Making Sense of Electoral Violence: The Narrative Frame of Organised Crime in Mexico

Abstract

Since the inauguration of Mexican democracy in 2000, organised criminal violence had been leaking into the political arena. Yet, it escalated in the 2018 elections, when dozens of local candidates were killed. In most of these cases, the concrete perpetrators and motives remained in the dark. How did Mexican society make sense of this opaque, unprecedented wave of electoral violence? On the basis of a qualitative content analysis of over 1,200 news reports, I examine the structuring power of a shared narrative: the frame of organised crime. By conceiving candidate killings as economic violence within the criminal community, this commonsensical frame of interpretation permitted Mexican society to ‘normalise’ these killings as ‘business as usual’ by criminal organisations.

Spanish abstract

Desde la inauguración de la democracia mexicana en 2000, la violencia del crimen organizado se fue filtrando a la arena política. Ahora bien, ésta escaló en las elecciones de 2018, cuando docenas de candidatos locales fueron asesinados. En la mayoría de casos, los perpetradores y motivos concretos se mantuvieron en la oscuridad. Cómo le dio sentido la sociedad mexicana a esta oleada turbia y sin precedentes de violencia electoral? En base a un análisis cualitativo de más de 1.200 informes de prensa, examino el poder estructurante de una narrativa compartida: el enmarcamiento del crimen organizado. Al concebir los asesinatos de candidatos como violencia económica al interior de la comunidad criminal, tal marco de interpretación común permitió a la sociedad mexicana ‘normalizar’ estos crímenes como ‘actos cotidianos’ de parte de las organizaciones delictivas.

Portuguese abstract

Desde a inauguração da democracia mexicana em 2000, a violência do crime organizado transbordou para a arena política. No entanto, aumentou nas eleições de 2018, quando dezenas de candidatos locais foram mortos. Na maioria desses casos, os perpetradores e motivos concretos permaneceram no escuro. Como a sociedade mexicana deu sentido a essa onda opaca e sem precedentes de violência eleitoral? A partir de uma análise qualitativa de conteúdo de mais de 1.200 reportagens, examino o poder estruturante de uma narrativa compartilhada: o quadro do crime organizado. Ao conceber os assassinatos de candidatos como violência econômica dentro da comunidade criminosa, esse quadro de interpretação de senso comum permitiu que a sociedade mexicana ‘normalizasse’ esses assassinatos como ‘negócios comuns’ das organizações criminosas.

Introduction

Since the closing days of its authoritarian regime in the late 1990s, Mexico has been living through a strange kind of civil war, the so-called ‘drug war’, a protracted lethal conflict among an ever-evolving multiplicity of armed business enterprises (‘drug cartels’). Over the past two decades, this ‘criminal war’¹ has claimed around a quarter of a million lives. While violence had been leaking into the political arena before, it rose to unprecedented levels in the 2018 general elections, when dozens of local candidates were killed. In most cases, the concrete perpetrators and motives remained in the dark. No one claimed responsibility and police investigations came to nothing. How did Mexican society make sense of this opaque, extraordinary wave of electoral violence?

In Mexico, there has been little controversy about the nature of the ‘narcoviolence’ that has been besieging the country. Early on in the war, a ‘narrative frame of organised crime’ has become dominant that serves both private citizens and public actors to comprehend violence as an income-maximising strategy among illegal business associations. It attributes violence to a set of quasi-extra-societal actors (‘los narcos’) who kill their competitors (as well as their defectors) in armed struggles for illicit markets. Arguably, this interpretative frame has contributed to ‘normalising’ the epidemic of violence. In the face of systemic opacity and impunity, it resolves the existential question of who kills whom and why in a way that allows ordinary citizens to keep violence at a symbolic distance. Since it conceives perpetrators as well as victims as organised criminals, rather than ‘decent citizens’, it offers an account of organised death that relieves ordinary Mexicans of risk and responsibility.²

Under conditions of factual uncertainty, societal actors cannot cope with political events by simply ‘relying on the facts’. They need to fill the holes in their common knowledge through shared assumptions about the reality they face. On the basis of a qualitative content analysis of over 1,200 print media reports, I examine the structuring power ‘the narrative of organised crime’ developed in the face of escalating electoral violence. That is, I examine its capacity to turn extraordinary acts of opaque political violence into meaningful, ordinary events. To what extent, I ask, did it permit Mexican society to ‘normalise’ candidate killings as ‘business as usual’ by criminal organisations? More concretely, to what extent did it shape the attribution of responsibility for electoral violence, the presumption of guilt of its victims, and the deflection of suspicion from political actors (‘depoliticisation’)?

My frame-analytic approach illuminates the dependence of electoral violence on social perceptions. Prevailing factual beliefs about its protagonists and their motives are not mere reflections of reality, as they cannot possibly be since core parts of reality itself are unknown. Rather, they are joint constructions of societal observers who mobilise available interpretative resources to make sense of the uncertain realities they face.

In the following, after sketching the intrusion of lethal violence into the Mexican 2018 elections, I outline the substantive contours of what I identify as ‘the narrative frame of organised violence’. I then describe the profiles of the 48 slain candidates and the dataset of news reports that serves as my empirical basis for qualitative content analysis, discuss the opacity of Mexico's electoral violence, and sketch plausible alternatives to the dominant interpretative frame. In my content analysis, I reconstruct three broad types of frame effects on public reports of electoral violence: the attribution of responsibility to organised crime; indirect efforts to dispel presumptions of guilt towards victims; and the deflection of responsibility from political actors.

