Why are Mexican mayors getting killed by traffickers? The dynamics ...

Why are Mexican mayors getting killed by traffickers?
The dynamics of corruption in illegal-drug markets
Viridiana Rios
Harvard University Department of Government ∗
November 4, 2010
Abstract
This is a paper about how organized crime and politicians interact, and the way in
which political institutions, and market dynamics impact this relationship. Recent years
have seen the transformation of Mexican drug traffic organizations (DTOs) into highly
violent political actors. Just in the first 9 months of 2010, 12 mayors have been killed
by organized crime, a sharp contrast 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 more is not because traffickers are more powerful -as
traditional accounts of corruption would argue– but because of (a) political decentralization,
and (b) changes in the structure of the illegal-drug industry (i.e. higher competition,
and a recently open domestic drug market). My conclusions are drawn from an extensive
fieldwork in Mexico’s drug war zone.
∗ 1737 Cambridge St; Cambridge, MA 02138; vrios@fas.harvard.edu; Telephone: (617) 997 2471
1

‘Well, first I pray to God every day and at every moment. Then, I do not denounce,
I do not see, I do no listen besides, I do not even have the [constitutional] mandate
to prosecute drug traffickers! Finally, at my home, I talk to God again 1 ”
Recent years have seen the mutation of Mexican organized crime into a politically violent
actor. In 2010 elections, Mexican drug traffickers made clear that organized crime was not going
to be a simple political spectator anymore. They attempted to change political preferences
by prohibiting the press to cover stories or by openly funding their favorite candidates; they
tried to reduce electoral turn out via blatant intimidation; they succeeding selecting candidates
either by assassinating the opposition, or by terrifying politicians until they withdraw from the
competition. If the previous actions were not effective, they just killed politicians. Political
violence became the norm. In 2010, 26 mayors were assassinated, at least 126 more were death
threatened, 8 journalists were killed or disappeared, and 411 policemen were murdered. Some
political parties were unable to compete in local elections because of “a lack of braveness (El
Universal 2010)” of the potential candidates. Complete towns remained without de facto authorities,
as mayors and policemen frequently ran out of office, terrified.
Yet, just recently Mexican drug traffic organizations (DTOs) were commonly referred as the
counter-case of political Colombian DTOs. Mexico’s organized crime was notoriously disinterested
in politics (Resa Nesateres, 2006). Under this traditional narrative, Mexican kingpins
were portrayed as unsophisticated but clever dealers who knew what was good for business
(Bunker and Sullivan 2001). They did not wanted people fighting in the streets, neither did
they wanted to take visible positions in politics; they were there for the money. Mexicans were
not politically involved criminals, they were businessman.
In this paper, I provide an explanation for the abrupt change in the relationship between
criminals and politicians. This is a paper about how organized crime and politicians interact.
I challenge the traditional “silver-or-lead” approach to understand the political/criminal nexus
–according to which as traffickers become more powerful, they also become more politically
visible, and they increasingly force politicians to decide between accepting bribes (silver), and
being assassinated (lead). I show that criminals’ political violence does not necessarily come
from an increase in the power of traffickers versus politicians. My story is able to explain why
Mexico’s corruption outcomes have changed even if, as I show, traffickers have actually become
weaker.
The standard Silver-or-bribe narrative fails to comprehend corruption outcomes because it
does not disentangle the factors that determine corruption supply versus those that determine
corruption demand. Corruption mechanics, I argue, are a bargaining game between a corruption
supplier (i.e. politicians willing/able to be corrupt) and a corruption demander (i.e. organized
1 Mayor of Pueblo Nuevo, Durango, when asked what he do to keep himself alive (Abrego 2010)
2

crime willing/able to corrupt) in which supply is largely determined by institutional factors,
and demand is determined by the industrial organization of illegal-drug markets. My main
argument is that the mechanics of corruption can only be understood if we analyze government
institutions (i.e. degree of centralization), and illegal-drug markets (i.e. degree of competition
and type of targeted market). Institutions and Illegal-drug markets determine corruption supply
and demand, and explain why corruption outcomes have changed.
The paper is divided in six main sections. I begin by briefly explaining why we would ever
expect to have corruption markets; this is to say, why traffickers would ever care about bargaining
with politicians instead of just disintegrating the state, or becoming the state themselves. A
second section presents a simple bargaining model of corruption dynamics, and argues in favor of
analyze corruption demand and corruption supply separately. A third section presents a model
which departs from previous literature (i.e. Silvel-or-lead approaches). I analyze the strategies
of corruption demanders with particular attention to the complexities of killing (i.e. violence is
costly, uncertain, and has a signaling value). I follow by analyzing the strategies available for
corruption suppliers noting that a weak state is not always the preferred outcome for bribers.
A fourth presents some comparative statics to show that sometimes bribing more is less, not
more expensive for traffickers; that sometimes politicians want small bribes; that even weak
traffickers can extort powerful states, and that politicians not always take a decision between
being bribed or killed. Killing can complement bribing, and bribes/killing may be completely
absent. A final interesting result refers to the quality of politicians. In a corrupt environment,
all available politicians will be corrupt unless the state has the ability to extort criminals. To
be completely virtuous, a politician needs to be capable of being completely corrupt. A fifth
section uses the previous model to explain why corruption outcomes in Mexico have changed.
I show that the recent emergence of political violence in Mexico can only be understood by
analyzing changes in the determinants of (a) corruption supply (i.e. political institutions), and
(b) corruption demand (i.e. illegal-markets structure). Finally, a sixth section concludes.
1 The utility of the political-criminal nexus
The first question to make when trying to understand why Mexican traffickers have become
more politically violent over time is why, in the first place, would traffickers ever prefer to make
deals with politicians, instead of just being violent, taking over the state and running it themselves.
Relating to politicians is costly and uncertain. Bribing requires the establishment of a
millionaire secret pact between two parts –two criminal parts of obviously dubious principles–
without any legal institution to appeal in case of default. Furthermore, the terms of the contract
are generally agreed verbally which allows for misinterpretations and cheating. It is not trivial
thus, to analyze when and why would an DTO ever prefer to bargain with politicians. Why
would traffickers ever demand corruption?
3

Current academic literature understands the relationship between criminals and government
as an evolutionary phenomenon. Criminals have as one of their main interests to control the
state in order to facilitate the conduction of their illegal business. Yet, in its first stages, when
crime is less pervasive, profitable and organized, controlling the state is impossible for DTOs.
It is only when DTOs “evolve” and become more powerful, that they are able to impose their
will into a relatively weaker state (Baylor and Godson 2001, pg 14-15). At the peak of their
evolution, criminal organizations become parallel states (Bunker and Sullivan 2010) and supply
services –such as protection– that the state is not able to provide. In short, literature argues
that traffickers demand corruption only when they are not powerful enough as to become the
government themselves.
This understanding of the political/criminal nexus however has failed to take in consideration
many other reasons why, I argue, even a highly powerful DTO may prefer to have a state
(rather than being the state themselves). A state benefits traffickers as it ensures larger and
cheaper trade.
A legitimate state is useful for traffickers as it (a) provides a platform of legal trade that
is used by traffickers to smuggle their products, and (b) a series of public goods and positive
externalities, such as a structure of law enforcement and infrastructure that traffickers can use
for their own advantage. A state is required for the survival of any type of business, illegal ones
included. Ironically, even criminals need politicians.
First, having a legitimate state secures an stable amount of commercial transactions and
trade. Legal trade facilitates illegal trade because a large share of illegal products cross borders
camouflaged as legal cargo. Actually, increments in US-Mexico trade during the late nineties
as a result of the North American Free Trade Agreement (NAFTA) proved to be a blessing
for Mexican DTOs. If Mexico became the main illegal-drug exporter into the US is precisely
because the country became an increasingly important commercial partner for the US 2 . Drug
lords love trade, and traders love governments. The circle is closed.
Furthermore, as trade increases, the burden of inspection increases exponentially. About 40
percent of all trade between Mexico and the US crosses via Nuevo Laredo, a small city at the
border of Texas (INEGI 2005), that is thousands of trucks per day. Mexican migration officers
2 During the nineties, Mexican DTOs were well aware of the value that Mexico would acquire if NAFTA was
signed. They saw the trade bonanza approaching. As soon as the trade agreement was signed in 1994, some of
them hired migration experts to advice them on which were the best type of cargo for introducing their products
into the US (Blancornelas 2001). Advisors were necessary because for health and other considerations, some
cargos were going to be (and currently are) inspected less carefully than others (e.g. perishable goods cross the
border faster than say, electronics). Knowing the probability of inspection in advance was a de facto reduction
in transportation costs for DTOs.
4

have neither the capability, nor the resources, to inspect all cargos. Inspecting all of them will
slow down trade and result in millionaire losses. As Mr. (XX- confidential), director of customs
at Laredo told me: “sometimes US inspectors seize drugs in cargos that we had just let in. We
are not doing this on purpose, as they think, it is just that we can only inspect some of the
cargos. We cannot do more than that. 3 ”
Axiom 1a: Illegal profits (M) increase in legal trade
Note that the expected profitability of illegal business is a necessary but not sufficient condition
for corruption markets to emerge. Very profitable territories, with a healthy and prosperous
legal trade, may not share the institutional and structural factors required for a corruption demand/supply
to exist. Most South American countries could be fertile territory for traffickers
and yet, DTOs have largely been confined to a bunch of countries (Colombia, Mexico, Peru and
Bolivia). It is only in these countries where institutions allow politicians to supply corruption,
and traffickers find politicians useful enough as to demand their participation.
Axiom 1b: M > 0 is a necessary but not sufficient condition for trafficking to exist.
A second reason why traffickers prefer to have a state is that a state provides public goods
–such as highways, airports, ports– that traders (both legal and illegal) need to perform their
operations. On the absence of a government, traffickers would have to take all of these costly
operations in their hands. Peru is an example. When in the early nineties traffickers started
using Peru as a drug production paradise, DTOs had to invest large amounts of money into
developing transportation systems and urban infrastructure (Lupsha 1995). That was certainly
not a good deal but nothing could be done given the de facto absence of Peruvian state.
That even very strong drug trafficking organizations need a state is not a secret; for a proof,
Haiti. When increments in law enforcement reduced the ability of Colombian DTOs to do
business, the most immediate and logical action to take was to reallocate operations in the
Caribbean. Back then, the most important drug corridor took drugs from Colombia, transporting
them via maritime, to introduce them into the US using Miami. Haiti, with a premier
location in the Caribbean, was the natural choice. Traffickers though, soon realized that Haiti
was not going to work. Haiti’s state was too weak, chaotic and uncertain, even for illegal business.
Government instability made bribe agreements to be short-lived, and required periodic
and burdensome renegotiations between DTOs and authorities. Traffickers quickly moved their
operations to Mexico and Peru. Traffickers as businessmen, value binding contracts and strong
3 Mexico has the capacity to inspect only 10% of all cargos that cross into the US. The numbers for traffickers
could not be better. Given the probability of inspection, seizure is just internalized as a cost of production. The
price of cocaine at least doubles at the moment that it leaves Mexico and touches American territory (RAND
2010). Even if Mexican DTOs smuggled drugs without any strategic consideration -say, by placing drugs in
every single cargo crossing into the US–, traffickers would still make significant profits.
5

