«My research agenda focuses on understanding a single question: If Mexico has long been a player in the illegal-drug industry, why is it only now that ...»
2.1 Where do drug traﬃckers operate? Mapping territorial areas of inﬂuence of Mexican DTOs Goal: Mapping areas of operation of Mexican DTOs at the municipal level with year to year variation Research design: Any eﬀort to understand why the Mexican drug traﬃcking industry became violent must start by knowing where traﬃckers operate. Up until now, public information about the industry came mostly from ethnographic and journalistic narratives that provide broad and vague mentions of the operations of DTOs at some Mexican states (Blancornelas 2002, Gomez 2005, Cruz 2009, Osorno 2009, Reveles 2009, Ravelo 2010, Rodriguez Castaeda 2010 and Hernandez 2010). Two other private sources of information on DTOs’ territoriality are also available. Intelligence oﬃces such as the Center for Research and National Security (CISEN) –Mexico’s intelligence service oﬃce– or the Drug-enforcement Administration (DEA), have generated maps. However, their contents are regarded as conﬁdential because of their usefulness for predicting national security policy and enforcement strategies. The second one, Stratfor, a consulting ﬁrm, has become the prime source of information on the topic (see for example NPR 2010), even though their identiﬁed territories are broad, quite imprecise, and rely on undisclosed sources as their inputs (Stratfor 2011).
As long as we know nothing about when and where a DTO started operating in a city/municipality, we won’t be able to accurately test any hypothesis regarding the eﬀects of enforcement or competition on drug-related violence. Consider, for example, studies of the impact of US-Mexico gun trade in fueling DTOs activities. To the extent of my knowledge, the only empirical model available on the topic controls for “cartel activities” using “dummies for cartel states”, an imprecise category including those states “where the leaderships of Mexico’s largest cartels were based (sic)” without an explicit mention to how they were picked (Chicoine 2011)7.
To overcome this severe data problem, I will develop a tool for tracking the territories in which a speciﬁed criminal organization operates using web information. Criminal organizations are Coauthoring with Michele Corcia, research fellow at Harvard’s Center for International Development Similar critics apply for studies on the impact of drug legalization in Mexico’s violence. As RAND Corporation (Kilmer, Caulkins, Bond and Reuter 2010) points in their own study, the diﬃculty with their projections is that they “require good information about [Mexican traﬃckers’] current level of involvement” in domestic territories.
traditionally tracked by intelligence services using information provided by local informants or individuals subject to criminal investigation. The tools are costly and require establishing longterm interactions at the ﬁeld with signiﬁcant risks for those collecting the information. My tool develops inexpensive intelligence information, taking advantage of the role that forums, blogs, and digital media have as depositaries of valuable, private information. I will adapt a search algorithm and use it to track all mentions done at news and specialized websites of the operations or presence of criminal organizations or any of their members. My program processes this information and returns a list of places and dates where a criminal organization was mentioned, which we later use as input to determine their main areas of operation. I also quantify the risks of missing and misleading information. Finally, I test our tool by tracking areas of operation of Mexican drug traﬃcking organizations (DTOs).
The results of my program will be complemented with a manual search of narco-messages8 in blogs, forums, local newspapers and general press [Figure 9]. Narco-messages will proxy for territorial presence by assuming that senders and recipients of them operate in the territory where it was left. I have already collected a dataset of about 1,400 diﬀerent narco-messages most of them providing explicit information about the operation of a particular DTO/traﬃcker in an area.
Empirics: Python-code for automatized searches in Google news and selected blogs and Geographical Information Systems Input: List of 350 drug traﬃckers, 54 gangs and their DTO of aﬃliation; other key search terms Output: Maps of areas of operation of Mexican DTO’s to be shared, as open source, to all researchers studying drug-related violence in Mexico.
Preliminary results: The diﬀerence between the information we have (Stratfor) and what I expect to get is pictured in Figure 10.
”Narco-messages” are a particular form of communication used by Mexican drug traﬃckers. They are billboards left on the streets to among other reasons, clarify the reasons why they assassinated someone, intimidate other potential victims, identify themselves or their victims, communicate with citizens around the area, or give instructions to the investigators or journalists who know will eventually come to record the messages. Narcomessages go all the way from maxims like “you cannot be in good terms with God and the Devil (Redaccion Vanguardia 2010),” to fortuitous expressions like “Merry Christmas, jo, jo jo (sic) (Bueno 2009),” to messages directed to “the brave, noble and loyal people” of a Mexican town, letting them know that “this [violent vendetta] is for the good of all (Delgado Aleman 2009)” Figure 8: Narco-message reading “So you learn to respect. Here I am leaving your garbage Miguel (...) Attn. G-1”
2.2 How do drug traﬃckers move? Identifying patterns of criminal territorial mobility9 Goal: Identifying the logic behind patterns of criminal territorial mobility.
Research design: Current research identiﬁes two distinct motivations determining criminal patterns of territorial mobility (Gambetta 2010). Either criminals move proactively in order to expand their business into potential new markets, or reactively, such as when they ﬂee from law enforcement operations. Using the maps of DTOs areas of operation created at section 2.1, I will tests these two hypotheses.
