The levels of violence in Mexico have dramatically increased in the last few years due to structural changes in the drug trafficking business. The increase in the number of drug trafficking organizations (DTOs) fighting over the control of territory and trafficking routes has resulted in a substantial increase in the rates of homicides and other crimes. This study evaluates the economic costs of drug-related violence. We propose electricity consumption as an indicator of the level of municipal economic activity and use two different empirical strategies to test this. To estimate the marginal effect of violence in the rate of homicides (per 100,000 inhabitants) we use an instrumental variable regression created by Mejía, Castillo and Restrepo (2012). For the average municipality, the marginal negative effect of the increase in homicides rates is substantive for earned income and the proportion of business owners, but not for energy consumption. Although negative and statistically significant, the effects are mild for labor participation. We also employ the methodology of synthetic controls to evaluate the effect that inter-narco wars have on local economies. The analysis indicates that the drug wars in those municipalities that saw dramatic increases in violence between 2006 and 2010 significantly reduced their energy consumption in the years after the change occurred, which is interpreted as a significant reduction in GDP per capita for these municipalities.
Gabriela Calderon holds a Ph.D. in Economics from Stanford University. Her research interests include policies that affect gender differences in developing countries, policy evaluation, violence in Latin America and the effect of institutions and governance on the provision of public goods and health/education outcomes. She did her master's degree in economic theory and bachelor's degree in economics at the Instituto Tecnológico Autónomo de México. Currently, in the Program on Poverty and Governance, her research analyzes the way institutions and democracy affect the provision of public goods, and the impact they have on health outcomes like infant mortality trends. She is also studying the effects of government interventions that combat drug-trafficking organizations over violence in Mexico.
Her research has focused on the topics of development, public finance, and the evaluation of public policy programs in Mexico. For example, during the summers of 2009/2010, she conducted a field experiment in Zacatecas, Mexico with Giacomo de Giorgi, an assistant professor from Stanford University, and Jesse Cuhna, a former Stanford student. The main task was to evaluate the impact of financial literacy classes on underprivileged women entrepreneurs in the region. To successfully complete an evaluation in an untreated region, they proposed collaborating with the Mexican NGO CREA on a joint project. They contacted local interviewers, trained them, and identified all women entrepreneurs in the 17 communities, in which we conducted the experiment. Preliminary results suggest that the female entrepreneurs who were randomly assigned to treatment earned higher profits, had larger revenues, and served a greater number of clients. They also found that they were more likely to implement formal accounting techniques.
She has also studied programs that are not randomly assigned as an experiment. For example, she has analyzed the effects of a national policy in Mexico of child care services, called Estancias Infantiles para apoyar a Madres Trabajadoras (EI), using administrative, census and household data. Her empirical research strategy identifies the effects of the program on both the men and women who were eligible for the program. She used time, location and eligibility variation, and considered a major threat to identification of the actual effects: for example, a manufacturer who moves into a municipality at approximately the same time as the EI program and who happens to disproportionately demand the skills of women who were eligible to the program happened to have. To ensure that such scenarios do not affect her results, she chose not triple difference strategy, in which all ineligible people are treated as “controls” for the EI-eligible families. Instead, she employs Synthetic Control Methods, using the same methodology as Abadie and Gardeazabal (2003) and Abadie, Diamond and Hainmueller (2010) to ensure that her control group has the same mix of skills and preferences as the EI-eligible group. She adapted the Synthetic Control Method to analyze repeated cross-sectional household data, which are data that are typically available in developing countries