«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 ...»
Research design: It has been claimed that none of the drug-related violence Mexico has experienced since 2007 would have existed if Mr. Calderon, president of Mexico, did not had launched an oﬀensive against drug traﬃckers (Aguilar and Castaeda 2009). Yet, we still lack solid empirical evidence of this assertion. Most literature has been descriptive, pointing to the arrest of middleman traﬃckers as Mr. Beltran Leyva or Mr. Osiel Cardenas as the starting point of the escalation in violence (Freeman 2006, Guerrero 2009). The only empirical test available (Poire and Trujillo 2011) ﬁnds no relationship between the detention of middleman traﬃckers and increases in violence. However, this test uses an inadequate econometric speciﬁcation and relies on undisclosed information sources. I will conduct my own empirical test.
I will explore two ways in which law enforcement may increase drug-related violence. First I will explore how law enforcement may directly increase violence by generating a higher number of deaths when enforcement operations claim the lives traﬃckers and law enforcement oﬃcials.
Secondly, I will explore how law enforcement may indirectly increase violence by further increasing the incentives for territorial competition (i.e. violence between DTOs generated when enforcement operations lead to a weakened DTOs being attacked by a stronger one) or by generating conﬂicts within DTOs when new leaders have to be picked to replace detained leaders.
Empirics: I will use HHI measures of periods before and after the capture of middleman traﬃckers for inputs. Each measure will be taken within the territories of operation of each middleman. I will test whether the capture of a drug-traﬃcker increases DTOs competition.
Note that because of changes in structural variables identiﬁed in section 3.2, I expect competition to increase even if law enforcement was absent. To avoid attributing the eﬀects of structural changes to law enforcement, I will use a diﬀerence-in-diﬀerence approach and F-tests. In other words, I will “extract” the eﬀect of structure by controlling for changes in competition that occurred in the areas where law was not enforced. For example, to test the eﬀect of the capture of Mr. Edgar Villareal, alias “La Barbie,” in triggering DTOs’ competition, I will control for changes in competition happening in non-Barbie territories before testing whether the capture of Barbie aﬀected violence within his territory.
4.2 Why have incentives for enforcement increased?
Goal: Understanding how large structural changes in Mexico’s illegal drugs industry are the root cause of increased law enforcement.
Research design: In this section I will provide a formal intuition to understand government’s prosecution eﬀorts. Prosecution value is determined by the value that the electorate gives to it. Thus prosecution increases when organized crime diversiﬁes into businesses that directly aﬀect population, such as extortion, kidnapping and domestic sells of drugs.
Following regulation theory, particularly the work of (Rowoski and Kayser 2002) based on Stigler (1971) Peltzman (1973), I will develop a formal model to fully explain my narrative. In my model, a politician decides whether to “collude” with traﬃckers and consent to corruption (not enforcing, accepting the cost) or “collide” and receive bribes (paying the cost of enforcement eﬀorts at the risk of running traﬃcker out of business and losing all future bribes). Its decision is based on the value that the electorate gain from enforcement, which is a function of how much traﬃckers aﬀect citizens.
Case study: I will show that Mr. Calderon’s enforcement eﬀorts have concentrated on regions where traﬃckers were diversifying their illegal business into endeavors that are more harmful to society, such as the sale of drugs domestically. In particular, I will present the cases of Michoacan, Chihuahua and Guerrero during December of 2006.
Section 5: Parallel research projects
5.1 What do international cases tell us about potential ways to reduce violence?
Goal: Comparative, descriptive studies showing diﬀerent policies that have resulted in decreases in organized-crime-related homicides/activites.
Research design: Study cases of Brazil, Colombia, USA, Russia and Australia. Descriptive statistics.
5.2 The role of local drug consumption in changing incentives for violence
Goal: Show the correlation between increases in the sale of drugs domestically and larger-scale drug-related violence.
Research design: I will present the ﬁrst estimates of drug-consumption in Mexico at the subnational level, from 1990 to 2010. Estimates come from overdoses and hospitalizations caused by illegal drug consumption. This unique data set will be used to understand the eﬀects of drug-related violence in the generation of DTO competition [Figure 11].
5.3 Who becomes a drug dealer and why?
Goal: Understand that the decision to become a traﬃcker depends more on their possibilities of getting a very large wage (i.e. length of right tail wage distribution) rather than in the average salary provided at the drug-traﬃcking industry -which are quite low and subject to large uncertainty.
