Killing Candidates in Mexico. The PRD in the 1990s.


The 2018 electoral process has been described as the most violent in the history of Mexico. According to records compiled by Estrategia Electoral, “between September 8th, 2017, and August 31st, 2018 (initial and final dates of the federal electoral process), at least 145 local political actors or people related to electoral processes were murdered”.1Animal Político, 6 de mayo de 2019.. The sociologist Arturo Álvarez Mendoza arrived at almost asimilar figure, counting 140 deaths.2Alvarado Mendoza, Arturo, “Violencia política y electoral en las elecciones de 2018”, Alteridades, 29, 2019.

In the recent year, political analysts seem to have discovered a new face of violence in the country: electoral violence. However, a reflection on Mexico’s political history shows that the episodes of 2018 must be inserted into a long chronology of violence that is detonated and resounds, above all, during electoral periods. The aim of this article is to show how the violence lived 30 years ago, during Carlos Salinas de Gortari’s presidency (1988-1994), which targeted mainly militants of the recently created Partido de la Revolución Democrática (PRD, Party of the Democratic Revolution), allows to trace a series of continuities and ruptures with the electoral violence witnessed in 2018.

Killing Candidates in ‘Silence’

 “Murdered militants? It’s an invention of the PRD”, told me one academic when I was documenting the early years of the party. During my doctoral research, I set out to analyze the impact of repression on the PRD’s construction, a topic virtually unexplored at the time. This was in the mid-to-late 1990s, precisely during Mexico’s democratic transition. At that time, speaking of violence in academic spaces – in sharp contrast to what happens today – was almost taboo. Political scientists of the time perceived political changes from the national perspective or, more exactly, from Mexico City, so they tended to consider the homicides that occurred in Guerrero or Michoacán a far-off geographic and social reality associated, historically, with the so-called “Dirty War” (Guerra Sucia) of the 1970s, not with the political aperture of the 90s.

In the 90s, homicides that occurred in Guerrero or Michoacán were a far-off geographic and social reality.

Speaking of violence in academic spaces was almost taboo.

Postures of this kind obliged me to adopt the most rigorous methodology possible to study this phenomenon. Thus, I elaborated a database of homicides, within which I carefully cross-referenced various sources to select only clearly documented cases. Those sources included a report by the Comisión Nacional de Derechos Humanos (CNDH, National Human Rights Commission); the magazine Proceso, which was one of the few publications to give broad coverage to this topic; and a collection of reports by the Fundación Ovando y Gil, a group dedicated to supporting the orphans of murdered militants.3Its name honors Francisco Xavier Ovando and Román Gil Heraldez, two collaborators of Cuauhtémoc Cárdenas (the founder of the PRD) who were executed in Michoacán during the 1988 electoral campaign.

My database contains records of 265 PRD militants who were killed in the 5-year period from 1989 to 1994.4Combes Hélène, ‘Gestion des manifestations dans le Mexique des années 1990’, in Olivier Fillieule, Donatella della Porta, Police et manifestants: Maintien de l’ordre et gestion des conflits, Paris: Presses de Sciences Po, 2006; Combes, Hélène. Faire parti: Trajectoires de gauche au Mexique. Paris: Karthala, 2011; Schatz, Sara,  Murder and Politics in Mexico. Political Killings in the Partido de la Revolucion Democratica and its Consequences, New York  : Springer-Verlag, 2011. The homicides of perredistas (members of the PRD) were concentrated in four states (77% of the total): Michoacán (27%), Guerrero (25%), Oaxaca (15%), and Puebla (10%). It is interesting to see that, 30 years later, these same states “present the highest incidence of assassinations of political figures”.5Animal político, art. cit.. But the collection of cases included in my database shows additional interesting features beyond this geographic specificity, for they reveal that violence occurred during electoral periods at the local – not federal – level. This finding is consistent with observations by Álvarez Mendoza in the decade of 2010, and shows the importance of the 2018 elections in terms of bringing local contexts into a national light.

Records show that 265 PRD militants were killed between 1989 and 1994.

64% of the militants were killed in electoral contexts.

From 1989 to 1994, 64% of the militants in the sample were killed in electoral contexts: 10% during campaigns and 54% in post-electoral incidents, which were numerous in those years.6Eisenstadt Todd, Courting Democracy in Mexico: Party Strategies and Electoral Institutions, Cambridge : Cambridge University Press, 2004. The following extract illustrates the temporal distribution of violence during the electoral cycle, which needs to be studied as a whole:7Norris, Pippa. Why Electoral Integrity Matters. New York: Cambridge University Press, 2014.