states which can central-enforce and provide public services.
In short, because legal trade increases illegal profits, even very powerful DTOs may find it
harmful to dismantel the state and/or to substitute it by a parallel criminal state. Actually,
the best scenario for DTOs is one in which they do not have to take the cumbersome and
overly public task of ruling a country themselves. Traffickers prefer to outsource politics to a
sufficiently strong and centralized government that will ensure the survival of illegal businesses.
The paradox of the political/criminal nexus is a simple but powerful one: having a state reduces
operational costs for drug traffickers (e.g. provides infrastructure, and increase legal trade) but
also weakens traffickers because dealing with the state may entail non-trivial monetary costs
(i.e. bribes).
Lemma 1: Organized crime (i.e. a group of drug traffickers (N)) prefers to have a legitimate
state (i.e. a group of politicians (A)) than to become the state, independently its relative
strength (P n ) versus that of the state (P a ).
Given Lemma 1, a question remains as to which would be the mechanism used by criminals
and politicians to interact. In the following section, I introduce the traditional way in which the
relationship between politicians and traffickers has been understood by the literature (“Silveror-lead”
argument).
2 Corruption mechanics: a bargain game between criminals
and politics
The political/criminal nexus between organized crime and politicians has so far being understood
as a very straightforward exercise described by the “silver-or-lead” dichotomy. Those
politicians ruling in areas of interests for traffickers have to decide between accepting bribes
(silver) and getting killed (lead). According to this narrative, bribes are determined by how
costly corruption is for politicians, as well as by the relative power of traffickers versus the state
(Bo etal 2006). The price of a bribe thus is the outcome of a rational calculation.
The price of bribes is determined by the relative power of the state versus traffickers. As
a government becomes more powerful (i.e. able to enforce law), its ability to extract traffickers’
profits via bribes increases. On one side of the spectrum, a powerful government extorts
traffickers keeping most of the illegal profits for itself. Such would be the case of Montesinos at
Peru, and the Communist Party at China. On the other side, traffickers retain almost all profits
for themselves with small –or non-existent– bribes for authorities. Such would be the world of
trafficker organizations that have managed to work in a virtual autonomy from the state, such
6

as FARC.
Definition 1: Corruption is a bargaining game in which N and A agree on the price of a
bribe (b).
Fixing the price of a bribe is not a trivial exercise. As an analysis of corruption dynamics in
Mexico shows, several mechanisms may be used to determine the price of bribes. Back in the
late eighties, the leadership of the notoriously corrupt Federal Direction of Security (Direccion
Federal de Seguridad, DFS) –the main institution for prosecuting DTO’s members– used to
sell positions inside the institution using auctions. The positions located in the most valuable
points for DTOs were priced the highest because such positions gave traffickers less bargaining
power to set the price of the bribe. As expected, the price of bribes was determined (as it still
is) by how much politicians could hurt, or help DTOs.
Bribes thus have a large variance. In perhaps one of the most scandalous cases of corruption
in Mexico, it was discovered that Juarez DTO operated in Cancun under the protection of the
governor, Mario Villanueva. The governor was arrested in 2001, and extradited to the US in
2007 after he acknowledged accepting millions of dollars in kickbacks and bribes (Morris 2010).
In 2007, another corruption scandal detonated in Cancun. This time the director of Cancun
airport was accused of accepting a BMW (plus 45 thousand dollars) per each flight loaded with
cocaine that was allowed to land (Gomez 2010). The price of bribes followed a clear trend:
Governors got million of dollars, airport directors only BMWs.
Axiom 2: b decreases in P n and increases in P a .
In this section, I defined the mechanics of drug trafficking corruption as a bargaining game
in which politicians and criminals agree on the price of a bribe. Until now, my definition follows
the traditional Silver-and-Lead narrative. The final price of the bribe is determined by the
relative power of the politicians and traffickers and is a non-trivial bargaining exercise. The two
following sections depart from the Silver-or-lead narrative by making an explicit differentiation
between corruption demand and supply. I draw upon some basic, logic axioms to present an
enriched understanding of the political/criminal nexus.
3 Corruption demand and supply
The mechanics of corruption can be better grasped by analyzing first, the strategies available
for corruption demanders (bribers, traffickers), and then those available for corruption suppliers
(bribees, politicians).
7

3.1 Demand side: Bribing, threatening or killing
Definition 2: N decides between paying b, and exerting political violence (k) (i.e. killing A)
The traditional Silver-and-Lead narrative understands killing as a substitute for bribing, and
predicts violence would be more common as traffickers become relatively more powerful than the
state. When traffickers are powerful enough as to posses a system capable of killing politicians,
and state is weak enough as to be unable to enforce law against assassins, costly bribes will be
substituted by killings. Political violence thus is a function of criminals’ power. As the relative
power of the state versus DTOs diminishes, organized crime is supposed to become more violent
(Godson 2003; Baylor and Pimentel (2001)) not only because (a) DTOs have the technology
to kill politicians but because (b) a weak state cannot stop crime organizations from fighting
among themselves. Under this logic, it is the state the one who can keep violence controlled;
an “unleashed” crime is always bloody.
In Mexico, the standard story argues, the arrival of opposition to power undermined the
ability of the government to keep corruption agreements between the hegemonic ruling party
(Institutional Revolutionary Party, PRI Partido Revolucionario Institucional) and criminal organizations.
When the government of Vicente Fox, an opposition leader from the National
Action Party (PAN Partido Accion Nacional) took office in 2000, it lacked of the expertise,
power and/or networks to force criminals to conduct their business in peace. By then, the PRI
had held power for more than seventy years, period at which it kept a pax mafiosa by extorting
criminal organizations (Astorga 2006, Mercurio 2006). When Mexican DTOs saw themselves
free from PRI supervision, violence exploited (Ravelo 2001, Gomez 1999).
Preferences for violence though, I argue, are much more subtle than the traditional narrative
argues. Violence is no free lunch. Three things have to be considered: (a) Killing is not always
cheaper than bribing, even for powerful traffickers; (b) criminals do not always find it attractive
to fight violently between each other; and (c) criminals cannot commit to kill with complete
certain, they can only commit to try (very hard) to kill.
Indeed, the balance of power between the state and traffickers matters for violence. Killing
is more common when such action is expected to cause less negative consequences. As Alfredo
Corchado, chief of the Mexican Buro from the Dallas Morning, and one of the most important
reporters of drug trafficking activities in Mexico, told me when I asked him what measures he
took to reduce his chances to be assassinated: “I make sure my batch from the Dallas is very
visible, they would not dare to kill a gringo, it would be too dangerous.” Being a first generation
Mexican American though, he joked: “the good news is that I am a gringo reporter, the bad
one is that nobody could tell (Corchado 2009).” Alfredo is right. Traffickers have not dare
to kill a foreign reporter, but have killed 11 Mexican reporters, and disappeared 7 more since
2007. They also target more local media than higher-profile national media. Only 4 out of the
8

18 cases of murder/disappearance were journalist working for the national media. “Freedom of
press can only be secured at the federal level; we local reporters do not have the international
exposure as to affect [traffickers]. We just listen to them [traffickers] and comply. Who care
about us? We are too little” accurately said Mr. (XX -confidential) director of a local in one of
the most important cities located at the US-Mexico border.
However, even powerful criminal organizations may find political assassinations unattractive.
First, killing politicians is costly negotiations are costly themselves. After an assassination, all
of the previous agreements, understandings and networks done with the defunct politician have
to be redone with the new authority. In an illegal business –full of secrecy and where money
laundering is inherently difficult– getting the cash and daring to expose in public to deal directly
with politicians (i.e. sending messages, visit them at their offices, or at their houses) is
not trivial. Making an agreement may also require several meetings over time, each one delaying
a bit more the continuation of business as usual.
Second, general violence between criminal organizations is also costly because the process of
employee substitution is problematic. At the base of the organization, defunct petty traffickers
are substituted for new ones rapidly, but these ones normally lack of the expertise to keep running
the business as usual. At the top of the organization, substituting defunct elite traffickers
is even more Challenging. Succession is never clear-cut established and normally causes intense
internal fights. This is well exemplified at current Mexico’s drug war. The assassination of
important drug lords, such as Beltran Leyva in 2009, and the capture of others, such as La
Barbie in 2010 (I consider capture a de facto equivalent to death), have generated important
spikes in violence because their successors have fought between each other.
Axiom 3: Negotiating b is costly.
One final reason why killing is costly for traffickers is that it is more uncertain than bribing.
A trafficker can commit to bribe. If he offers a bribe in exchange for a service, the bribe will
be paid unless the trafficker himself decides not to. The trafficker controls bribes. A trafficker
though cannot commit to kill. Even if the trafficker decides that killing is the best strategy to
follow, killing is difficult. It requires planning and it is always subject to in-the-field mistakes.
Traffickers cannot entirely control who dies and lives. They can only try (very hard) to kill but
assassinations can always get it wrong.
Actually, evidence demonstrates that mistakes in assassinations ordered by Mexican traffickers
are quite common. Perhaps, the most important one happened in 1993 when the Archbishop
Juan J. Posadas was assassinated by the Tijuana DTO. He was mistakenly identified as a drug
lord from the competing Sinaloa DTO. His assassination brought an important increase in
federal enforcement. The Mexican government set a series of military raids in which several
high-profile traffickers were assassinated and imprisonment.
9

