Most of my work analyses non-violent behaviour of criminal organizations and armed actors.
One focus of my research is the motivations behind the provision of goods and services by criminal organizations, a form of criminal behaviour that has become more frequent and visible across communities in Latin America but that has received limited scholarly attention.
A second research line examines novel methodologies for data collection and analysis, applying computational techniques to study difficult-to-observe phenomena such as organized crime.
My third and related line of work looks at the consequences of subnational violence in public life, with particular focus on the effects of violence on local political scenarios.
Below you can find the latest projects I am working on.
Gifts or Gambits? Explaining the Provision of Goods and Services by Criminal Organizations
Under Review
Under what conditions do criminal organizations provide goods and services to civilians? This study argues that the provision of goods and services is a strategic response to criminal competition. In contested environments, criminal organizations deploy visible and episodic distribution of material goods to secure civilian cooperation and consolidate territorial control. To evaluate this argument, the study introduces an original dataset of criminal provision events across Mexican municipalities during the COVID-19 pandemic (2020-2023).
Leveraging the pandemic as a common exogenous shock, the study examines variation in the provision of goods and services across Mexican municipalities. Results show that higher levels of criminal competition are associated with significant increases in both the likelihood and frequency of provision, while violence, measured by homicide rates, is not a robust predictor. These findings identify criminal competition as a central driver of non-violent criminal strategies.
Community Dependence and the Provision of Goods and Services: The Gulf Cartel and Los Zetas in Veracruz
This study asks why, under similar competitive conditions, some criminal organizations provide goods and services to civilians and others do not. I argue that competition is necessary but not sufficient, and that it is the interaction between competition and the community dependence of an organization’s illicit markets what ultimately drives the provision of goods and services by criminal organizations. I theorise that this interaction produces four patterns of provision: strategic, residual, incidental and absent. I illustrate the examining the cases of the Gulf Cartel and Los Zetas in the state of Veracruz, Mexico.
Extracting Data on Organized Crime with Large Language Models: The Organized Crime Behavioural Record
Measurement remains one of the main challenges for the analysis of organized crime. Criminal activity is hard to observe and difficult to systematically record at scale. To address this, I develop a four-stage protocol for extracting structured data on criminal behaviour from Spanish-language newspaper text using large language models. The protocol comprises 1) a theoretically motivated codebook, 2) the constrained extraction of the coded fields, 3) validation against human coding, and 4) post-extraction processing into an analysis-ready dataset.
I apply the protocol to 48,113 articles published by 77 Mexican outlets between April 2017 and September 2025, and present the resulting dataset, the Organized Crime Behavioural Record, Mexico (OCBR-Mex). Resulting counts are consistent with official homicide records and with previous event data.
This study provides a workflow that produces usable measures of criminal behaviour while identifying where those measures are reliable and where they are not, it also develops a process that can be adapted to other contexts, languages and phenomena.