The Irruption of Electoral Violence

Since its official inauguration with the 2000 alternation in presidential power, Mexican democracy has been sliding into the so-called drug war, an ‘economic civil war’ in which ever-changing criminal organisations have been battling among themselves as well as against the state and civil society.³ While lethal violence forms only the tip of the iceberg of cruelty and destruction, the war's death toll has been staggering. During the first two democratic decades, an estimated 180,000 homicides and 70,000 disappearances have been attributed to the conflict.⁴

Organised criminal violence has also produced a steady stream of political victims. To cite just some fragmentary data: 92 journalists were assassinated from 2009 to 2019,⁵ 178 local officials from 2004 to early 2018,⁶ 63 human-rights activists during the first five years of the Felipe Calderón government (2006–11),⁷ and 14 environmental activists in 2018 alone.⁸

During the 2018 elections, violence against political actors escalated. In a historical first of electoral synchronisation at all levels, over 18,000 elected positions were at stake: the federal presidency, both national legislative chambers, nine governorships, 27 local legislatures and 1,612 city councils.⁹ During the ten months preceding election day on 2 July 2018, 104 elected officials and 48 candidates were assassinated.¹⁰ In some countries, as in Pakistan and the Philippines, electoral violence is endemic, and it is habitual to see elections produce high numbers of lethal victims.¹¹ In Mexico, however, electoral violence had been rare in the early 2000s and no more than episodic after the escalation of the drug war under President Calderón (2006–12).¹² The scale of electoral violence in 2018 was unprecedented. It turned the multi-level contest into ‘the most violent election in Mexico's modern history’.¹³

The Frame of Organised Crime

Even though the creation of multiple private armies by drug cartels in the 1990s had sown the seeds for the epidemic of violence to come, nobody had foreseen the catastrophe.¹⁴ Its rapid descent into the everyday horrors of a ‘violent democracy’¹⁵ took Mexican society by surprise. It has been struggling since then with how to understand ‘the hell’ it found itself dragged into.¹⁶ The public vocabulary of violence has been unstable and contested. Academic analyses and media reports commonly refer to ‘drug violence’, ‘organised crime’ or ‘organised violence’. Abstractions like ‘violence and insecurity’ are common and so is the language of war: ‘the drug war’, ‘the criminal war’, ‘the war among cartels’.¹⁷ However, beneath conflicting conceptual and terminological choices, we can discern a common narrative that has emerged early on in the war, a dominant frame of interpretation that conceives certain types of violence as routine activities by criminal business firms.

Frame analysis is ‘a vibrant field within the social sciences’.¹⁸ The bulk of framing research in the social sciences studies the ways in which political elites (mass media, governments, parties and candidates) present societal problems, such as child abuse or obesity, or policy issues, like nuclear energy or monetary politics. Policy frames are ways of conceptualising political issues. They define problems, causes and remedies, weigh normative trade-offs, assign responsibility to actors, and assess their moral worth as well as their empirical constraints and opportunities.¹⁹

In Mexico, what I propose to conceive as ‘the frame of organised crime’ serves to make sense of the bewildering intrusion of senseless violence into national life. It delineates a category of violence that is neither anomic nor incomprehensible but calculating and strategic. At its core, it defines a field of actors and their relations: perpetrators, victims and politicians.

(i) Rational perpetrators. The frame of organised violence pictures its protagonists as rational barbarians, employees of ruthless and faceless business firms who use physical violence as a means of maximising illicit profits and rents. They are neither revolutionaries nor psychopaths. Their cumulative mass murder (one by one) obeys the impersonal, economic imperatives of competitive illicit markets.²⁰ They pursue their business in recognisable ways, either by murdering their victims with assault rifles in public or by kidnapping and disappearing them.²¹

Most scholarly explanations of lethal violence against political actors in contemporary Mexico have embraced this motivational assumption of entrepreneurial rationality as well. To decipher the drivers of political homicides, they reconstruct the expected utility calculus of ‘rational actors’ who engage in violence as a ‘routine activity’ in the pursuit of material gain.²² The list of potential economic payoffs of political violence is large. Violence against local officials, for instance, promises to give criminal organisations control over local law enforcement and access to public revenue.²³

(ii) Suspicious victims. According to the prevalent understanding of ‘drug violence’ in Mexico, most of it is ‘internal’ to the criminal community. It is neither ‘insurgent violence’ against the state nor ‘predatory violence’ against citizens but ‘competitive violence’ that takes place either within or between criminal organisations.²⁴ Hence, most of its victims are self-selected and thus self-responsible: they get killed because they were part of the criminal underworld. They are guilty, or at least suspicious, unless proven innocent.²⁵ The presumption of guilt applies to political victims of organised violence as well. Invariably, there are two possible explanations for their murder: they may have been targeted because they ‘refused to establish pacts’ with criminal groups or because they were ‘allied with rival criminal groups’.²⁶ Almost invariably, available evidence does not permit adjudicating between these competing hypotheses, so that a shadow of doubt falls over victims.

(iii) Subordinate politics. The relations between criminal organisations and Mexican politics (governments, parties and state agents) are complex and dynamic. Neither side is homogeneous or unitary, both are fractured and multi-layered. Between them, some relations are confrontative, others cooperative, some stable, others fragile. The logics of war, collusion and conflict avoidance coexist, shift and intersect.²⁷ Still, both political and academic observers are likely to agree on one structural property of the conflict: the independent agency of criminal organisations.

Up until the 1990s, the cartels served as subordinate partners to political patrons who protected them from prosecution and competition.²⁸ Yet, as these protection rackets broke down in the wake of democratisation and alternation in state governments, the cartels ‘emancipated’ themselves by creating their autonomous military capacities.²⁹ Their transformation from dependent bandits into ‘warring oligarchs’³⁰ did not lead them to renounce their ties with the state, though. Rather, after each breakup of established alliances with public actors, they strive to recapture state agencies through the time-tested combination of corruption (‘plata’) and violence (‘plomo’). In contrast to the heyday of hegemonic authoritarianism, it is not civilian authorities anymore, but criminal organisations, who initiate and control these relationships of ‘coercive corruption’.³¹ The dominant narrative of organised crime is a narrative of organised-crime dominance.³²

If, as I posit, the ‘master frame’ of rational criminals, voluntary victims and passive politicians has served Mexican society to render the war-like epidemic of violence intelligible (as well as bearable), did it serve as well to ‘make sense’ of the wave of lethal violence that swept through the 2018 general elections? Or did the shocking irruption of electoral violence crack, or even break, established ontological certainties about the nature of violence?

The Coverage of Candidate Killings

To examine the public treatment of lethal violence against candidates in the 2018 elections, I analyse news reports on all 48 cases registered in the dataset compiled by Víctor Hernández (with a few light name and date corrections).³³ The victims were assassinated between September 2017 (the official start of the federal electoral period) and election day on 2 July 2018. Of the victims, 17 were so-called ‘pre-candidates’ competing for their parties’ nomination and 29 were ‘official’ candidates already chosen by their parties.³⁴ Table 1 provides the full listing.

Table 1. Murdered Candidates in Mexico's 2018 Elections

Notes: ^a Status: P = pre-candidate, C = candidate; ^b Coverage: C = central, P = peripheral, S = succession, Total = sum of articles.