Axiom 4: k is affected by random factors (r).
In short, we can find scenarios in which bribing may be preferred to killing, even by powerful
organizations. Bribing is less disruptive and more certain. Actually, evidence shows that in
states completely controlled by drug traffickers, such as Colombia during the seventies (Ledebur
2005) or some regions of Brazil, political assassinations are limited. In Rondonia, a Brazilian
region completely controlled by drug lords during the late eighties, political assasinatios were
not even considered because to begin with, “no public figure [could ever] hope to get elected []
without the explicit or implicit consent of traffickers (Geffray 2002)”.
Most of the episodes of violence between Mexican drug traffic organizations have been ended
by traffickers themselves. Keeping violence low has been a constant worry of Mexican DTOs.
We can count at least 5 summits in which elite cartel members got together to discuss mechanisms
and rules to avoid violent competition.
Lemma 2: k does not necessarily increase in P n
This is not to say though, that violence will be avoided always. Selective and controlled
violence is beneficial for crime as it opens the possibility of using threats instead of bribes.
Threatening is the cheapest strategy available for criminals (it is the equivalent of a zero-cost
bribe) but to be used requires reputation.
Reputation matters because of the interaction between politician and bribers are subject to
information/commitment problems. Politicians normally ignore the actual ability of traffickers
to exert violence, and criminals cannot commit to kill even if such is their preferred strategy.
Selective killing is beneficial in the long term as it provides the required reputation for threatening
in the future. Only criminal organizations which have established a reputation of compliers
will deliver credible threats.
Politicians built beliefs about traffickers’ power based on information collected by intelligence
agencies and observable actions of criminals. A trafficker is non-violent either because he
is unable to assassin, or because he does not want to. Politicians never know. Criminals can affect
politicians believes by killing. Violence sends an unequivocal sign that they are not kidding,
that traffickers are actually capable of killing. Violence thus is useful because it increases the
leverage of criminals in the bargaining process, deters potential competitors, and discourages
current corrupted officials to cheat on the agreements. Paradoxically, present violence is needed
to reduce the need to exert future violence.
Lemma 3:Because k builds reputation (π), k reduces b
10

Note that because killing is a sign, direct violence against high-level politicians can be substituted
with slightly less politically visible killings of similar difficulty level. Actually, threats
followed by intimidating actions are common in Mexico’s political/criminal relationships. When
in February 2009, “La Linea” -a DTO operating in Juarez- decided that it was in its best interest
to take Roberto Orduna, chief of the Ministry of Public Security (Secretaria de Seguridad
Publica SSP), out of their way, the DTO did not had to kill him to accomplish its goal. A
credible threat with effective side killings was sufficiently efficient. La Linea started by killing
a middle-level SSP agent, as well as three of Mr. Orduna’s bodyguards. A day later, the
traffickers posted a billboard, directed to Mr. Orduna which read “if [you] do not resign, we
will kill one agent every 48 hours.” Two days later, after a policemen and a guard had been
executed, Mr. Odonez decided to leave office “not because [Mexico] is losing the war [on drugs]
but precisely because they wanted to win it (Davila unknown year)”.
When the mayor of Juarez accepted Mr. Ordonez resign, he was immediately congratulated
by La Linea for his “good” decision to “allow tainted Orduna to leave office.” La linea also left
him very clear that “if you keep assisting more queers [sic], and helping those who you know,
we won’t ask you to resign, we will cut your head and the heads of all your family, even if they
live in El Paso, Texas (Davila unknown year)”
La linea does not need to kill high-profile politicians to make them compliers because this
DTO has heavily invested in its reputation. Perhaps, one of the most exemplary displays of
their power happened again in Ciudad Juarez. When at the beginning of 2008 the predecessor
of Mr. Odonez, Mr. Prieto was in office, La Linea posted a billboard (this time at the policemen
memorial, along with funeral flowers) in which the DTO listed 17 authorities who would be executed
to instruct “those who still do not believe [in La Linea’s threats] (Topete 2008)”. One to
one the authorities in the list were assassinated. By April, eight had been already killed, about
one per week. When in may 2008, the second on board at SSP -and the closest collaborator of
Mr. Prieto- was assassinated, the resign of Mr. Prieto arrived the office of the mayor (Gomez
2008). La Linea won. Traffickers had “killed” Mr. Prieto by killing lower-key political figures.
To summarize, this section discussed the strategies available for corruption demander. The
demander can choose between two strategies: bribing or killing. Killing is not always picked,
even if the demander is powerful, because it is costly and uncertain. Selective killing though
may be beneficial as it can be used to build a reputation. Reputation opens the possibility of
a third strategy: threatening. Threatening is the preferred strategy of demanders because is
equivalent to a zero-cost bribe. The following section analyzes the second side of corruption
mechanics: corruption supply.
11

3.2 Corruption supply: Prosecuting, ignoring or protecting
Definition 3: A decides between rejecting b and prosecute N (e ≥ 0), and accepting b and
protect N (e ≤ 0).
Politicians decide how much they will enforce law given their capacity constraints. Law can
be enforced against traffickers (i.e. prosecute) or in favor of them (i.e. protect). Politicians
can also decide not to enforce (i.e. ignore). No enforcement means traffickers won’t be actively
hunted; a de facto equivalent to decriminalization. Not enforcing is less compromising than any
other option but gives fewer profits for politicians.
The capacity to enforce of the state is a double-edge knife for traffickers. A strong government
is an exceptional criminal protector but also, a threatening institution. A certain level of
government strength is beneficial for traffickers because it allows corruption to be efficient by
securing that bribes will indeed buy what they are meant to buy: impunity.
A very weak state may be entirely willing to contribute to traffickers’ wellness, but may be
de facto unable to deliver what traffickers require. Securing say, an oligopolistic market for a
DTO by prohibiting other organizations to sell or cross drugs in an area, requires at the very
least a government able to crush potential competitors and identify which members belong to
each DTO. Tijuana local police, for example, famous during the 90s for protecting the Arellano
Felix DTO, regularly beat up local gang members who were trying to become small drug traders
in Mr. Arellano territory. Policemen were effectively protecting the oligopolistic benefits that
Mr. Arellano’s bribes had paid for (Proceso 1995). If Tijuana government would have been
unable to track and hunt gang members, Mr. Arellano would have found it completely useless
to bribe them and would have to take care of the cumbersome task of protecting his territory
himself.
The ability of the state to enforce along all its territory and levels is also beneficial for organized
crime; such state keeps the price of bribing under control and ensures that agreements will
bind. Bribing such state may indeed be expensive but it will never be as expensive as bribing a
fractionalized state. When bribing one fraction of the state does not ensure that the agreement
will bind in all the territory (as would be the case during a civil war), more bribes are needed 4 .
In short, a strong state with ability to enforce and vertical control over its bureaucracy will
ensure that traffickers bribe only one actor, keeping the price of bribes under certain affordable
limits. If instead, several agents have legitimacy over one single territory, drug trafficking organization
may need to spend more money to get the impunity and monopoly that they need
4 Note that keeping bargaining power fixed, more bribes do not necessarily mean fewer profits for traffickers
(e.g. a same bribe may be divided among two bribees) but it does mean that trafficker’s profits will not be more
than with a single actor.
12

for their business. The simplicity of closing a deal with a single actor is always preferred to the
chaotic dynamic of multi-agent bribing processes.
Lemma 4: N has concave preferences over P a
To summarize, this section showed the strategies available for corruption suppliers. They
can prosecute, protect or ignore criminals. Both prosecution and protection require a strong
state which means that, traffickers may not always want to have a weak state. The following
section presents some comparative statics and shows interesting results of the previously defined
model of corruption.
4 Equilibrium results: How much silver, how little lead?
Definition 4: A corruption interaction finishes when A accepts a bribe, and a level of political
violence is exerted.
Corruption is a dynamic, non-zero cost negotiation between a briber (trafficker) and a bribee
(politician) which has as observable outcomes either (a) impunity towards traffickers with or
without political violence (corruption agreement), (b) active prosecution of traffickers with political
violence (if no agreement), or (c) no trafficking.
The size of the bribe, as a share of total illegal profits, will be determined by politicians’
costs of corruption (given by relative power of the state versus traffickers, the value politician
attributes to virtuous behavior, and the probability of being prosecuted), and traffickers’ benefits
of corruption (given by the profitability of the business, their relative power of the state
versus traffickers, and the costs of killing). Politician’s power refers to their ability to prosecute
criminals; trafficker’s power refers to their ability to exert violence against politicians.
The relationship between the price of bribes and the capacity of enforcement of the state
is pictured in Graph 1. The normal line pictures all equilibriums in which bribes are priced to
pay exactly the benefits of no-enforcement; I call this the “fair trade bribe”. The actual bribe
paid can be above or below fair trade, depending on the value that politician give to virtue, the
probability of politicians to be prosecuted, and the cost of killing. The cost of killing depends
on expected state reaction (i.e. how much retaliation will increase after the killing, and the
value that current corruption agreements with the politician (if such are in place).
Several cases could be portrayed. The dashed line represents the bribe paid to politicians
who are very well connected into the political arena or in very politically relevant positions.
Killing well connected, relevant politicians is costly. How crucial buying those politicians is for
13

