Since the early 2000s, Mexican cartels have increasingly threatened, attacked, and assassinated government officials. The literature contends that electoral incentives, political vulnerability, and government crackdowns account for these killings. We propose an additional explanation: the radical transformation of Mexican cartels that coincided with the increase in violence against the state. As large cartels fragmented, smaller and more localized cartels emerged, creating cartels with distinct capabilities. Fragmentation also multiplied inter-cartel conflicts, creating incentives for different types of cartels to target the state. Using both original and existing datasets on cartel presence and violence against government officials, we find that larger cartels and inter-cartel conflicts -- initially those involving major cartels but subsequently those also including smaller cartels -- are the strongest predictors of violence against the state. Small cartels alone do not appear to explain assassinations. These findings suggest that transformations in the criminal underworld can result in violent political consequences.

Criminal organizations kill tens of thousands each year. This article argues that criminal organizations under women leadership are less violent due to women's gendered pathways to power. We examine Mexico’s Tijuana Cartel and leverage the 2008 ascension of a woman to leadership following her brother's arrest. Synthetic control results show that the Tijuana Cartel reduced visible violence and territorial expansion under a woman's leadership, but slightly increased hidden forms of violence. Original data on all male cartel leaders in Mexico shows systematic gender differences in pre-leadership violent socialization and expertise that help explain the results. We probe generalizability by constructing 44 case studies of women criminal leaders across 18 countries since 1850 and find that gender profoundly shapes their trajectories, with our findings likely extending to women who inherit leadership from male kin. Our findings highlight the need to incorporate gender into research on organized crime.

Criminal organizations (COs) often fragment, triggering inter-CO wars that are now frequently deadlier than civil wars, but this process remains largely understudied. This article develops a theory in which opportunistic potential defectors, differing in relative power vis-à-vis their parent CO, strategically weigh two concerns---economic viability and surviving retaliation---when considering fragmenting. These calculations are conditioned by illicit markets, geography, and alliances. Leveraging an original dataset of Mexico's drug cartels and their fragments from 2000 to 2018, we first show that defections occur across ranks and that kingpin strategies only partially explain fragmentation. We then show that larger fragments can operate near their parent cartel, compete in similar markets, and form military alliances with other cartels to withstand retaliation, while smaller fragments survive by operating in distant areas and more local markets. Further evidence suggests that fragments are more violent and predatory.

Multi-method research combining quantitative evidence of causal effects with qualitative evidence of causal mechanisms has gained prominence in political science, yet its practice remains informal and inconsistent, presenting challenges for conducting and interpreting this type of research. This article introduces a unified framework that formally integrates static, within-unit causal mechanisms into the potential outcomes framework, bridging the divide between quantitative and qualitative approaches. By deriving the role of qualitative mechanisms, the framework provides clear guidance on multi-method research designs, their underlying assumptions, and their application and interpretation. To demonstrate its utility, I apply the framework to a recently published study in comparative politics, illustrating how it can help guide and strengthen the ontological consistency and transparency of multi-method research. This work advances methodological debates and provides practical tools for improving multi-method research in political science.