Research design: It has been claimed that generating employment is the only way out of Mexico’s violence in the long term (Fregoso 2009). Employment, it is traditionally argued, will allow young Mexicans to avoid the criminal industry.
I disagree. Jobs are required but not all jobs will do the trick. We need to generate jobs that are attractive to those in danger to become traﬃckers, a subset of the population that psychological research has shown are more entrepreneurial, less dependable, and posses strongs leadership skills (Fairlie 2002, Fairlie and Woodruﬀ 2004). The jobs that are attractive for these people are those that provide a chance, even if minimal of becoming the boss, being independent and An older draft where this idea was ﬁrst explored is available here
Figure 10: Diﬀerent measures of local drug consumption in Mexico
have a large income. They do not care about the average wage they will get, they care about how far they can get if they become the best among their peers. They are sure they are. Jobs providing low/average salaries without chances of promotion will commonly be dismissed for jobs providing bad average salaries (even no salary at all) but with large changes of advance up.
In short, what matters is the length of the right tail distribution, independently of its thickness or of the mean of the whole distribution.
In this section, I present the formal model underlying my argument, combining two main areas of labor economics literature: (a) the classic approach of Becker (1968) in which individuals face single criminal vs. legitimate market wages and (b) the Roy (1951) model in which skills yield diﬀerent returns in the criminal vs. legitimate labor market sectors.
My model also generates an intuition for understanding why traﬃckers have become more violent than ever before. People in the ﬁeld claim that a new generation of traﬃckers have emerged in the last years, a generation that disdains old maﬁa laws and is more violent. I claim this is not a generational factor but the result of the new ways in which Mexican traﬃckers are recruiting their members. As competition in and among DTOs has increased, so has the demand for new DTO members. As a result, recruiters have started searching for their new employees in non-blood-related circles. This new labor pool is more violent-prone and entrepreneurial.
5.4 Why are politicians being killed now more than ever?11
Goal: Understand why recent years have seen the transformation of Mexican DTOs into highly violent political actors. Research design: since 2004, 27 Mexican mayors have been killed by organized crime, 15 just in 2010, a sharp contrasts with a history of complete absence of political violence. I present a formal model of corruption dynamics to understand these outcomes and show that, if mayors are getting kill is not because traﬃckers are more powerful than the new political class emerged after Mexico’s democratization -as traditional accounts of corruption would argue but because of political decentralization My model develops as follows. Let a corrupt market be one in which a briber (corruption demander) and a bribee (corruption supplier) agree on the price of a bribe. The bribe buys protection and monopolistic trade rights over the region controlled by the bribee. Bribers maximize proﬁts given by illegal revenues minus bribes. Bribees maximize utility given by bribes An older draft where this idea was ﬁrst explored is available here and honesty. If a potential bribee is not bribed, enforcement happens in the region with some probability which depends on the size of bribee’s honesty.
Bribers may kill the bribee if he cheats (either by enforcing or allowing another briber to proﬁt in the region).
First, a single bribee and a single briber had agreed on the price of a bribe. Yet unexpectedly another bribee gets into the game. Protection is now oﬀered by two diﬀerent bribees and both are partial complements. The traﬃcker may bribe one bribee only but traﬃckers’ proﬁts are much higher if they bribe both bribees.
There are two ways in which both bribees may coordinate to set the price.
• With centralization. One bribee sets the price of the bribe to maximize total bribe revenue and gives the second bribee a share of total revenue. Incentives are set such that second bribee has no motivation to charge a bribe on top of the one that the ﬁrst bribee is already collecting.
In practical terms, bribing one bribee assures compliance of both bribees.
• With decentralization. Both bribees set the price of bribes simultaneously. They play diﬀerentiated-goods Bertrand. Unlike in (a), bribing one bribee does not assure compliance of both bribees. If only one bribee is being bribed, there is a probability of second bribee to enforce in the region.
Note that under the decentralized agreement, a bribee may cheat on an agreement by asking the second bribee to enforce in the region. The ﬁrst bribee would get both the bribe, and the payoﬀ of honesty (because traﬃckers are going to be prosecuted).
Case study: Explanations of each of the mayors that have been killed.