Time series applications to intelligence analysis: a case study of homicides in Mexico

摘要

The scale of lethal violence in Mexico seen in the past decade has been a pressing concern for both Mexican and US officials, including law enforcement organizations, intelligence agencies, and policy makers. With much of the homicides being a result of the trafficking of illegal drugs, it has been suggested that the homicides in Mexico follow seasonal patterns tied to the drug trade, specifically to the cultivation of heroin. In this paper, conventional econometric time series methods are applied to test this hypothesis. Results demonstrate that not only do the drug-related homicides in Mexico display evidence of seasonality, but also that seasonality appears empirically related to the heroin trade.The paper makes the larger argument that time series and other statistical methods are an untapped resource that can complement standard intelligence analysis to support defensible judgments based on the scientific method of inquiry. However, a fuller integration of statistics and traditional analysis would require sufficient support structures be developed to encourage and promote such analysis.