I was aware of the high levels of violence in Colombian society since I was a child. I lived through various historical turning points in Colombia’s armed conflict. First, there was the assassination of Pablo Escobar and the downfall of the Cali cartel. Then I witnessed a surge in paramilitary violence at the height of its political and military power. Simultaneously, the FARC entered into peace talks with Colombia’s then-President Andrés Pastrana’s administration. Eventually, such negotiations collapsed, and the fight against insurgent groups intensified, primarily in response to Alvaro Uribe’s Seguridad Democrática approach. A new peace agreement with the FARC insurgency, this time fruitfully concluded, impacted the lives of many people, particularly those living in areas traditionally afflicted by armed conflict. Even in that case, there is much to be done in order to properly implement the final agreement signed on December 2016. Along with this very brief and superficial description of Colombia’s civil war, all of these events have had an impact on violence trends. Why is violence so common in Colombian society? The answer to this question is not so simple to obtain. That is why I decided to focus my PhD research on understanding the dynamics of violence in Colombia as they relate to the evolution of its armed conflict. The chapters presented here, in particular, elaborate on the unintended consequences of policy interventions aimed at reducing violence. These interventions, on the contrary, had the opposite effect. These chapters investigate the impact of the war on drugs and the peace agreement with the FARC insurgency on violence dynamics at the local level in Colombia. In Chapter 1, we argue that peace agreements may inadvertently increase selective violence against civilians when they are incomplete in two key dimensions. First, only a fraction of the existing armed groups participates in the agreement. Second, the legitimate government fails to establish an institutional presence in the areas previously controlled by those who do participate. Under these two conditions, the resulting vacuum of power may attract active armed groups who engage in selective civilian victimization to obtain control. Studying the recent Colombian experience, we find that the permanent ceasefire declared by the FARC insurgency in 2014 led to a surge in the targeting of community leaders in former FARC strongholds, perpetrated by armed groups excluded from the peace process, with the goal of consolidating their dominance in those areas. Critically, selective victimization is attenuated by some dimensions of state capacity and exacerbated in places that are more valuable as proxied by the existence of recent land conflicts. In Chapter 2, I analyze whether the use of violence in a civil war is bounded by the need of armed actors to secure territorial control. Specifically, the paper examines the causal effects of aerial spraying operations launched to eradicate coca production on patterns of civilian victimization committed by the FARC insurgency and paramilitary groups, two of Colombia’s most well-known non-state armed organizations. The results show that aerial eradication operations trigger violence against civilians perpetrated by these two groups. In general, the findings suggest that territorial control disputes appear to be influenced by the presence of government security forces rather than the potential negative economic impact of aerial eradication operations. Repertoires and targets of violence are dictated by two main factors: the strategic importance of the territory and the level of cooperation between combatants and civilians. In Chapter 3, we investigate the consequences of expressing preferences about peace and conflict in Colombia, where the majority of people voted against the final peace agreement reached between the FARC insurgent and the government in a referendum in October 2016. We employ a regression discontinuity design with the referendum vote share as the score variable, and we identify a differential increase in violent events in municipalities that did not approve the peace agreement. This effect is concentrated primarily in former FARC strongholds and municipalities where coca is cultivated, and mining is practiced. These findings support the view that the peace agreement was regarded as an opportunity for non-state armed organizations that had not yet been disarmed to potentially contest economic rents previously controlled by the FARC insurgency. Such organizations use the outcome of the peace referendum as a signal to determine what FARC strongholds control
