Tamaulipas. If there was a combat, group 1 is the group that initiated the action Origin of atribution of responsibility: For each group the database determines the origin of the information that led to identification in the following categories.   388   a. Divergence: The cell contains a 1 if there was divergence about the groups involved, 0 otherwise. b. Acusation: The cell contains a 1 if the report explicitly acused one group, 0 otherwise. c. Suspicion: The cell contains a 1 if the group was only mentioned as a potential perpetrator. d. Responsibility: The cell contains a 1 if the group itself claimed responsibility, 0 otherwise. Source: When there was suspicion or accusation, the origin of the information was included in this column. The possible sources of information are: civilian, state, armed group, journalistic, police or military. Method to claim responsibility: This cell is filled when the cell responsibility contains a 1. The possible methods are: public statement, note in the location of events, note on corpses, direct comunication with a person, direct communication (of the perpetrator) with a person like public authority or journalist, propaganda action (if the perpetrators claim out loud their responsibility). If there was a corpse with a note, but the note did not claim responsibility, this cell was left blank, except when the note threatened a specific group and it could be assumed that the source was another organized crime group. In such case the origin of attribution was classified as suspicion, the method was a note, and the source was journalistic. 5. METHOD: This column codes the methods used to perpetrate lethal actions. When one action combined different methods, the cell coded the most visible one. List of methods: a. Head in cooler (when a head was found in a cooler, separated from the rest of the body) b. Car bomb c. Wrapped in blanket: when the corpse was wrapped in a blanket. If the corpse had signs of torture, then the method was torture, but if the body had signs of shots, then the method that prevailed was “wrapped in blanket.” Includes corpses found in plastic bags, bedsheets. d. Combat with explosives e. Combat with fireguns f. Combat with knives g. Frozen corpse h. Use of explosives and fire guns (without combat). When more than three people perpetrated an attack, the method classified was sicariato, and the type of arm was explosives. i. Execution j. Grenade (when a grenade was used in the attack) k. Fire (of a place) l. Mutilation or incineration with a note m. Simple mutilation or incineration (without note). n. Forced disappearance: when the person disappeared. Includes methods   389   identified as paseo or levantones. If a person disappeared but then was found alive, the type of action was kidnapping and there was no method classified. If a disappeared person was found with signs of torture, disappearance prevailed in the classification. If a body was not recovered, both the type of action and method were forced disappearance. There are three possible combinations of type of action-method in this case. If the person disappeared and the body was found, the type of action was murder, the method disappearance. If the person appeared alive, the type of action was kidnapping and the method disappearance. If the person never appeared, both type of action and method were classified as forced disappearance. o. Sicariato: Is the use of paid assassins. Given the difficulty of determining if the perpetrator was a paid assassin, sicariato was coded depending on whether the event suggested planning and previous identification of the target: the criteria used to determine this was if the action involved more than three people, a moving vehicle, or attacks from one moving vehicle to another. In Colombia, an action that included two people on a motorcycle was coded as sicariato. p. Torture: if the note referred mutilation, then the method classified was mutilation, and if there was mutilation and a note on the body, then the latter method prevailed. Torture usually involves the finding of naked bodies, with hands tied, and signs of heavy beating. q. Simple use of knives r. Simple use of fire gun: Generally when a single individual or two attacked, using few shots, generally in private locations s. Explosives t. Sexual violence (any instance in which a murder was accompanied by signs of sexual violence, the latter took precedence) u. Strangulation and asphyxiation v. Simple use of blunt objects w. Corpse with a note (when there were no mutilations) x. Poisoning Type of arm used: This cell contains the arms used to cause the death. Unlike the method, the arm coded was the one that caused the death, not the most visible. a. Knives b. Fire guns c. Fire elements (gasoline) d. Explosives e. Assault weapons f. Blunt objects g. Rope, duct tape h. Plastic bags 6. TARGET: The apparent target of the violent attack. Except in cases where the event was clearly a result of interpersonal violence, the classification assumed the target using the occupation of the victim, i.e, a murder of a cop was interpreted as police or military target. Considering the discussion in chapter 2, the information   390   about target was mostly speculative and therefore for the analysis in the dissertation we have preferred to use the category type of victim, explained below. The categories included are: Civilians, Rival organizations, common delinquent, civil servant, public infrastructure, private infrastructure, members of the same organization, criminal organization (when it was not clear whether it was rivals or members of the same organization), police or military. Street level drug distributors were categorized as criminal organization. a. Source of attribution: As in the group variable, this cell contains information about whether the target was identified through suspicion, accusation, direct responsibility of the perpetrators, or if there was divergence among sources. b. Source: Origin of the attribution (as in the Group variable) 8. VICTIMS: These cells contain information on the number and type of victims. The number in each cell corresponds to the number of victims in each of the following categories: a. Dead b. Wounded c. Detained (when law enforcement detained people) d. Retained (when an illegal organization retained people) e. Dissappeared (when a person was taken and there was no information about captors or whereabouts) f. Not wounded (as when people escaped from an attack without wounds) The dead and wounded categories in turn contain subcategories that include the number of victims per subcategory: a. Member of criminal organization b. Relative of member of criminal organization c. Law enforcement (police and military) d. Public authority (politicians or civil servants) e. Common Delinquent g. Female and males over 25 years h. Children (male and female under 25) 9. OCUPATION: Profession, ocupation, or job of the victim (only for fatal victims, if an action only wounded people there was no classification of ocupation). If there was more than one victim and more than one ocupation identified, each different ocupation was determined in a second or third column. The possible categories were: Lawyer, attorney’s agent, peasant, merchant or businessperson, social leader, high school student, university student, civil servant, mechanics, members of criminal organization, journalist, police or military, politician, presumed delinquent, professional, non-professional worker (drivers, general services employees, guards, construction workers, employee of a store), human rights worker, informal economy worker (street vendors). When a clipping identified the ocupation of the victim and a presumption of delinquency, the ocupation   391   was classified in this column, and the presumed delinquency in the subcategories of victim. 10. SOURCE OF ORIGIN OF INFORMATION a. Name of newspaper b. Section c. Page d. Date of publication (month, day, year) 11. OBSERVATIONS: contains a summary of the press clipping for coding reliability.   392