profits’ levels determines the baseline price of the bribe, that is to say, the origin of the fairtrade-bribe
line. The bold-dashed line represents a well-connected, politically relevant politician
in a non-crucial position. The attenuated-dashed line represents a non-connected, politically
irrelevant politician in a crucial position.
[Graph 1 about here]
The cost of corruption for politicians is larger as bribes increase in size because hiding a
large amount of illegal profits is increasingly difficult. This explains why some politicians get
moderate bribes even if they could get much more. Mr. (XX -closed source), high-level authority
at Tijuana’s office of the Attorney General (Procuraduria General de Justicia del Estado
de Baja California, PGJE) protected Arellano Felix DTO and yet, had a very modest car and
lived in a very small apartment. He was using his moderately priced bribes only to take their
daughters to private universities (XX -closed source). For politicians thus, the marginal costs
of corruption are increasing which means their fair trade bribe is a convex function.
Contrary, the cost of corruption for traffickers is smaller as bribes increase. Bribing one
politician reduces the cost of bribing others because it shows the last ones that the probability
of prosecution is low. If others do it, it should not be that dangerous. Corruption is contagious.
Furthermore, corrupt authorities may pressure new authorities to be corrupt. As (XX
-closed source), ex-member of Tijuana police acknowledge, “it is impossible not to [be corrupt],
if you do not get [bribes] your boss will be pissed off because he won’t get his share of it.” In
very a very well structured corruption market, as is Tijuana local police department, bribes
are commonly distributed along the hierarchy by chunks. Each member gets a share of what
their subordinate got, and gives a share of his profits to his supervisor. Supervisors ask for a
“required quota” to each subordinate. As long as the quote is fulfilled, subordinates can keep
the rest for themselves. Failures to get the quota are punished. Non-corrupt officials thus may
become liabilities in corrupted institutions. Because for bribers the marginal costs of corruption
is decreasing, the fair trade bribe for traffickers is a concave function.
Result 1: The concavity of fair trade bribes (b ′ ) is undetermined; “Sometimes bribing more
is less, not more expensive, and sometimes politicians want small bribes”
Uncertainty plays an important role when determining whether actual bribes will be above
or below fair trade bribes. If the briber is paying a higher bribe than the benefits he gets from
no enforcement, is because the bribee is more powerful than the briber (i.e. the state extorts
traffickers and takes part of their profits). The opposite is not true. We can imagine a situation
in which the briber is paying less than the benefits he gets from no enforcement (i.e. traffickers
are extorting the state) without the briber being necessarily more powerful than the bribee.
Because of information issues (Axiom 5), and the randomness of violence (Axiom 4), politicians
are never certain about the real power of traffickers, and traffickers may get lucky in a political
14

assassination. Given this, it may be that politicians miscalculate their power (versus that of
traffickers) and “under-price” their value.
Result 2: If b < b ′ , P n has not to be larger than P a ; “Even weak traffickers can extort
powerful states”
Other interesting results come from analyzing the relationship between killing and bribes
(see Graph 2). Violence relates to bribes in two ways: it substitutes them when these are too
high, and it reduces them by signaling traffickers’ power. Killing is a substitute for bribing,
when bribing politicians is too expensive (because for example, they put al lot of value into
being honest). Yet, when bribes are very high, traffickers also find it more difficult to kill. As
bribe increases thus, violence is more wanted but more difficult to attain for traffickers
[Graph 2 about here]
Note that it is the relationship between capability to kill and willingness to kill what determines
whether violence would be positively or negatively correlated to bribe size. The traditional
silver-or-lead approach fails to understand corruption because it always assumes a relationship
in the form of the dashed line. The relationship, I argue, can have the form pictured with the
bold line because violence has a random factor, and because quietness has an intrinsic value.
The bold line describes the behavior of a not very powerful organized crime. They will kill
less (not more) when bribes are high because they are incapable to kill highly-bribed, powerful
politicians. They may want to kill politicians more when bribing is too expensive, but they do
not do it because it is costly. If any killing at all, they will kill the underdogs. The key here is
that the reason why they kill underdogs is because such killings signal the top guys, and reduce
their demands. This proves result 3:
Result 3: Correlation between b and k must not be assumed to be negative; “Silver or
lead, Silver and lead”
Graph 2 also shows that politicians will always receive a share of illegal profits unless they
are scared enough as to not ask for money. Scaring politicians requires credible threatening.
Note that there is always a point at which killing more is not valuable. When politicians are
intimidated enough as to accept reducing enforcement without bribes, killing only remains there
as a residual; as a reminder for politicians of the power of traffickers. Thus, if traffic is done
and bribes are zero, it is because selective killing is working as a mechanism to keep politicians
intimidated.
Result 4: When M > 0, k > 0 is required for b = 0;“No-silver and lead”
To understand why result 4 holds, consider an alternative scenario in which a politician does
15

not receive illegal profits (b=0), and there are no killings (k=0). This would only be the case if
he places a lot of value into virtue and has the force to control traffickers’ violence. Under this
circumstance, there will be not illegal profits because trafficking would be controlled. It won’t
be worth for traffickers to go into illegal business because the state is too powerful and virtuous.
Result 5: When M > 0, k = 0 if and only if b = M is a viable equilibrium;“Silver and
no-lead”
If organized crime is in business, political violence will only be absent (k = 0) if the corrupt
state is powerful enough as to take all of the illegal profits for itself. Recall that traffickers always
have an incentive to kill either to (a) reduce the price of bribes, or (b) keep their reputation.
Killing thus will only be zero in the very extreme case that traffickers are unable to kill. This
is not easy given than even weak traffickers may get lucky at times. Only a very powerful state
will be able to have a peaceful drug market yet, this requires the state to be strong enough as
to extort criminals. Such strong state can decide between eradicating crime at all, or take all
illegal profits for itself. This is the case of a potentially extortionist state (i.e. b=M is a viable
equilibrium)
Results 4 and 5 have important implications for the quality of politicians. When running for
office, politicians know that if trafficking is in place is because either there has been a corruption
agreement (b > 0), or because politicians are too scared to (fully) enforce against crime (result
4).
Result 6: When M > 0, only corrupt A’s will take office, unless b = M is a viable equilibrium;
“To be completely virtuous, one need to be able to be completely corrupt”
Those politicians who take office at corrupted places know they will be bribed because otherwise
they will be killed. Because of result 6, we know that the only chance for political violence
to be absent, given that traffic is present, is to have a very powerful state. Rational politicians
deny running for office if they have a positive probability to be killed 5 . Weak politicians who
decide to enforce have a positive probability to be killed. Thus, we will only have non-corrupt
candidates if the incumbent is powerful enough as to be completely corrupt. Quiet drug markets,
like the US, require politicians to be potential extortionists 6 .
5 The rationality of politicians may be debatable as it may be affected by the value that they put into virtue.
A very virtuous politician could be argued, will take office even if it is to be a martyr. I do not discard this
possibility but find it improbable given that, if the ultimate motivation of a virtuous politician is to reduce
crime, his assassinated will not put him any closer to achieve his goal. Nobody can change the world from the
grave. Martyrs though may still give their live for a uncertain, further ulterior purpose. A politician may think
that even if killed, his assassination will generate international attention and a increase in retaliation against
crime.
6 Enforcement by a weak politician is only possible if such enforcement does not really hurt traffickers, at
16

To summarize, the current section has analyzed corruption mechanics and the outcomes
that the political-criminal nexus is expected to generate. I defined a simple model of corruption
bargaining, and used a small set of logically derived axioms to show that corruption is far more
than a decision between accepting a reasonable bribe (silver) and getting killed (lead). Silver
may actually come hand with hand with lead, there may be no offer of silver at all, there may
be no chances of lead, politicians may want to have fewer silver than available, and traffickers
may want politicians to take their silver. Furthermore, I presented and logically proved the
validity of some apparently counter-intuitive statements. I showed that to be non-corrupt one
has to be able to be corrupt, that traffickers want politicians to be around, and that power is
not a necessary condition for criminals to be able to extort politicians. In the following section
I present a study case to asses the validity of the results presented so far. In particular, I draw
attention towards the mechanics of sub-national corruption in Mexico, and explain its particularly
puzzling outcomes.
5 Silver and lead in real life: Why Mexican majors are
getting killed like never before?
The puzzle
Political-criminal interactions and outcomes have abruptly changed in recent Mexico. Drug
traffickers, previously known for their relative distance to politics, have become increasingly
politically violent. Traffickers have become active electoral participants by killing candidates,
intimidating voters, and assassinating the press.
Political assassinations -which were pretty rare in Mexican history- have recently become
alarmingly common. Local governments, in particular mayors, have taken most of the burden.
Just in the first ten months of 2010, 12 majors were killed; a considerable figure if we note that
from 1999 to 2003 not even a single one was executed (See figure 1). Both the rapid increase in
assassinations over the last five years, and the total absence of political violence during the previous
five years are puzzling corruption outcomes. Killing should not be absent. As explained
by the model above, we know that killing not only substitute bribes but, because of its value as
a signal, it complements them. In a corrupt system, we should expect thus to have some spare
violence. Complete quietness is not then norm. The steep increase in assassinations during the
last five years also cries for an explanation. Why, if drug trafficking has long been present in
least not as much as to make them angry in a meaningful way. Traffickers are not hurt by enforcement only if it
is done in non-valuable territories or targets non-valuable assets. I do not consider this possibility because (a)
I assume that corruption demand will only exists in valuable territories, and (b) I believe selective-enforcement
is a form of corruption.
17