Source: Hernández, ‘Candidatos asesinados en México’ (2020); author dataset of print news on candidate assassinations.

As Table 1 shows, with one exception (Fernando Purón, who ran for a seat in the Federal Chamber of Deputies), all of them competed for local offices: 28 aspired to be mayors, eight hoped to be municipal councillors and nine aimed to be state legislators. In terms of party membership, 12 of the murdered candidates belonged to the ruling Partido Revolucionario Institucional (Institutional Revolutionary Party, PRI); ten to the left-wing Partido de la Revolución Democrática (Party of the Democratic Revolution, PRD); seven to the conservative Partido Acción Nacional (National Action Party, PAN); six to Andrés Manuel López Obrador (AMLO)'s Movimiento de Regeneración Nacional (National Regeneration Movement, Morena); and all others to minor parties. They were aged between 23 and 65, with a mean of 44 years (N = 38). Most of them were men (83.3 per cent). In geographical terms, the state of Guerrero took the brunt of lethal electoral violence (14 cases or 29.2 per cent), followed by Michoacán, Oaxaca and Puebla (5 cases each) and Jalisco and the State of Mexico (4 cases each).

The ‘universe’ of reports I collected consisted of all articles that contained the name of individual victims and that were published between the day of their assassination and the subsequent three months in any of the print media included in the digital archive EMIS API Infolatina,³⁵ which holds major national as well as a broad array of local Mexican newspapers. The ensuing dataset comprises 1,263 news reports, an average of 26.4 per case. Table 2 shows the distribution by news sources. Given its breadth, I expect this article sample to provide a comprehensive picture of published societal responses (by public officials, security agents, party politicians, and friends and family) to individual candidate assassinations (as covered by either national or local print media).³⁶

Table 2. Distribution of Articles by News Sources

Source: Author dataset of print news on candidate assassinations.

Less than half of these reports (542 or 42.8 per cent), though, centre their attention on the case at hand: the act of killing, the investigations, political and private responses, tributes, and funeral proceedings. All others (699 or 55.2 per cent) mention the victim in passing only. Such ‘peripheral’ notes often include the assassination as a mere exemplar in a larger string of similar events. A small portion of articles (25 or 2 per cent) report on debates or decisions on successor candidates. Given the syndicated nature of numerous local newspapers, the dataset contains numerous perfect or near repetitions. Searching primarily for similarities, rather than frequencies, I excluded these repetitive notes from my analysis.

The last four columns of Table 1 show the highly unequal coverage of individual cases. The Infolatina database does not contain any reports on the murder of six candidates, yet over 20 ‘central’ notes and over 30 ‘peripheral’ ones regarding about a dozen of the cases. Figure 1 traces their distribution over time (days after assassination). Like most reported events, candidate assassinations in 2018 had a short news cycle. Almost two-fifths of all central articles appeared in the first three days after the event (423 or 78 per cent).

Figure 1. Temporal Distribution of News Reports on Candidate Assassinations

Source: Author dataset of print news on candidate assassinations.

Frames shape perceptions of facts. Thus, when we wish to establish their interpretative weight, the first question concerns relevant facts. When facing the 2018 wave of electoral violence, did Mexican society enjoy a high-information environment of known perpetrators and motives? How transparent was anti-candidate violence, or how opaque and thus dependent on interpretative frames?

The Opacity of Electoral Violence

Political elections are a civilising device. They open peaceful roads to state power. Still, a significant minority of elections worldwide, some of them democratic, others authoritarian, are afflicted by violence.³⁷ Conventional definitions of electoral violence include its timing and its targets. It takes place during elections (before, during or after election day) and targets the electoral process (its actors, institutions or infrastructure).³⁸ Though its goals may be systemic (an attack against elections by their ideological enemies), most electoral violence is competitive (an attack against electoral contenders by their adversaries). Its ‘main organisers [are] political parties’³⁹ in the pursuit of ‘electoral advantage’.⁴⁰ Incumbents tend to use it as a tool of exclusion, opposition parties as a tool of mobilisation.⁴¹ However, since violence is an illicit strategy of electoral competition, its use is ‘typically not widely advertised’.⁴² Its masterminds often seek to maintain deniability by delegating it to ostensibly non-state, non-partisan agents, like groups of thugs, youth groups, rebel groups and private militias who preserve varying degrees of autonomy from their political sponsors.⁴³ Identifying the authors of electoral violence is therefore notoriously difficult and often shrouded in controversy.⁴⁴

In ‘criminal wars’ like the Mexican one, where presumably non-ideological ‘violent entrepreneurs’⁴⁵ compete for material gains, not partisan advantage, problems of attribution are even more acute. As a rule, criminal organisations exercise ‘violence without clear political goals’.⁴⁶ They lack political affiliations or programmatic affinities with political actors. They do not serve as armed wings of political parties. When they succeed in infiltrating or capturing political parties or state agencies, those involved would not usually advertise their alliance, but do everything to keep it secret. Furthermore, unlike terrorist groups or guerrilla fighters, criminal organisations seldom issue public claims of responsibility for the political assassinations they carry out. The high ‘visibility of criminal political partnerships’ that Juan Guillermo Albarracín found in Brazilian cities,⁴⁷ where political entrepreneurs establish ‘overt electoral alliance[s]’⁴⁸ with criminal groups who grant them electoral, social and armed assistance, is almost unthinkable in Mexico.

In general, the Mexican ‘drug war’ unfolds within a thick fog of ignorance and impunity. Since about 90 per cent of all homicides remain unresolved, concrete knowledge about perpetrators and their motives is scarce.⁴⁹ Did the irruption of violence into the electoral arena produce a systemic shock that pushed the police to dissipate ‘the fog of criminal war’⁵⁰ and clarify these political crimes? As far as media reports allow to infer, the answer is negative. Within three months after the murder, print media informed of the identification and detention of suspects in ten of the 48 registered cases, a little over one-fifth (20.8 per cent). In five of them, authorities declared that they were able to identify the concrete criminal organisations behind the murder of candidates but would refuse to name them.