Empirics: Network analysis and OLS estimation. Using an analogy of the Product Space, a tool used to study the evolution of comparative advantage and countries’ productive structures (Hidalgo, Klinger, Barabasi, and Hausmann 2007), I will identify the common patterns that DTOs have followed when moving along Mexico’s territory.
I will calculate network proximity measures between all 2,500 Mexican municipalities using information of traﬃcker’s presence year by year as my input. The result will be a 2,500*2,500 matrix of proximity measures. Drawing upon Product Space terminology, if many DTOs (i.e countries) operate at the same municipalities (i.e. produce the same products), those municipalities (i.e. products) will be close to each other. To visualize the results, I will create a network with 2,500 nodes (one per municipality), linking only those where proximity measures Coauthoring with Michele Corcia, research fellow at Harvard’s Center for International Development
Figure 9: Estimates of drug-related violence for Mexico City (DF)
are above a deﬁned threshold. I will name it the “map of mobility patterns.” All things being equal, a traﬃcker operating in a municipality will have higher chances to move her operations into municipalities linked to that municipality than into those that are not.
Finally, using this network as my dependent variable, I will test whether municipalities are closer because of market dynamics (proactive criminal movements) or law enforcement dynamics (reactive criminal movements).
Section 3: Traﬃckers’ contribution to violence
3.1 The eﬀects of DTOs’ territorial competition in violence Goal: Empirically testing the eﬀects of territorial competition in drug-related homicides.
Research design: The Mexican government has repeatedly claimed that violence has erupted as a result of increased territorial competition between DTOs (Poire 2011), but iit is precisely because of a lack of public data on when and where Mexican drug traﬃckers operate that the accuracy of such interpretation has never been tested empirically. Using the maps of DTOs areas of operation created at section 2.1, I will test this hypothesis.
Empirics: DV: Drug-related homicides (oﬃcial, press and own measures –as calculated in Section 1) IV: The eﬀective number of DTOs operating in a single municipality at a particular time calculated as a Herﬁndahl-Hirschman Index (HHI) of market concentration. I expect higher concentration to be associated with lower violence. The share of the market that each DTO controls in a municipality will be a function of the share of municipalities that such DTO controls in the state where the municipality is located. Following standards of the Antitrust Division of the Department of Justice, I will consider indices between 0.1500 and 0.2500 to be moderately concentrated and indices above 0.2500 to be concentrated.
3.2 Why have incentives for competition increased?
Goal: Understanding how large structural changes in Mexico’s illegal drugs industry are the root causes of higher territorial confrontation.
Research design: In this section I will provide a formal approach to understand, based on industrial organization entry models (Green and Porter 1985, Rotemberg and Saloner 1986) and principal-agent models, why incentives for competition increased. This will be the narrative behind results given at Section 3.1.
Traﬃckers ﬁght more because of institutional and structural changes in Mexico’s illegal drugs industry, particularly in what I identify as the evolution of DTOs into higher-proﬁt, morediversiﬁed ﬁrms.
Table 1: Selected study cases Case Study Outcome was violent Outcome was peaceful Baja California Teo vs Tijuana, 2008 Tijuana and Sinaloa, 1998 Michoacan Familia vs. Zetas, 2006 Familia and Sinaloa, 2010 Tamaulipas Golfo vs Zetas, 2008 Golfo and Zetas, 2001 The reason why Mexican traﬃckers became more competitive can be found in institutional and structural changes, particularly in what I identify as the evolution of Mexico’s DTOs into higherproﬁts, more-diversiﬁed ﬁrms. The combination of increased law enforcement in Columbia and decreased law enforcement in Mexico made Mexican traﬃckers wealthier and allowed them to extend their operations into other illegal businesses such as kidnapping, extortion, robbery and domestic distribution of drugs. With higher proﬁts and with the newly acquired ability to get revenues from other illegal activities that do not require the same level of structural organization or long-lasting corruption agreements that importing drugs into the US requires, middleman traﬃckers who were previously forced to work under the supervision of traditional Mexican kingpins realized the proﬁtability of working independently. They became increasingly autonomous in their actions and ﬁnances. Competition increased when middleman split from their original organizations because market conditions decreased the payoﬀs for remaining loyal.
Case Studies: I will present study cases to provide evidence supporting my narrative. In all cases an external shock altered oligopolistic market arrangements, but the change resulted in violent confrontation in only half of the cases. The diﬀerence between those that had violent outcomes and those that did not lies in the degree of competition that the shock generated.
Competition was larger when the amount of proﬁts and the degree of diversiﬁcation of the local criminal industry was larger.
The inputs for my study cases will come from (a) an extensive analysis of my narco-messages dataset described in Section 2.1, (b) interviews and ﬁeldwork conducted at Mexico’s Drug War Zone (Rios 2008, 2009 and 2010), and (c) the reinterpretation of journalists’ accounts that attribute violence to the demise of the PRI (Blancornelas 2002, Gomez 2005, Cruz 2009, Osorno 2009, Reveles 2009, Ravelo 2010, Rodriguez Castaeda 2010 and Hernandez 2010).
Section 4: Government’s contribution to violence
4.1 The eﬀects of government’s enforcement in violence Goal: Empirically testing the eﬀects of law enforcement operations in drug-related violence.