Mexico, it is only now that has become violent?
[Figure 1 about here]
In this section, I present a detailed analysis of the causes behind each of the 26 political
assassinations that have happened in Mexico since 1999. My case by case study shows how
the theoretical implications of my model of corruption are able to make sense of the observed
trends in violence. Political violence is not a simple decision between silver and lead –in which
increments in the power of traffickers increase levels of violence– as the traditional narrative
has argued. Silver and lead complement each other, and depend on more factors that just the
power of traffickers.
The methodology
Getting to know the reasons why traffickers have killed politicians has not been an easy task.
The following are the results of an extensive fieldwork during which I conducted over a hundred
interviews in Mexican cities known by their strong presence of illegal trafficking -Ciudad Juarez,
Nuevo Laredo, Tijuana, to name a few. Accompanied by the most renowned journalists covering
Mexico’s drug traffic -most of them already threatened by traffickers-, I repeatedly visited
each city, slowly gaining the confidence of local journalist, citizens and politicians to speak up
their first-hand experience on the mafia/politics relation. My fieldwork was complemented by
the brave reporting of Mexican local newspaper journalists. Local journalists are well informed
sources which have a fair amount of corruption stories that remain unknown due to their lack of
national relevance. I went over the local and national press myself, reviewing and corroborating
data that anonymous oral sources told me. This was a deep jump into Mexican corruption
mechanics.
Sometimes though, even the most tenacious research efforts were unsuccessful. “You cannot
know that much,” an important human rights advocate with first hand information on DTOs
once told me (XX -confidencial). I humbly agree. I cannot claim that I know the exact reasons
why mayors are killed, but I can claim that, to the best of mine and my sources knowledge, what
I present here is a fair portray of Mexico’s corruption dynamics. However, this is a research
topic that will always have some degree of speculation. The key, I believe, is to accept what we
do not know, and work creatively with what we know.
Not quite silver OR lead
The reason why, if Mexican organized crime has long been in business, it is just now that it
has become politically violent, cannot be found by using the traditional Silver-or-lead narrative.
18

Such theory, which assumes that bribes (silver) and political violence (lead) are substitutes,
can only make sense of recent political violence trends by arguing that traffickers have become
relatively more powerful overtime, which has lead to substitute bribes for killings. This thesis
is highly contestable. Mexican DTOs are currently more powerful than a couple of decades ago
but, if anything, during the last five years –period at which assassinations have increased– they
have weakened.
Mexicans indeed get a larger share of illegal revenues than during the eighties –when illegal
drug production and distribution was controlled by Colombians. Colombians weakened
when the US set up “Plan Colombia,” an important policy to provide that Latin American
country with about 1.5 billion dollars a year to fight drug production and transportation. Intensive
drug eradication, a clever policy to promote voluntary surrender (i.e. no extradition,
shorter sentences for confessed traffickers), and higher surveillance over the Caribbean diminished
the power of Colombian DTOs, and consequently empowered Mexicans (UNODC 2010,
XX).Empirically testing the Colombia-Mexico shift is not easy but evidence on the relationship
between the press and organized crime can give us an insight of it. Mexican DTOs, for example,
have become increasingly violent against the press at the same time at which Colombian’s have
become more pacific (Figure 2) 7 .
[Figure 2 about here]
Matching this trend, drugs have become increasingly available in Mexico as Mexican traffickers
have taken over Colombians. Using data of deaths caused by illegal drug consumption
to proxy for the presence of DTOs in Mexico, figure 3 shows how in the late nineties, Mexicans
stopped being a mainly opium producer country and entered the cocaine business. The industry
of illegal drugs changed, Mexico became a Cocaine producer, and relative importance of
Mexico as an opium producer diminished versus Afghanistan and China. The entrance into the
Cocaine business –and the small but stable business of marihuana that Mexican DTO’s already
had– supported Mexicans as their relative importance as opium producer diminished. In 2004,
Mexican DTOs were at the peak of their power selling a fair amount of opium and marihuana,
transporting cocaine into the US, and booming in the methamphetamine business.
[Figure 3 about here]
Figure 3 though, also shows how Mexico’s organized crime faced a large decrease in its power
starting in 2005 8 . It is 2005, the year at which a large territorial battle between Sinaloa and
7 Other variables, such as (a) the number of asylum petitions to the US, as well as (a) the number of criminals
caught at American territory, follow the same negative correlation.
8 Power is defined here, as wall as in the rest of the paper, as violence capacity. The assumption is that higher
sales increases traffickers’ ability to be violent because higher revenues can be invested in improving killing
technology.
19

Golf DTOs emerged. The battle has as a goal to get control of Nuevo Laredo, the main port
of entry of legal merchandise into the US. This year, marks the beginning of a drug war that
still remains alive and that has taken the live of at least 26,739 individuals in the last three
years (Figure 4). The territorial war between Golf an Sinaloa DTOs was triggered, among other
things because of a temporal instability in Golf’s leadership caused by the imprisonment of the
head of Golf organization, Mr. Osiel Cardenas (Ravelo 2009). The war was further increased
when Mexico’s Federal government launched, in 2007, the most important offensive against drug
trafficking that the country has seen. Mr. Felipe Calderon, elected president of Mexico in 2006,
called during his first month in office for a “war on drugs,” and sent troops to control the most
important trafficking points. This increase in enforcement reduced the ability of traffickers to
keep doing business as usual, and further destabilized the internal dynamics of DTOs when
drug lords were captured and extradited.
[Figure 4 about here]
Changes in the power of organized crime are compared with changes in the capability of the
Mexican government to prosecute in Figure 5. State capacity is here measured as the number
of law enforcement officials working at the Federal Agency of Investigations (Agencia Federal
de Investigaciones, AFI), and the number of active elements at the army and the military
(Secretaria de Defensa Nacional (Sedena) and Secretaria de Marina (Semar)). This measure
is imperfect as it only measures capacity, not capability/willingness to enforce. Other aspects,
such as leadership, coordination, training, and physical resources may be more important to
determine the real power of the government 9 . Yet, even if these figures were not completely
accurate, there is no doubt that, at least since Felipe Calderon took office in 2007, enforce
against drug traffickers has increased in Mexico.
[Figure 5 about here]
To summarize, this is not the story of an increasingly powerful Mexican organized crime
that became more violent. This is the story of an organized crime that boomed quietly, and
weakened loudly. If anything, during the years where political violence has increased (2004 and
on), Mexican DTOs have diminish their power as a result of intra/inter DTO’s territorial battles
and increases in law enforcement. Empirical shows a Mexican government that has systematically
increased its human resources to fight drug traffic, and (since 2007) its willingness to do it.
Political violence thus, has increased as traffickers have become less powerful, a result at odds
with traditional understandings of corruption. Mexican politicians seem to be increasing their
capability to get silver, at the same time that they are getting more lead 10 .
9 Using statistics on detention and drug seizures though, is not recommendable as those figures are endogenous
to the size of DTOs.
10 An alternative explanation is that of general disorder. Current drug war may have diminished the cost of
killing because homicides have increased overall. Authorities are not able to punish homicides, and impunity has
20

Insight 1: Traffickers are not stronger, government is not (necessarily) weaker, and yet
politicians are getting killed more.
Understanding corruption outcomes in Mexico: demand and supply
dynamics
My model of the dynamics of corruption can yield some insights into the reasons for this puzzling
result. Corruption supply and demand are the key. Mayors were not killed before (1999 - 2003)
because during the strongly centralized PRI-regime there was not a market for sub-national corruption.
Mayors are killed because the illegal-drug industry has changes. Markets have become
more competitive, and DTOs have turned into multiproduct, domestic sellers. Such changes
have increased the demand for sub-national corruption, made killings more valuable as signals,
and increased the price of bribes. In short, corruption outcomes changes in Mexico because of
(a) the decentralization of political institutions, and (b) changes in the illegal-drug industry.
First, the decentralization of Mexican government has increased political violence because
it opened a new source of corruption supply (the sub-national supply). The rupture of the
hegemony of PRI as the dominant party shifted the balance of power from the national to the
sub-national level. The PRI, which ruled Mexico for more than 70 years, had a very centralized
form of government. All policy decisions were taken at the center and were not questioned by
state politicians, mainly because a well instituted system of political clientelism was in place.
All the political careers of state-level politicians were determined by direct appointments coming
from the national level. Beginning in the late eighties this started to change. State politicians
became increasingly independent when the first victories of the opposition broke the ability of
the federation to keep political clientelism working 11 . The recent independence of sub-national
figures made sub-national corruption a possibility.
If majors are being killed more now, is not (necessarily) because traffickers are more powerful
than before (which may not be) but because, mayors now have the political power to influence
the trafficking business by endorsing or rejecting federal policies. It is not a coincidence that
the number of mayors assassinated has increased with decentralization.
increased. Indeed, political assassinations have increased following national homicide rates (see Appendix 5). I
doubt of the explanatory power of general disorder theories because (a) mayors have not been assassinated in the
most violent municipalities, (b) neither during the peak of violence of their states (Appendix 4) or municipalities
(Appendix 5)
11 Note that this is not an argument about democratization but about centralization. Even under an authoritarian
regime, sub-national politics could have been decentralized, and mayors could have supplied corruption.
Democratization and centralization are not interchangeable terms.
21