In only three cases (6.2 per cent) did the identification of suspects allow for clarification of the motives of the crime. One was a story of resistance, in which a candidate (Purón) was killed by criminal organisations which he had confronted, and which had publicly announced his assassination two years before. One was a story of ordinary crime, in which a candidate (Addiel Zermann) was apparently killed in a botched burglary attempt. And one was a story of agency rebellion, in which a candidate (Maribel Barajas) was killed by the woman she had hired to kill the ex-girlfriend of her partner. In all other cases (93.7 per cent), three months of crime investigations did nothing to dispel the initial uncertainty about who killed the candidates and why. No case was brought to court.⁵¹

In the absence of shared information, actors need shared interpretative frames to make sense of emergent events.⁵² To what extent did the available frame of organised crime allow Mexican political society to ‘make sense’ of the intrusion of lethal violence into electoral competition, despite common ignorance about concrete motives and perpetrators?

Common Sense and Its Alternatives

Students of political frames are generally interested in frame variance, in strategic choices among competing frames by elites or counter-elites. They do not study frames as cultural givens but as strategic constructions by collective actors like mass media, electoral campaign teams, or social movements. In contradistinction, Erving Goffman, in his foundational Frame Analysis, did not conceive frames as elite projects or as choices, but as conventions.⁵³ Instead of conceiving them as ‘alternative ways of defining issues’,⁵⁴ he took them as ‘the basic frameworks of understanding available in our society for making sense out of events’.⁵⁵ In his perspective, frames are commonsensical ‘principles of organisation’ of experience that allow actors to reach ‘seeming agreement’ about the ‘social reality’ they inhabit.⁵⁶ Rather than alternative modes of presenting shared realities, they are ‘conventional understandings’ that are constitutive of shared reality, allowing us ‘to cope with the bizarre potentials of social life’.⁵⁷ In ordinary life, they provide answers to the question: ‘What is going on here?’⁵⁸

My treatment of ‘the frame of organised crime’ embraces Goffman's notion of frames as commonsensical ‘definitions of a situation’.⁵⁹ I wish to reconstruct narrative common sense rather than narrative choices.⁶⁰ However, to assess the strength of this hypothetical frame, that is, the extent to which it served to render the eruption of electoral violence intelligible, we need a sense of plausible alternatives. Which might be competing stories of candidate killings? Without pretensions of completeness, we might consider at least five alternative accounts of actors and motives:

1. (i) Ordinary crime: ‘disorganised’ crime perpetrated by individuals for either economic or personal motives.

2. (ii) Irrationality: the commission of crime by ‘crazy’ individuals who suffer from mental illness (‘psychopaths’) or ‘evil’ people who suffer from moral disabilities (‘sociopaths’).

3. (iii) Structural causes: the attribution of crime, not to actors, but structural conditions, like poverty, the dissolution of communal bonds, the demand for illicit goods, or the availability of weapons.⁶¹

4. (iv) Ideology: a violent campaign by enemies of democracy who wish to derail electoral processes.

5. (v) Political competition: the elimination of candidates by political contenders who seek electoral advantage.

The latter possibility, the suspicion that candidate assassinations might serve as tools of electoral competition, is the most disquieting. It shifts the locus of agency from criminal organisations to political actors. Rather than assuming that criminals kill politicians for economic purposes, it surmises that politicians are hiring criminals for political purposes. And it implies that political contenders violate the most fundamental norm of democratic conflict settlement: the renunciation of violence. In general terms, the hypothesis of competitive violence appears improbable, yet not impossible.

Surely, Mexico's criminal groups are not ideologically aligned with political parties. And, clearly, since candidates from all parties fell victim to violence, there was no single party waging a national campaign of violence against all others. However, while supporting a presumption of innocence, the non-ideological orientation of drug cartels and the political heterogeneity of their victims did not foreclose the possibility of political complicity. Cartels do not share ideological affinities with political parties, but they do not have any ‘ideological constraints’⁶² either. And the absence of a nationwide pattern of candidate killings did not preclude the possible use of violence by electoral contenders at the local level.

Still, even in the unlikely event that a political party knew with relative certainty that organised criminals had ordered and executed the assassination of one of its candidates, it would be left wondering about the relationship between the perpetrators and its political adversaries. Criminal authorship is compatible with multiple motives. A ‘cartel’ may kill a candidate because it does not like him or her, or because it likes other candidates more. Regardless of its concrete motives, though, the electoral contenders of its victim are likely to benefit from the ensuing reduction of electoral competition. In a context where violence has turned into a widely available ‘social resource’,⁶³ the crime cries out for clarifying their role. Did they possibly arrange, encourage, sanction or tolerate the murder? Given the fact that the boundaries between private and public criminals are often blurred and porous in Mexico,⁶⁴ it seems hard to discard, without further inquiry, the possibility that ‘criminal gangs [were] allied with political actors’.⁶⁵

In fact, both before and during the 2018 elections, some scholars issued rather bold claims about the partisan nature of electoral violence. Laura Ross Blume read ‘narco-assassinations of Mexican politicians’ as ‘evidence of an alarming and persistent pattern in Mexico of politicians enlisting criminal organisations to eliminate their political competition’.⁶⁶ The consulting firm Etellekt described the 2018 wave of electoral violence as a ‘strategy of the State’ directed ‘against opponents of the governing parties or coalitions’ at the state level.⁶⁷ Yet, beyond simple correlations of political killings with vertical divisions of power (opposition control of municipalities), neither of the two provided empirical evidence for their provocative assertions.⁶⁸ Nevertheless, even if distressing, implausible and unsupported by evidence, the hypothesis of competitive violence among political contenders in 2018 lay within the realm of the possible.

In the remainder of this article, I will assess the strength of the frame of organised crime in public responses to candidate killings against these alternative narratives. Striving to not just discern the mere presence of the frame but also document its concrete manifestations in a precise and nuanced fashion, I coded my sample of news reports⁶⁹ by five main thematic categories: the attribution of blame, the denial of guilt, the communication of injustice, the communication of innocence, and the political logic of violence. I then dissected the microstructure of the resulting thematic text fragments, coded their granular analytic components, and arranged the corresponding quotes in extensive tables of structured primary material. While my dataset contains the raw material of my analysis (the full text of news reports), these quotation tables put the entirety of my empirical, textual evidence on display. For the purpose of methodological transparency, they are available in full in the online Appendix. In their depressing repetitiveness, they make for fascinating readings.⁷⁰

In addition to tracing statements that indicate the presence of the dominant frame, I screened news reports on candidate assassinations for deviations from it, above all, by searching for instances of blame attribution to partisan actors (mutual suspicions among adversaries). In doing so, I do not wish to suggest that political contenders, rather than criminal organisations, were the authors of candidate assassinations. In the absence of reliable judicial, journalistic and scientific evidence, any such claim would be frivolous. I do no more than trace public expressions of political suspicion. I have no ambition to adjudicate the authorship of these homicides.