[In] the municipality of Tixtla [Guerrero], half an hour from the highway to the state capital, it was believed (…) [that] the leader of a perredista committee, Guillermo García Tiledano. (…) was the first to be shot down, on November 27, 1989, a few days before local elections (…). This triggered indignation and [people] closed ranks. Then with the elections came fraud (…) People in Tixtla decided to seize and occupy the town hall [presidencia municipal] and name their own commissioners [comisarios] in each community. No one imagined this action would cost three more lives.8Fundación Ovando y Gil, p.11.

In the context of a democratic transition carried out at the national level, violence was clearly linked to resistance by local apparatuses of the Partido Revolucionario Institucional (PRI, Institutional Revolutionary Party) in various states, as they struggled against the construction of the PRD and the advent of electoral pluralism. Despite efforts to relate those homicides to “organized crime” (La Jornada, 7 July 1988), they were clearly products of political tensions.

This record of violence, however, goes beyond electoral cycles, for 6% of the perredistas were killed while serving as members of the Party’s municipal teams. Another 22% were murdered while performing routine Party activities or were identified as perredistas by their killers during the attack. In the case of Tixtla mentioned above, a PRD militant was killed as he left a Party meeting. The “aggressors walked through the town shouting their impunity [saying] they could kill as many perredistas as they wanted, since they were protected by the PRI”.9Fundación Ovando y Gil, p.12.

In that period, violence affected mainly men, likely because in the 1990s local political life was practically closed to women participation. This explains the latter’s low level of mortality (only 3% of the cases recorded). Event today, despite the clear feminization of local politics since then, however, in 2018 86% of victims were still men.10Álvarez Mendoza, p. 67.

The homicides of perredistas (members of the PRD) were concentrated in four states (77% of the total): Michoacán (27%), Guerrero (25%), Oaxaca (15%), and Puebla (10%).

During the study period – the 1990s – 55% of victims were active militants of the PRD, 5% were candidates for public office, 15% were Party officials, another 15% were popular representatives, and 5% were family members of militants. We found that 33% of these homicides occurred outside Party activities. While this fact could open the possibility that Party membership was a fortuitous, non-determinant, trait of victims and victimization, the context –recorded in circumstantial narratives presented to the CNDH– allows us to establish a causality between Party affiliation and homicide.

A Causality Between Party Affiliation and Homicide?

The histories of both political parties – the PRI with its semi-corporativist system; the PRD, which was formed at the local level, supported by social organizations –11Combes 2011. show that conflicts extended to social organizations and associated professional groups.

In the municipality of Chichihualco (Guerrero), for instance, conflicts after the 1990 municipal election became entangled with a conflict among the board of the Cattle-breeders Union (Unión Ganadera) that resulted in two deaths. Then, “violence returned to Chichihualco four years later (…) under the same signs of electoral fraud and public protests (…) and the PRD’s municipal committee was ousted from the town hall”.12Comité PRD Local.. Involved in those events was a conflict between transport workers (the Union of Democratic Taxi Drivers, a “movement of PRD cab drivers”) and “pirate” taxi drivers linked to the PRI. This second episode resulted in the deaths of 6 more militants.

Repression varied according to local political spaces. For example, more homicides occurred during demonstrations and marches in Guerrero and Oaxaca than in Michoacán.

Yet, in most cases in the sample, the link between homicides and militant activities is direct, for 13% occurred during Party meetings or as participants were dispersing. Another 9% of victims were killed while performing their functions as local representatives, and 8% in scenarios when town halls were occupied as a protest measure. Then, marches account for another 9%, and 6% took place at sit-ins (plantones). It is interesting to notice that certain forms of action – “repertoires”, to use the concept coined by the historian Charles Tilly – are met with greater repression than others. These include demonstrations (manifestaciones) and, above all, situations of “parallel governments” (municipios paralelos). In fact, in 1989, 23 of Michoacán‘s 113 municipalities, and 20 (of 75) in Guerrero, were occupied in this manner due to suspicions of fraudulent elections.13Rámirez Sevilla (Luis), Dibujo de sol con nubes : una aproximación a los límites y potencialidades del PRD en un municipio de Michoacán, Zámora, El Colegio de Michoacán, 1997.

In those cases, municipal administration was taken and assumed by teams organized by the PRD for periods ranging from a few weeks to a few months. Our sample reveals, moreover, that repression varied according to local political spaces. For example, more homicides occurred during demonstrations and marches in Guerrero and Oaxaca than in Michoacán, where victims were largely popular representatives (75% of cases). In Guerrero, deaths occurred more often outside Party activities, but the context of chronic violence in that state, exemplified by the two municipalities discussed above, made it possible to link them, as well, to tensions triggered by political pluralism.