The case of Mr. Alfonso Pena, the mayor of Tepehuanes, a municipality of Durango where
poppy has long being growth, is an example of the power that mayors have recently got. In
2008, just a couple of months after Mr. Pena took office, his entire police department resigned.
His police though was integrated by only five members. They were scared because traffickers
had killed one policeman, kidnapped another, and wounded one more. After the massive resignation
nobody in the municipality wanted to be a policeman anymore. Mr. Pena decided to
become the police himself. With his armored hummer and his brother company, he patrolled
Tepehuanes for a month until help, brought from the federal government, arrived town. Traffickers
did not liked this. “Allow us to work bastard, and take out the military or do you want
to lose your head?” were the type of calls that Mr. Pena started getting (Almazan 2010). Mr.
Pena disappeared on July 26th, 2010 (Hernandez 2010) and has not being founded yet.
Even if Mr. Pena lacked of a real enforcement capacity himself (at most five policemen), he
had the possibility to call the federation for support. This possibility would not have been an
option during the PRI hegemony because all enforcement decisions were taken at the center;
mayors lacked of a voice. It is the recent capacity of Mr. Pena to supply corruption (by deciding
whether to work or not with the federation), what made him (and other mayors 12 ) a target of
traffickers.
Insight 2: Decentralization made sub-national politicians potential corruption suppliers
Second, mayors are being killed more because changes in DTOs target consumers and products
have increased the demand for sub-national corruptions. DTOs now need mayors more.
As Figure 6 shows, Mexican DTO’s have shifted from being drug exporters only, to be drug
dealers inside Mexico. The exporting business only requires control of highways and ports of
entry into the US –which are all federally regulated. Domestic sells though, require a tighter
control of local authorities; drugs have to be sold in street corners and local policemen have to
agree on it. Corrupting mayors in more profitable because by gaining control over a territory,
DTOs can both use the highways to introduce drugs into the US (old consumers) and sell to
Mexican consumers (new consumers).
[Figure 6 about here]
Furthermore, Mexican DTOs have also become multiproduct firms. Mexicans became
methamphetamine producers at the beginning of the 00s. The production of synthetic drugs
requires building laboratories which significantly increase the value of corruption agreements
with local authorities. The laboratories have to remain clandestine. Out of all laboratories that
12 Mr. Luis Carlos Ramirez, mayor of Ocampo, Durango, was assassinated on July 1st, 2009 for denouncing
traffickers to federal authorities. As the mayor of Pueblo Nuevo, another municipality of Durango, candidly
acknowledged: “[he] was killed because he denounced an armed group [], the mayor of El Mezquital, I do not
know what [he did to] bother them [traffickers], but they killed him” (La Loncheria, 2010).
22

have been discovered by the federation, 56% of them were located in politically violent states,
even though these only represent 34% of all the states.
Insight 3a: Changes in Mexico’s illegal-drug industry (i.e. DTOs becoming multiproduct
and domestic sellers) created a demand for sub-national corruption.
Mayors are also being killed more because Mexico’s illegal industry became more competitive,
and with competition the price of bribes have increased, and killing has become more
valuable as a signal. In short, the utility of political assassinations increased. Because corrupting
the government is a necessary input for drug production, competition for territory makes
bribes very expensive for the newcomer DTO. If a mayor works for Sinaloa DTO, he must
refuse to accept bribes from another DTO, as this will create a large incentive for Sinaloa DTO
to kill him. Traffickers have made it very clear that making deals with two DTOs is a death
sentence. “[The mayor] has been killed [] because he betrayed his people. He [] became the
principal operator of a group of assassins and kidnappers called Los Zetas (Proceso 2010)” was
the message that Golf DTO left when Mr. Leal, mayor of Hidalgo, Tamaulipas was killed in
August 28th, 2010.
Bribing a mayor that is already being bribed is very expensive (or impossible). Newcomer
DTOs face the logical equivalent of an infinite bribe price. Kill the mayor thus, becomes the
only possibility to compete into the market. Killing is a high entry cost to pay but, never as
costly as remaining out of business.
Figure 7 shows the detailed causes of mayor assassinations. Competition, which takes form
as violent territorial fights, is perhaps one of the stronger determinants of political violence.
From the 26 cases of assasinated mayors from 1999 to September 2010, 24 are directly related
to drug trafficking. I could find the causes for 20 cases. Out of these, 14 are directly related to
territorial wars. Historically, we can see how the number of mayors assassinated has increased
as the number of competing organization has increased. The first rupture of oligopolies in the
00s came in December of 2004, the exact year in which the first two mayors were assassinated.
Since then, inter-DTOs conflicts have increased in number. While in 2007 there were only four
competitive markets (Sinaloa vs. Tijuana; Sinaloa vs. Juarez, Sinaloa vs. Golfo-Zetas, and
La Familia vs. Zetas), in 2010 the number increased to eight (Sinaloa vs. Tijuana, Golfo vs.
Juarez, Sinaloa vs. Golfo, La Familia vs. Zetas, Sinaloa vs. Beltran Leyva, Sinaloa vs Linea,
Golfo vs. Zetas, and Pacifico Sur vs. La Barbie) (Guerrero etal 2010 complemented by author).
[Figure 7 about here]
Mayors are well aware of the negative consequences that competition may bring them and
thus have changed their strategy from taking bribes and not enforce, to not taking bribes and
not enforce. When the mayor of (XX- confidential) took office in 2007, he received as a gift a
23

iefcase full of money. He rejected it because he knew that under the strong competition that
the municipality faced, any deal done with a DTO will immediately put him in danger to be
killed by the other. The only viable solution for mayors ruling in competitive municipalities is
to reject bribes at all and not enforce. Enforcing is twice as dangerous when two enemies are
being fought.
The mayors of Durango, a state where seven mayors have been killed since 2004 –three on
row at the municipality of Otaez–, have been very explicit about their approach to corruption.
First, “pretend to be stupid” claimed one of them (XX-confidential). Second, never enforce
“Nooo!” said the mayor of Santiago when asked whether he prosecuted traffickers, “I never
bother anybody. My police do not prosecute drug traffickers, they only arrests boozers and
petty thieves.” Third, better avoid becoming a mayor. “This job of being a mayor is not recommendable
anymore” concluded the mayor of Guacevi. The advice, given by the Mayor of
Poanas, to those that already have the job was “to trust in God.”
Note that assassinations will only happen when mayors get trapped into a bad equilibrium.
When they took office competition was not in place, which allowed them to made explicit arrangements
with one DTO. Later though, competition started and they were unable to either
break the agreement that they had, or made a new one with the new DTO. The only chance to
remain alive in such case is to run away, but even that is difficult given that traffickers can still
take revenge with whoever remains in town (i.e. family, friends). In the words of the mayor of
Pueblo Nuevo: “if I leave now, they [drug traffickers] will think that I’m guilty.”
With competition the value of killing as a sign of power increases. Killing is a way to generate
a reputation. Reputation double pays in competitive environments because it not only
reduces the price of bribes but, deters potential competitors. At least 2 of the 20 cases studied
here are explained as signaling cases. The mayor of Mezquital was killed, an intelligence report
(XX-confidential) claim, as a threat to the governor of Durango; the mayor and the governor
were very close friends.
Insight 3b: Changes in Mexico’s illegal-drug industry (i.e. DTOs becoming more competitive)
increased the price of bribes, and made killing more valuable as a signal.
To summarize, this section has presented the results of a case by case study of assassinated
mayors. I showed that, unlike silver-or-lead narrative would predict, political violence does
not necessarily increase as traffickers become more powerful. Mayors are getting killed now
because (a) decentralization open a new sub-national corruption market, and (b) recent market
competition between DTOs made bribes get higher, and increased the value of killing as a signal.
24

6 Conclusion
The simplicity through which drug-trafficking and mafia-related corruption has been understood
by current literature oversimplifies the decisions, interactions and strategies available for
the actors. We have not being able to get a complete understanding of corruption outcomes
because we have failed to consider how institutions and market changes affect the dynamics of
corruption. In this paper I showed that the state is useful for traffickers, and that violence does
not always increase as the power of traffickers increase. Killing is costly and uncertain, and
dismantling the state may entail higher costs (because of reductions in legal trade) than any
benefit brought by the elimination of bribes.
A simple bargaining model to explain political/criminal interactions was presented. In my
model, bribes can be substituted by killings (as traditional understandings of corruption would
argue) but can also be complemented with killings. Killings may complement bribes because
politicians are never sure about the actual power that traffickers have to kill. Killing functions
as a signal to keep bribes affordable.
The comparative statics of the model showed that sometimes bribing more is less, not more
expensive for traffickers; that sometimes politicians want small bribes; that even weak traffickers
can extort powerful states, and that politicians not always take a decision between being bribed
or killed. A final interesting result refers to the quality of politicians: In a corrupt environment,
all available politicians will be corrupt unless the state has the ability to extort criminals. To
be completely virtuous, a politician needs to be capable of being completely corrupt.
My model of corruption dynamics was originally motivated by the need to explain recent
puzzling changes in the relationship between Mexico’s politicians and criminals that I believe
the traditional silver-or-lead approach cannot fully account for. Mexico shifted from being a
completely peaceful country (no politicians were killed from 1999 to 2004), to be a mayhem (12
mayors were killed in the first 9 months of 2010).
Drawing on my own fieldwork experience in Mexico’s drug war zone, I use the results of my
model to explain political violence. The reasons for such radical change, I have argued are not
to be found in increases in the power of traffickers –as traditional narrative would argue– but
in analyzing political institutions and features of the illegal-drug industry. Mayors are being
killed because of (a) as the country has decentralized a market for sub-national corruption has
emerged, because (b) Mexican DTOs have become multiproduct firms which sell drugs inside
Mexico (not only transport them to the US), and because (c) the illegal drug-industry have
become more competitive. These changes have made mayors’ killings more prevalent because
mayors have increased their relevance as political actors, bribes have become more expensive,
and signaling has become more valuable.
25

Corruption is a complex game where signals, beliefs and expectations play a large role in
determining outcomes. Bribe and getting bribed are actions subject to many subtleties. One
side has to avoid to get killed; the other has to keep a fragile equilibrium between killing, and
not-killing; both sides cannot get too powerful neither too weak. Power sometimes hurts, and
weakness sometimes benefits. The game of corruption is without any doubt a difficult one to
play. In the words of Mr. (XX -closed source), local politician: “I do not know [what one has
to do to remain alive], one just has to be very, very clever.”
26

Graphs, Figures and Appendixes
Graph 1: Determining the price of bribes
(1a) Different bribes for different politicians/territories.
Bribe
Fair-trade b'
State protection (negative enforcement)
b' for a relevant/connected politician
b' for relevant/connected politician at non-crucial territory
b' for unrelevant/disconnected politician at crucial territory
Traffickers pay politicians a bribe in exchange for state protection. The “fair-trade” price of a bribe (b’) is
determined by (a) how much protection the trafficker gets, (b) how relevant/connected the politician is,
and (c) how crucial the territory in question is for illegal businesses. The slope of b’ increases as the
politician is more relevant/connected; the point at which b’ crosses the y access increases as the territory
becomes more crucial. In very relevant territories even no-enforcement policies require the trafficker to
pay a positive bribe to the politician.
All bribes above the b’ line represents the politicians extorting traffickers and viceversa. Note that
trafficker extortion may happen with more frequency because unlike traffickers, the state never
completely known the power of his enemies.
1