Besides, striving to trace discursive consequences of ‘the frame of organised violence’, I remain silent about its sources. I discuss neither its empirical plausibility nor its possible persistence through fear. Both media and political actors may refrain from inquiring into political complicities with candidate assassinations to avoid becoming victims of criminal violence themselves. Just as fear of criminal punishment has shaped civilian responses to organised violence⁷¹ as well as prevalent reporting routines by the media,⁷² it likely informs responses to and coverage of electoral violence as well.⁷³

The Attribution of Blame

At the peak of media attention, in the days after the murder, everybody knows: a candidate was killed. Outside the realm of perpetrators, however, nobody knows who did it and why. How do political actors (public officials, party politicians and the social environment of the victim) deal with these factual uncertainties about concrete authors and motives of candidate assassinations? What I propose to conceive as the ‘narrative frame of organised crime’ allows them to ignore their collective ignorance. It distributes blame among three sets of actors: criminal groups, the state and victims. First, it offers a theory of criminal authorship that attributes responsibility to abstract forces: criminal organisations. Second, by blaming private actors, the frame exempts public actors from direct responsibility; the state appears responsible for crimes of omission, yet not for those of commission. Third, while creating abstract certainty about perpetrators, the frame sheds a shadow of suspicion on victims: they may have been killed because they were involved in criminal activities.⁷⁴

Blaming Crime

One might expect that actors who are habituated to the presence of violent criminal organisations would name these groups as authors of electoral assassinations. But no, not primarily. Rather, they tend to move to a higher level of abstraction and cite the prevailing context of ‘violence’, ‘insecurity’ and ‘crime’ to account for candidate assassinations. In their efforts to make sense of these criminal acts, they refer to ‘violence and crime’ (774),⁷⁵ ‘the violence in the country’ (626), ‘the insecurity we live in’ (19), ‘the situation of insecurity’ (394) or ‘the security situation’ (772) as objective causal conditions. Sometimes, they grant explanatory roles to local contexts of hyperviolence: the crime happened in ‘the most violent town of the country’ (384) or ‘the most violent state of Mexico’ (1189). Knowledge about generic circumstances substitutes for knowledge about concrete responsibilities.

Speakers often describe ‘the violence’ like a natural phenomenon whose origins, reproduction and growth follow autonomous logics beyond the control of actors. Like a ferocious dog, ‘violence was unleashed’ (232) and has ‘gotten out of control’ (1130). The metaphor of a ‘violent wave’ (286) or ‘wave of violence’ (895) is recurrent: a natural catastrophe that floods the country. Others, by contrast, attribute agency to ‘violence’ as if it were a purposeful actor: ‘the wave of violence in our country has taken another life’ (191), it ‘has raged against candidates’ (1239) and ‘lashed out against small towns’ (552).

Talk about general ‘insecurity’ is similar. It has its own expansionary dynamic. Once ‘unleashed’ (208), it ‘has been growing alarmingly’ (36), ‘in a terrible manner’ (772). And despite its abstract nature, it kills in person: ‘the insecurity has left about 24 politicians dead’ (552), it produces lethal ‘victims’ (1087), we mourn ‘a loss caused by the insecurity’ (206). In similar anthropomorphic language, the abstract phenomenon of ‘crime’ appears as the great executioner who ‘exacts … human lives’ (298), chooses its victims (‘crime targets municipal powers’ (465)), plans strategically (‘criminal hands act in premeditated, despicable manner’ (1215)) and overwhelms actors on the ground (‘criminality … has surpassed us’ (206)). The individual actors behind the aggregate trend, ‘the criminals’, the ‘masters of life’ who ‘commit crimes with complete impunity’ (3) and ‘rob us of our tranquillity’ (1331), commonly remain outside the field of vision.

The vocabulary of violence, insecurity and crime does not draw an explicit distinction between ‘ordinary crime’ and ‘organised crime’. Given the ordinary nature of organised crime, speakers may not feel the need to draw a distinction. Or they may feel safer not making it explicit. Still, numerous statements do cite the ‘presence of organised crime’ (19) as an explanatory factor of candidate assassinations (516). They either mention the abstract phenomenon of ‘organised crime’ (36 quotes) or refer to general categories of criminal collective actors: criminal ‘groups’ (56 quotes) or criminal ‘gangs’ (7 quotes), though not ‘organisations’ (1 quote). They are the dominant actors in a field of lethal abstractions. It is them who ‘dye this electoral process in blood’ (649) and ‘keep sowing fear’ (473).⁷⁶

In a few candidate killings that occurred in so-called ‘hot zones’ (1201), media reports would go beyond the abstract invocation of ‘criminal groups’ and name concrete organisations who were known, or believed, to dispute the territory in which the victims had been pursuing their political careers: Jalisco Nueva Generación vs. Nueva Familia Michoacana (Stalin Sánchez), Los Rojos vs. Los Ardillos (Francisco Tecuchillo), or La Línea vs. Gente Nueva (Liliana García). At times, such local contextual explanations cite previous strings of assassinations in the personal or political environment of the victim to add plausibility to the authorship of ‘criminal organisations’ (27 and 688).

Blaming the State

In a world where lifeless abstractions and faceless corporations take human lives, the anger and frustration of friends and family understandably turn against those actors and institutions that do have identifiable names and responsibilities: governments and state agencies. In the face of lethal violence, it is clear that they have failed their basic duties of protection. Hence, again and again, mourners accuse both elected and non-elected public officials of crimes of omission. They designate governments as ‘co-responsible’ (823) or even ‘solely responsible’ (774) for lethal violence. They denounce the ‘absolute failure’ (771) of security policies, their inefficiency (‘we want results’ (304)), the ‘incompetence and irresponsibility’ (304) of public officials, their self-complacency (298). Their clamour is desperate: ‘the Mexican State must assume its responsibility to grant security to all citizens’ (1130).