Who Killed the Candidates?

The data collected provide some information on the (alleged) perpetrators of these crimes. While these data must be taken with caution because they do not come from formal judicial processes, it is interesting to examine the profile of the killers, as they were identified by people close to the victims (often eyewitnesses of attacks).

In 20% of cases, perpetrators were identified as “gunmen” (pistoleros); in 27% as police officers; and in 31% as PRI militants. In the remaining 19%, no “profile” was recognized.

In 20% of cases, perpetrators were identified as “gunmen” (pistoleros); in 27% as police officers; and in 31% as PRI militants. In the remaining 19%, no profile was recognized. These figures also fluctuate by state. The cases registered in Guerrero indicate heavy involvement by state police, while PRI militants stand out in Michoacán, and gunmen are more common in reports from Oaxaca. Moreovoer, in Guerrero, killers circulated in small groups of 2-4 individuals, and in 43% of cases they murdered several people on the same day (as occurred in Chichihualco). Then, in 7% of the cases, victims showed signs of torture, and in 2% their bodies had been mutilated. 

While one could argue that those close to victims would have incentives to point to their political enemies as the perpetrators, this possibility in no way weakens the argument I wish to establish. In fact, whether those political enemies were the victimizers or not, the descriptions speak eloquently of the political (and politicized) character of the violence of those years.

This brief account of the decade of the 1990s shows some continuities with 2018: first, the dimensions of electoral violence; second, that, broadly speaking, the same states are affected. These findings led us to question the chronologies of political-criminal recompositions in those territories, for the moment of greatest risk in the electoral cycle appears to have changed over time, from the post-electoral phase to the period when pre-candidates vie for their parties’ support (Álvarez Mendoza, 2019). This is related to the processes of transition, the organization of new work teams, and the periods of candidate selection, and indicates the need to ponder the degree to which intra-party processes may explain part of the violence.

The temporality of electoral violence seems to have changed. In the past, the post-electoral phase was the most dangerous. Today, it has evolved towards the period when pre-candidates vie for their parties’ support, and the moment of the campaign itself.

Then, this brief account of the decade of the 1990s shows some continuities and ruptures with 2018. First, the statistical relevance of electoral violence. Second, a relative geographic continuity in terms of states affected, which should lead us to question the narratives and the chronologies of political-criminal recompositions in those territories as they are being told. Yet, the temporality of electoral violence seems to have changed. In the past, the post-electoral phase was the most dangerous. Today, it has evolved towards the period when pre-candidates vie for their parties’ support, and the moment of the campaign itself.14Álvarez Mendoza, 2019. It seems that we have transitioned from the periods of political transition between two teams, to the periods of candidate selection.

Finally, while all parties were affected by violence in 2018, a certain continuity can still be detected. The parties most often victimized are those that form the opposition, as occurred with the PRD in the 90s. Yet, an analysis of the information available in the collections of “Political and Social Information” in Mexico City’s ‘Archivo General de la Nación’ (IPS, AGN) since the decade of 1940 identified that electoral violence in certain states – Guerrero, Oaxaca, and Puebla stand out– has been a long-lasting issue. At that time, however, violence affected mainly priistas (members of the PRI) themselves (or dissidents), especially during the candidate selection process. This indicates the need to pay more attention to whether intra-party processes may explain part of the contemporary violence.


If the lack of attention paid to electoral violence by analysts in the 90s can be linked to the national context – as I suggested at the outset – it is also clearly inscribed in the analytical frameworks that were being applied at that time worldwide, mainly inspired by “democratic transitions” theories. They largely set aside topics of local politics and electoral violence. Current work on electoral integrity, in contrast, has returned attention to violence as a significant variable in electoral cycles. This help explain the interest of analysts and Non-Governmental Organizations (NGO) in this field.

Yet, our research identifies a challenging issue that must be analyzed most rigorously: political-electoral violence since the 1910 Mexican Revolution. This enormous task would require cross-referencing diverse sources of information to elaborate an “X-ray” from the local perspective. This scientific rigor could lead us to question both the chronology of violence in electoral processes that has been associated with drug-trafficking, and the fact that today all types of violence tend to be attributed to criminal activity. Research of this kind will be especially important because a historical perspective may well show that today’s violence is, in many cases, inserted in long local histories.

Click here to discover the Elections & Violence in Mexico Project.

Click here to go back to the Mexico and Central America Program.