(1b) What should the price of the bribe be according to trafficker’s/politician’s logic?
Bribe
State protection (negative enforcement)
according to trafficker's logic
according to politicians's logic
Politicians want bribes to be relatively higher as they protect traffickers more because hiding a lot of
protection is more difficult. Traffickers believe bribes should be relatively lower as state protection
increases because a corrupt state prosecutes traffickers less. The actual form of the line is undetermined.
2

Graph 2: The relationship between bribes and killings
Politicians' power
Trafficker's power
Bribes
If killing substitute bribes
If killing complements bribes
Bribes decrease as the power of trafficker’s increases (normal line) up to a point in which they are zero. At
this point politicians do not enforce (and sometimes even protect traffickers) without getting bribes
because they are too scared. Once this “threatening level” is crossed, the number of killings gets stabilized
(zero slope). Killing does not become zero because some selective killing has to be done to keep
politicians scared over time (i.e. trafficker’s power is unknown for politicians; there is an information
problem).
The traditional understanding of corruption dynamics (Silver-or-lead) argues that killings increase with
trafficker’s power because bribes get substituted by political violence. Powerful traffickers do not care
about making deals with politicians; they just kill them (dashed line). I argue that killings we can also think
on killings as complement of bribes (bold line). When bribes are high, traffickers want to kill more because
killing diminishes the price of bribes by changing politicians’ beliefs about traffickers’ power. Lower bribes
give traffickers fewer incentives to be violent.
3

Figure 1: Mexican mayors assassinated by DTOs
14
12
10
8
6
4
2
0
0 0 0 0 0
12
6
3
2 2
1
0
2004 2005 2006 2007 2008 2009 2010
Mexico’s total lack of political violence until 2004 is as puzzling as the following spike.
Source: Author based on press and official reports
Figure 2: Drug trafficking moves from Mexico to Colombia (proxies)
(2.1) Journalist murdered by confirmed/unconfirmed motives
10
8
6
4
2
0
-2
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Mexico
Colombia
Mexican press has been increasingly assassinated by organized crime, as trafficking has moved from
Colombia to Mexico (See Appendix 2 for a complete list of assassinated Mexican journalists assassinated
during the last four years by name, media and place)
Note: Not counted are other seven journalists which have disappeared in Mexico since 2007. The direct
implication of DTOs has not been officially proved in all instances given that most of the judicial cases are
still open.
Source: Lauria and Mike O'Connor (2010) complemented by author.
4

Figure 3: Mexican DTOs presence, by drug type (1990 – 2008)
Overdosis deaths
50
45
40
35
30
25
20
15
10
5
0
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008
Opium Marihuana Cocaine Total
Mexican DTOs’ presence increases from 1990 to 2004, and sharply decline from 2005 to 2007. Profits
from opium have diminished over time, marihuana was profitable only during the late nineties, and
cocaine experienced a boom from 1998 until 2005. Profits from hallucinogens (not pictured) and other
synthetic drugs, boomed in the early 2000-2003.
Note: I use number of overdoses deaths in Mexico as a proxy of DTOs presence (Rios 2010a). I rely in two
assumptions: (a) overdoses happen with a fixed probability among consumers, and (b) the correlation
between domestic and international profits is positive and fixed over time.
Source: Author compilation based on Mexican census information (Rios 2010a).
Figure 4: Drug-related deaths (2007 – August 2010)
1400
1200
1000
Drug-related homicides
800
600
400
200
0
J F M A M J J A S O N D
2007 2008 2009 2010
Drug related deaths increased in 2008, and 2010, and remained stable in 2007 and 2009. From 2007 until
August 16 th , 2010 official intelligence reports (XX –closed source) had accounted for 26,739 drug related
assassinations, most of them in Chihuahua. Independent media accounts, less accurate, counted 21,806
assassinations for the same period. Source: Duran and Rios (2010) and (XX –closed source)
5

Figure 5: Strength of Mexican traffickers versus the state (1995 – 2009)
2
1
0
-1
-2
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Strenght of Drug traffickers (market presence)
State capacity to prosecute (Military personel)
State capacity to investigate (AFI agents)
Traffickers were more powerful than the state in the late nineties, and the period immediately before the
drug battles exploited. Mayors have been assassinated the most in the periods where drug traffickers
have been the least powerful.
Note: The strength of drug traffickers was calculated using drug consumption measures (see Table 4).
State capacity to investigate was measured as the number of law enforcement officials working at the
Federal Agency of Investigations (Agencia Federal de Investigaciones, AFI). State capacity to prosecute
was measured as the number of active elements at the army and the military (Secretaria de Defensa
Nacional (Sedena) and Secretaria de Marina (Semar)). All measures were normalized for comparison.
Figures for 2009 are linearly projected, based on monthly averages.
Source: Guerrero etal (2010) and author compilation based on Mexican census information (Rios 2010a).
Figure 6: Size of Mexican domestic illegal-drug market (1990 – 2008)
600
500
400
300
200
100
0
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008
Municipalities with openned domestic drug markets
Overdoses fatalities
A domestic market is considered to be opened when at least one overdoses fatality happened in the
municipality. This measure underestimates consumption for municipalities with fewer population.
Source: Author compilation based on Mexican census information (Rios 2010a).
6

Figure 7: Causes of political assassinations, case by case (1999 – 2010)
Name Municipality State Date Territorial fight Reason?
Jose Manuel Soto Ortiz Otaez** Durango July 8, 2004 Sinaloa vs Linea Had nexus with Sinaloa
Esteban Estrada Corral Otaez Durango October 5, 2004 Sinaloa vs Linea Retailiation for his debts or nexus with a DTO, both motives were investigated
Fernando Chavez Lopez Buenavista Michoacan July 8, 2005 Refused to pay quota to Golf
Raul Delgado Benavides Cuauhtitlan de GB Jalisco July 16, 2006 Not-drug related
Walter Herrera Ramirez Huimanguillo Tabasco November 16, 2006 Milenio vs Valencia
Marcelo Ibarra Villa Villa Madero Michoacan June 1, 2008
Homero Lorenzo Rios Ayutla de los Libres Guerrero September 25, 2008 Yes (not specified)
Salvador Vergara Ixtapan de la Sal Mexico October 4, 2008 Refused to be extorted by La Familia
PRD says that groups linked to PRI are guilty; others argue he was linked to
trafficking but no evidence.
Claudio Reyes Nunez Otaez Durango February 6, 2009 Sinaloa vs Linea Retailiation for asassination of mayor Manuel Soto
Raul Mendivil Sotelo Guadalupe y Calvo Chihuahua February 17, 2009 Sinaloa vs. Linea Signaling to federal government
Octavio Carrillo Vista Hermosa Michoacan February 24, 2009 Retailiation for firing policemen
Luis Carlos Ramirez Ocampo Durango June 1, 2009 Sinaloa vs. Linea Denounced to federation
Manuel de Jesus Angulo Topia* Durango June 3, 2009 Sinaloa vs. Linea Linea vs Sinaloa
Hector Ariel Meixueiro Namiquipa Chihuahua July 14, 2009 Sinaloa vs Linea Linea vs Sinaloa
Bernardo Mecinas Lopez Animas Trujano Oaxaca February 19, 2010
Manuel Estrada Mezquital Durango February 22, 2010 Sinaloa vs Linea Signaling to governor
Jose Santiago Agustino Zapotitlan Tablas Guerrero April 28, 2010
Manuel Lara Rodriguez Guadalupe Chihuahua June 19, 2010 Sinaloa vs. Linea Had nexus with Linea; her daughter was dating a trafficker
Oscar Venancio Martinez San Jose del Progreso Oaxaca June 19, 2010 Not-drug related
Nicolas Garcia Ambrosio Sto Dom de Morelos Oaxaca June 30, 2010 Not-drug related
Alfonso Pena Pena Tepehuanes* Durango July 26, 2010 Sinaloa vs. Linea Denounced to federation
Edelmiro Cavazos Leal Santiago NL August 18, 2010 Zetas because he was interfering with business
Marco Antonio Leal Hidalgo Tamaulipas August 28, 2010 Golfo vs Zetas
Alexander Lopez Garcia El Naranjo SLP September 8, 2010 Called federal forces
Prisciliano Rodriguez Dr Gonzalez NlL September 24, 2010 Golfo vs Zetas
Jesus Montejano Tancitaro Michoacan September 28, 2010
Mayors began to be killed/disappeared in 2004; 26 have been killed until September 2010, 24 in relation with drug trafficking; I found the cause of the murder
for 20 of the 24 drug-related cases. Note: * disappeared, ** elected but had not taken office yet; the DTO to which the killing was attributed is underlined.
7

Appendix 1: Notation
N – Drug trafficker
M – Illegal profits
A – Politician
Pn – Trafficker’s strength (power of violence)
Pa – Politician’s strength (power of violence)
b – Bribe (aka silver); given as a share of total drug trafficking profits
k – Kill (aka lead); either 0 or 1
e – Enforce; positive if enforcing law, negative if protecting criminals
pi – Politician’s believes; probability of trafficker’s power of violence to be higher than
politician’s.
DTO – Drug traffic organization
Appendix 2: Definitions, Axioms, Lemmas, Results and Empirical Insights
Definition 1: Corruption is a bargaining game in which N and A agree on the price of a
bribe (b).
Definition 2: N decides between paying b, and exerting political violence (k) (i.e. killing
A)
Definition 3: A decides between rejecting b and prosecute/ignore crime crimes (e>=0),
and accepting b and protect/ignore crimes (e0 is a necessary but not sufficient condition for trafficking to exist.
Axiom 2: b decreases in Pn and increases in Pa.
Axiom 3: Negotiating the value of b is costly.
Axiom 4: k is affected by r
Lemma 1: N prefers to have an A than to become the state themselves, independently
of Pn and Pa.
Lemma 2: k does not necessarily increase in Pn
Lemma 3: Because k builds pi, k reduces b
Lemma 4: N does not always prefer to face a small Pa
Result 1: The concavity of b’ is undetermined. “Sometimes bribing more is less, not more
expensive, and sometimes politicians want small bribes”
Result 2: If b0, k>0 is required for b=0 “No-silver and lead”
Result 5: When M>0, k=0 if and only if b=M is a viable equilibrium “Silver and no-lead”
8