Blaming Victims

Media reports on candidate assassinations contain scant reflections on the concrete reasons that might have motivated them. In a handful of interviews on electoral violence, experts speculate on its general causes. Commonly, they just state the obvious: cartels try to exert political influence. Criminals who use violence as a ‘mechanism of candidate selection’ (1193) attempt to ‘control’ municipal governments (63). General hypotheses, however, cannot explain specific cases. ‘Who is killing pre-candidates and why? This is the crucial question’ (471). ‘[N]o one knows whether [they] ran into problems for collaborating with some criminal group or for refusing to do so’ (19). In the end, though observers may have ‘no doubt whatsoever’ that these assassinations derive from strategic ‘calculations of organised crime’ (1208), the concrete logic of victim selection remains unclear in all except a few cases. While the narrative frame of organised crime provides certainty about perpetrators, it creates uncertainty about their motives and thus the status of their victims (innocence or guilt).

Of the 48 murdered candidates, only three got cleared of criminal entanglements with organised crime: Zermann (C32), victim of a derailed robbery attempt, Purón (C41), victim of his year-long resistance against organised crime, and Barajas (C29), victim of her own criminal ambitions. In three other cases, by contrast, either criminal investigations or media narratives suggested a ‘guilty’ verdict. José Aguirre (C36) had stood at the centre of a criminal network. Two other candidates attracted suspicions because of their personal closeness to known criminals: political superiors in one case (Antonia Jaimes, C23), the father in another (Pamela Terán, C40). At the latter's funeral, her grieving family members did little to dispel suspicions about their criminal entanglements as they launched dramatic threats of vengeance: ‘We will finish those damned dogs, one by one. Now the massacre begins!’ (963). All other 42 slain candidates (87.5 per cent) remained vulnerable to doubts about their proximity to crime.

Dispelling the Presumption of Guilt

By creating not just clarity about perpetrators (in the abstract) but also suspicions about victims, the frame of organised crime compels those who were close to the victims to defend their innocence. Explicit denials of guilt are rare, though. In fact, we observe them only in response to explicit claims of guilt. For instance, family and friends of Aguirre (C36) forcefully rejected informational leaks about his involvement in the organised theft of petroleum (‘huachicoleo’). They portrayed him as ‘an honourable man, a family man, a well-regarded person in his community’ (790) and protested against his ‘criminalisation’ (771), the diffusion of ‘information about unconfirmed facts’ (771), libellous efforts to ‘stain the image, the memory, and the honour of a person who dedicated his life to his brothers and sisters’ of his hometown (788). As they stated, ‘it is not fair … to revictimise the victims’ (790). Cognisant of the general tendency to blame the victims of organised crime in Mexico, some mourners would defend their victims pre-emptively: ‘We will not permit that [the dead candidate] shall be simply and plainly criminalised as happens to thousands of victims in this country because authorities are incapable of granting justice’ (94); ‘without doubt, once again they will want to invent a story to justify his murder’ (1192).⁷⁷

Most of the time, though, both the hypothesis of guilt and the defence of victims’ decency remain implicit. In the absence of solid evidence, making the hypothesis explicit would be unfair to the victims, while making explicit claims of innocence would be ineffective towards the sceptics. So, how can actors resist, break or at least soften the suggestive power of the frame of organised violence? How can they persuade neighbours and readers that the candidate who fell victim to organised crime was nevertheless innocent of his fate?

Charges of Injustice

When members of criminal organisations die by the hands of criminal colleagues, it is inviting to conceive their death as part of a ‘just world’.⁷⁸ They share responsibility in their own fate. They appear as voluntary victims who accepted certain professional risks when entering the criminal world and, in the end, suffered the consequences. Their ostensible consent in their own death diminishes its injustice. As they themselves accepted its possibility, others may well accept its reality. Accordingly, Mexican citizens often find it easy to meet news about narco-style killings with a shrug: ‘surely, the guy was into something’. The unnatural death of ‘a criminal’ does not appear a tragedy but a natural event, a simple empirical regularity: what else would one expect?⁷⁹

Therefore, if supporters of victims want to counteract tacit suspicions of their criminal involvement, the first thing they need to do is to persuade the world that the murder constitutes an act of injustice. If their fellow citizens might celebrate the death of a criminal as an act of higher justice, or shrug it off as a professional accident, those who mourn the death of an innocent victim need to communicate that they are responding to a disruption of the just world, to an act of injustice that upends fundamental expectations of fairness. As the extensive lists of quotations in the online Appendix show,⁸⁰ this is exactly what they do. Instead of keeping their distance, as they might towards someone considered ‘guilty’, actors offer their condolences, sympathies and solidarity to the family and friends of the victim. They express their feelings of shock, pain and sadness. They publicly lament the assassination, condemn it, express their indignation.

Too, over and over, they demand justice: expedient and exhaustive investigations, the establishment of truth, the identification and capture of the murderers and their masterminds, and their punishment. Their litany of repetitive demands is nothing less than depressing. In all likelihood, people make them in full knowledge that they are as good as condemned to failure. Their endless, repetitive nature indicates their structural ineffectiveness. In the end, their only value seems to be symbolic: the quest for justice reaffirms the injustice of it all.

At times, the supporters of killed candidates do more than declare their consternation and call for justice. They take collective action to underline their just anger. In various cases, they suspended campaign activities for some period. Exceptionally, their constituents gathered in public places to lend weight to their clamour for justice. In one case, local residents threatened with armed rebellion if authorities failed to capture the murderers.⁸¹

Claims of Innocence

As stated above, when honouring murdered candidates, speakers rarely offer explicit claims of innocence or express denials of criminal involvement. Often, though, they affirm their innocence in more subtle and implicit manners.⁸² Frequently, they describe victims as members of a larger community of innocent victims: ‘all those families who weep for their disappeared, violated, and murdered children’ (194), ‘the thousands of Mexicans who live with this form of violence’ (201), ‘the numberless list of victims of citizens who have lost their lives due to the uncontrolled violence’ (771). They express the inexplicable nature of the killing of someone innocent by asking for explanations: ‘Why my daddy?’ (343), ‘We are hurt, indignant, Addiel did not deserve to die this way. Why did they attack him with such fury? Why would they do that?’ (653). To signal the distance of the victim from the criminal world, numerous statements mention the absence of previous threats, which renders the killing enigmatic as it rules out both involvement and resistance: ‘he had never received any threats’ (1218), ‘his family has not received any threats’ (343). Repeatedly, speakers formulate such claims with epistemic caution. They point to the uncertainty of public knowledge and specify that they lacked reports (16), alerts (191), complaints (314), knowledge (1215) or information (1217) on possible threats.

Signs of Innocence

When someone dies, it is bad taste to speak badly about them. When public figures pass away, those who survive them tend to pay them tribute and honour them as exemplary people whose death constitutes a significant loss. In the context of institutionalised suspicions against victims, such eulogies acquire additional significance. They turn into counter-narratives to the notion of the criminal victim. In numerous news reports, speakers portray slain candidates as people of admirable personal virtues and high social esteem (in their private as well as their public lives). These portrayals carry a tacit message: The victims did not belong to the criminal world; they were good people who belonged to our world, the world of decent citizens.

Private virtues: Against the idea of the criminal who seeks quick money instead of earning his livelihood with good, honest work, speakers describe the dead as people of work (‘a beloved woman, honest, hardworking’(891), ‘since he was little, he liked working’ (333)); who lifted themselves out of humble origins (‘he knew the struggle of many Mexicans of getting ahead’ (191)); and as decent men dedicated to their families (‘a good man’ (1131), ‘an honest man’ (333), an ‘exemplary family man’ (573)).

Public virtues: Against the idea of the criminal as a ruthless maximiser of personal utility who is willing to destroy everything in his path that threatens his private material gains, speakers praise victims as publicly spirited citizens of firm principles (‘a committed man’ (10), ‘a participatory person’ (310), ‘a leading woman with an impeccable political career, of struggle, work, and strong beliefs in favour of the people’ (396), ‘a man loyal to the party and its principles’ (298), someone who ‘strongly believed in democracy’ (191)); solid personal values (‘he was very honest, very transparent’ (191)); firm local roots (‘a man who worked with his heart and his principles, always concerned about his beloved village’ (201)); a clean public record (‘an exemplary public servant’ (1008), ‘a political leader with an impeccable track record of struggle, work, and conviction’ (396)); and a spirit of personal sacrifice (‘an open and kind man with a permanent willingness to serve the people in his community’ (343), ‘with an enormous love for his people’ (121)).

Private esteem: Mourners would indicate the private esteem that victims enjoyed, above all, by describing them as friends, family members and carriers of other social roles: ‘a friend, godfather, good neighbour and family man’ (311), ‘a great leader, excellent person, but above all a great friend’ (664), ‘the best of friends’ (311).

Public esteem: Frequently, media reports would also indicate the public esteem in which victims were held. They might simply claim they were popular (‘very much loved by the people’ (1218)) or mourned by their constituents (‘Michoacán is in mourning’ (10)). More often, they would indicate the breadth of public support by describing the variety or sheer numbers of people attending their funeral: ‘About 350 people … accompanied the funeral procession’ (493), ‘A crowd accompanied Paco’ (343), ‘business people, politicians, family members, former mayors, council members, merchants, social leaders, workers, secretaries, public servants, department heads, local employees, all in mourning’ (311), ‘the building turned out to be insufficient to accommodate the 3,500 to 4,000 people’ (652).

Almost all these statements and observations on the virtue and esteem of killed candidates appear in funeral reports, which are a rare original contribution the print media sometimes make to stories of candidate killings. Attending funerals, in fact, is their only independent act of investigation. And funeral reports are the only places where victims emerge as individuals, embedded in their personal environments, so that readers may grasp the ‘social loss’ of their death.⁸³

The Depoliticisation of Electoral Violence

A candidate is shot dead. Over a ten-month electoral process, 48 candidates are shot dead. What did this do to relations among competing parties? In principle, the frame of organised crime deflects attention from them. By attributing authorship to criminal society, it shields political society from suspicion. And by enveloping victims in a cloud of uncertainty, it places the burden of proving their innocence on them, rather than their political adversaries. Does the systematic revision of the news coverage of candidate killings confirm these expectations of depoliticisation? Did the frame of organised crime indeed protect the electoral arena from reciprocal suspicions among contending parties and candidates? Tracing all hints of political suspicions in the universe of newspaper reports, I did not, in fact, find many. By its dominant reading, lethal electoral violence reflected a cleavage between the criminal and the political world, rather than conflicts within the political world. Departures from this interpretation were scarce and, with one single exception, not even antecedents of political conflict were able to shake it.⁸⁴

Non-Partisan Violence

Given the political role of the victims, it comes natural to political observers to speak of candidate killings as acts of ‘political violence’ (334, 669) or speculate about their ‘political motives’ (650, 232). Yet, even when admitting that these killings were political, speakers would generally conceive them as non-partisan.

Most references to ‘political motives’ and ‘political violence’ remain vague. They indicate the political role of the victims but remain silent on the political role of violence. What is the purpose of violence in elections? Some state the obvious: killing a candidate alters the menu of electoral choice, ‘the criminals want to decide who can participate in the elections’ (669). Others are more specific: criminals want to be decisive. When leading candidates get killed, their prospects of victory appear as a plausible motivation of their assassination: ‘the candidate who passed away was leading in the polls and was therefore murdered’ (774), ‘he was poised to win the city hall’ (788). Still others would describe electoral violence as a strategy of ‘intimidation’ (1084), designed to ‘generate fear or uncertainty in this electoral process’ (1084).

Speakers would not spell out the underlying logic, however. Who wants to select or intimidate candidates and why? Almost no one would point to political adversaries. Even when they see candidate assassinations for what they are – objective instances of ‘political’ or ‘electoral’ violence – political actors did not describe them as instances of partisan violence. Electoral violence hit candidates of all parties to a roughly similar extent. Political actors read this fact, not as a possible sign of spreading mutual violence among political adversaries, but as proof of its quasi-impartial nature: ‘this violence hits all parties’ (389) and ‘candidates from all parties’ (522, 1222), ‘it makes no distinctions in terms of political parties or ideologies’ (19). The notion that candidate assassinations constituted a form of criminal violence that was directed against political parties and the political class in general formed part of its common description. Indicating its non-partisan nature, actors would routinely talk about ‘the wave of violence against candidates’ (626), ‘the wave of violence against political leaders’ (283), ‘against those who aspire to elective positions’ (1239).

Frame Fissures

The common assumption that electoral violence had been untouched by the logic of partisan competition was not without fissures, though. Some observers admonished that the bitterness of inter-party competition had spilled over into physical violence. Violent rhetoric, they warned, ‘generates hate, resentment and confrontations and is no more than a prelude to other much more violent acts … against the integrity and life of candidates’ (522). Condemning ‘any form of violent attack, be it verbal or physical’ (1085), they issued general exhortations of moderation: ‘ideological differences must not constitute motives for aggressions’ (772), ‘the electoral process [must] unfold peacefully and within a framework of respect and tolerance’ (1083), without ‘undue passions’ (‘apasionamientos’) (890).