Result 6: When M>0, only corrupt A will take office, unless b=M is a viable equilibrium
“To be completely virtuous, one need to be able to be completely corrupt”
Insight 1: Traffickers are not stronger, government is not (necessarily) weaker, and yet
politicians are getting killed more.
Insight 2: Decentralization made sub-national politicians potential corruption suppliers
Insight 3a: Changes in Mexico’s illegal-drug industry (i.e. DTOs becoming multiproduct
and domestic sellers) created a demand for sub-national corruption.
Insight 3b: Changes in Mexico’s illegal-drug industry (i.e. DTOs becoming more
competitive) increased the price of bribes, and made killing more valuable as a signal.
Appendix 3: Mexican journalists murdered
(a) Motive confirmed
Journalists slain in direct relation to their work:
1 Rodolfo Rincon Taracena, Tabasco Hoy, January 20, 2007, Villahermosa
2 Amado Ramirez Dillanes, Televisa and Radiorama, April 6, 2007, Acapulco
3 Alejandro Zenon Fonseca Estrada, EXA FM , September 24, 2008, Villahermosa
4 Armando Rodriguez Carreon, El Diario de Ciudad Juarez, November 13, 2008, Ciudad
Juarez
5 Eliseo Barron Hernandez, La Opinion, May 25, 2009, Gomez Palacio
6 Norberto Miranda Madrid, Radio Vision, September 23, 2009, Nuevo Casas Grandes
7 Bladimir Antuna Garcia, El Tiempo de Durango, November 2, 2009, Durango
8 Valentin Valdes Espinosa, Zocalo de Saltillo, January 8, 2010, Saltillo
Media support workers slain in the course of their duties:
9 Flor Vasquez Lopez, El Imparcial del Istmo, October 8, 2007, between Salina Cruz and
Tehuantepec
10 Mateo Cortes Martinez, El Imparcial del Istmo, October 8, 2007, between Salina Cruz
and Tehuantepec
11 Agustin Lopez Nolasco, El Imparcial del Istmo, October 8, 2007, between Salina Cruz
and Tehuantepec
Journalists missing
12 Gamaliel Lopez Candanosa, TV Azteca Noreste, May 10, 2007, Monterrey
13 Gerardo Paredes Perez, TV Azteca Noreste, May 10, 2007, Monterrey
14 Maria Esther Aguilar Cansimbe,, El Diario de Zamora and Cambio de Michoacan,
November 11, 2009, Zamora
15 David Silva, El Manana and La Tarde, March 2010, Reynosa
16 Pedro Argüello, El Manana and La Tarde, March 2010, Reynosa
17 Miguel Angel Dominguez Zamora, El Manana, March 2010, Reynosa
18 Ramon Angeles Zalpa, Cambio de Michoacan, April 6, 2010, Paracho,
(b) Motive unconfirmed
Journalist possibly slain for their work:
19 Saul Noe Martinez Ortega, Interdiario, April 2007, Nuevo Casas Grandes,
9

20 Gerardo Israel Garcia Pimentel, La Opinion de Michoacan, December 8, 2007,
Uruapan,
21 Mauricio Estrada Zamora, La Opinion de Apatzingan, February 12, 2008,
Apatzingan,
22 Teresa Bautista Merino, La Voz que Rompe el Silencio, April 7, 2008, Putla de
Guerrero,
23 Felicitas Martinez Sanchez, La Voz que Rompe el Silencio, April 7, 2008, Putla de
Guerrero,
24 Miguel Angel Villagomez Valle, La Noticia de Michoacan, October 10, 2008,
between Lazaro Cardenas and Zihuatanejo,
25 Jean Paul Ibarra Ramirez, El Correo, February 13, 2009, Iguala,
26 Carlos Ortega Samper, El Tiempo de Durango , May 3, 2009, Santa Maria del Oro,
27 Juan Daniel Martinez Gil, Radiorama and W Radio , July 28, 2009, Acapulco,
28 Jose Emilio Galindo Robles, Radio Universidad de Guadalajara, November 24, 2009,
Ciudad Guzman,
29 Jose Alberto Velazquez Lopez, Expresiones de Tulum, December 22, 2009, Tulum,
30 Jose Luis Romero, Linea Directa, January 2010, Los Mochis,
31 Jorge Ochoa Martinez, El Sol de la Costa, January 29, 2010, Ayutla de los Libres,
32 Evaristo Pacheco Solis, Vision Informativa, March 12, 2010, Chilpancingo
Appendix 4: Drug-related murders in municipalities with/without political violence
(2007 – September 2010)
40
35
30
25
20
15
10
5
0
1
1
1
3
2
1
2
1
1
1
2
3
Jan, 2007
Feb, 2007
Mar, 2007
Apr, 2007
May, 2007
Jun, 2007
Jul, 2007
Aug, 2007
Sep, 2007
Oct, 2007
Nov, 2007
Dec, 2007
Jan, 2008
Feb, 2008
Mar, 2008
Apr, 2008
May, 2008
Jun, 2008
Jul, 2008
Aug, 2008
Sep, 2008
Oct, 2008
Nov, 2008
Dec, 2008
Jan, 2009
Feb, 2009
Mar, 2009
Apr, 2009
May, 2009
Jun, 2009
Jul, 2009
Aug, 2009
Sep, 2009
Oct, 2009
Nov, 2009
Dec, 2009
Jan, 2010
Feb, 2010
Mar, 2010
Apr, 2010
May, 2010
Jun, 2010
Jul, 2010
Aug, 2010
Sep, 2010
Total
Mayor (drug related)
Drug-related murders at non-politically
violent municiaplities
1400
1200
1000
800
600
400
200
0
Jan, 2007
Feb, 2007
Mar, 2007
Apr, 2007
May, 2007
Jun, 2007
Jul, 2007
Aug, 2007
Sep, 2007
Oct, 2007
Nov, 2007
Dec, 2007
Jan, 2008
Feb, 2008
Mar, 2008
Apr, 2008
May, 2008
Jun, 2008
Jul, 2008
Aug, 2008
Sep, 2008
Oct, 2008
Nov, 2008
Dec, 2008
Jan, 2009
Feb, 2009
Mar, 2009
Apr, 2009
May, 2009
Jun, 2009
Jul, 2009
Aug, 2009
Sep, 2009
Oct, 2009
Nov, 2009
Dec, 2009
Jan, 2010
Feb, 2010
Mar, 2010
Apr, 2010
May, 2010
Jun, 2010
Jul, 2010
Aug, 2010
Sep, 2010
Political violence increases as general violence increase but does not necessarily happen in the most
violent periods. Drug-related deaths in non-politically violent municipalities follows an increasing trend
not matched by politically violent municipalities.
Source: Rios 2010a and (XX –closed source)
General
10

11
Appendix 5: Drug-related deaths at states where mayors were assassinated
Chihuahua
0
10
20
30
40
50
60
70
80
90
42/07
46/07
52/07
3/08
7/08
11/08
15/08
19/08
23/08
27/08
31/08
35/08
39/08
43/08
47/08
51/08
3/09
7/09
11/09
15/09
19/09
23/09
27/09
31/09
35/09
39/09
43/09
47/09
51/09
3/10
7/10
11/10
15/10
19/10
23/10
27/10
31/10
35/10
Durango
0
10
20
30
42/07
46/07
52/07
3/08
7/08
11/08
15/08
19/08
23/08
27/08
31/08
35/08
39/08
43/08
47/08
51/08
3/09
7/09
11/09
15/09
19/09
23/09
27/09
31/09
35/09
39/09
43/09
47/09
51/09
3/10
7/10
11/10
15/10
19/10
23/10
27/10
31/10
35/10
Guerrero
0
10
20
30
42/07
46/07
52/07
3/08
7/08
11/08
15/08
19/08
23/08
27/08
31/08
35/08
39/08
43/08
47/08
51/08
3/09
7/09
11/09
15/09
19/09
23/09
27/09
31/09
35/09
39/09
43/09
47/09
51/09
3/10
7/10
11/10
15/10
19/10
23/10
27/10
31/10
35/10
Estado de Mexico
0
10
20
30
42/07
46/07
52/07
3/08
7/08
11/08
15/08
19/08
23/08
27/08
31/08
35/08
39/08
43/08
47/08
51/08
3/09
7/09
11/09
15/09
19/09
23/09
27/09
31/09
35/09
39/09
43/09
47/09
51/09
3/10
7/10
11/10
15/10
19/10
23/10
27/10
31/10
35/10
Michoacan
0
10
20
30
42/07
46/07
52/07
3/08
7/08
11/08
15/08
19/08
23/08
27/08
31/08
35/08
39/08
43/08
47/08
51/08
3/09
7/09
11/09
15/09
19/09
23/09
27/09
31/09
35/09
39/09
43/09
47/09
51/09
3/10
7/10
11/10
15/10
19/10
23/10
27/10
31/10
35/10

12
Nuevo Leon
0
10
20
30
40
50
42/07
46/07
52/07
3/08
7/08
11/08
15/08
19/08
23/08
27/08
31/08
35/08
39/08
43/08
47/08
51/08
3/09
7/09
11/09
15/09
19/09
23/09
27/09
31/09
35/09
39/09
43/09
47/09
51/09
3/10
7/10
11/10
15/10
19/10
23/10
27/10
31/10
35/10
Tamaulipas
0
10
20
30
40
50
60
70
80
90
42/07
46/07
52/07
3/08
7/08
11/08
15/08
19/08
23/08
27/08
31/08
35/08
39/08
43/08
47/08
51/08
3/09
7/09
11/09
15/09
19/09
23/09
27/09
31/09
35/09
39/09
43/09
47/09
51/09
3/10
7/10
11/10
15/10
19/10
23/10
27/10
31/10
35/10