On counted occasions, the left-wing PRD would accuse state governments of crimes of omission or commission against their adversaries. After suffering a whole string of candidate killings in crime-ridden Guerrero, a state with a long tradition of political violence against the Left, the PRD started seeing an ominous pattern of partisan selectivity: ‘a series of despicable assassinations of PRD candidates’ (195) and ‘PRD members’ (669), ‘the escalation of violence [is] directed against left-wing leadership’ (533). In Jalisco, some party leaders held their political adversaries to be unable to protect them: ‘The PRI is incapable of guaranteeing the safety of its opponents in the states it governs’ (191). They demanded protection by the federal government against ‘the executions of party members’ (195). Only in one registered statement would a PRD leader go further and accuse the local government of active involvement in electoral violence, denouncing a campaign of ‘extermination of PRD members as it happened in the 1980s’ (115).

Histories of Conflict

To discern potential criminal motives, homicide investigators routinely inquire into victims’ history of conflicts. In the case of candidate assassinations, antecedent political conflicts might reveal political motives that put into question the attribution of authorship to criminal organisations. In a few cases, media reports did indeed mention pre-existing conflicts with four categories of actors outside the criminal underworld: societal actors, state agents, intra-party competitors, and partisan adversaries. Except in one case, however, these conflict histories were unable to derail the narrative of organised crime.

Societal conflicts: When environmental activist Salvador Magaña (C12) was murdered, media reports did not raise concrete accusations against anybody, yet they did point at the interests he had threatened in his year-long fight against ‘the privatisation of beaches’ (100) and ‘abuses of authority’ (94). Another candidate, Javier Fragoso (C33), was reportedly murdered after attending a community meeting on a valuable public property (664). In other occasions, by contrast, speakers would stress the absence of societal conflicts: there were ‘neither problems nor frictions’ (14) between the candidate and local communities. The candidate ‘had no problems with anybody’ (389).

Conflicts with state actors: In two cases, reports mentioned preceding conflicts with state officials as possible homicide motives. Gabriel Hernández (C17) had a record as a human-rights activist. His murder might have been committed ‘because of his work … as an activist’ (288). Fernando Ángeles (C47) had, after a lifetime without political engagement, campaigned against ‘corrupt’ municipal authorities, which ‘cost him his life’ (1217).

Intra-party conflicts: Many of the candidates who were murdered early in the electoral cycle were not yet official candidates but ‘pre-candidates’ campaigning to be nominated by their political party. In such contexts of intra-party competition, internal adversaries seem ‘natural’ addressees of suspicion. Yet, in only a small handful of cases did reports allude to internal party conflicts: One candidate had switched internal party factions in the past (386), another had switched parties (769). In a third case, one report mentioned that ‘the assassination took place in the midst of the internal process [his party] was carrying out to elect its candidates’ (486). In two further cases, Dulce Rebaja (C24) and Homero Bravo (C26), media reports explicitly stated that candidate selection processes within their parties had already concluded at the time of their murder (454, 473 and 534).

Inter-party conflicts: In the news, all these various conflicts were alluded to as possible motives of homicide, yet nothing else. No one would investigate, broaden accusations, specify them, follow the lead. These blurry allusions were left dangling in the air and left dissipating without any follow-up. The same happened with public suspicions against political adversaries. Only ‘days before his death’ (94), environmental activist Magaña (C12) had been denouncing ‘the diversion of public funds and an electoral offence’ (96) by local authorities. Again, an innuendo without sequel. Aguirre (C36), one of the few candidates with documented criminal involvement, had initiated libel proceedings against political adversaries who had accused him of criminal involvement (780). They proved right and had no part in his subsequent assassination.

In the end, in only one case did the presumption of non-partisan criminal authorship fracture. When Aarón Varela (C25), opposition candidate for mayor of Santa Clara Ocoyucan in the federal state of Puebla, was murdered, his supporters immediately attributed responsibility to the political organisation that had been controlling local power for decades: Antorcha Campesina (Peasant Torch). A corporatist survivor from the authoritarian past, the group had built a profitable empire of local political, economic and cultural power. Renowned for its brand of exploitative, contentious clientelism, it had a proven record of authoritarian conduct, including intimidation and violence.⁸⁵ Given the incumbent's solid anti-democratic credentials, the accusation against it had strong mobilisational resonance, driving people into the streets in protest: ‘Aarón did not die, Antorcha killed him’ (498).

The exception proved the rule, however. In all other cases, the narrative frame of organised crime prevailed. By locating electoral violence within the larger context of criminal violence, it provided a ready-made, non-political interpretation that allowed actors to comprehend lethal violence, not as a breach of basic norms among political actors, but as an external aggression on political actors.

Conclusion

During more than a decade, political violence had been knocking at the door of Mexico's fledgling democracy, claiming the lives of dozens of mayors, journalists and activists. In the 2018 general elections, however, it rose to new heights when an unprecedented wave of lethal force against local candidates burst through the floodgates of the electoral arena. In objective terms, these crimes were shrouded in mystery. No one claimed responsibility and police investigations yielded little information about their masterminds, perpetrators or motives. The established frame of organised crime, however, allowed Mexican society to ignore its factual ignorance and ‘make sense’ of these crimes despite their opacity.

Political violence creates hard, material facts. These facts, however, do not speak for themselves. Observers need shared frames of interpretation to form common understandings of ‘what is going on here’. In Mexico's 2018 general election, the narrative of organised violence allowed political actors to exclude a broad range of interpretative possibilities and ‘normalise’ electoral violence as ‘criminal business as usual’. Instead of treating candidate killings, for instance, as the result of structural forces, like anomie or poverty, or as acts of mentally deranged or ideologically radicalised individuals, they were able to comprehend them as the work of armed business groups in the pursuit of illicit material gain. In consequence, the 48 dead candidates constituted no more than a sad footnote to the 2018 general elections, the ‘dark side’ of an otherwise vibrant democratic election.⁸⁶ By incorporating their assassinations into the narrative of organised crime, political actors were able to render them normal and intelligible without requiring much factual knowledge about concrete cases. The established narrative offered a solid bridge over the chasm of systemic opacity.