Mexican president Felipe Calderon's time in office—from 2006 to 2012—was marked by violence as soldiers took to the streets and forests, making the war on drugs an especially tragic one. That war was an abject failure, and one with a serious human cost: Military intervention led indirectly to roughly an additional 200 homicides per year nationwide, according to an analysis published last month in The American Statistician.
Researchers Valeria Espinosa, now at Google, and Donald Rubin, a statistics professor at Harvard University, were not the first to wonder whether militarizing the fight against drug cartels brought with it some collateral damage. An analysis in Nexos magazine suggested that the drug war had pushed non-combat homicides higher, though neither that report nor an earlier visual comparison of homicide rates and military action, also published in Nexos, was very rigorous. Visual comparison isn't really a recognized statistical technique, and neither analysis took into account a variety of potentially important factors, such as when military interventions began in each of the 18 Mexican states involved.
Murder rates in Mexican states where the military had fought cartels went up by 11 people per 100,000, an increase about two and a half times the total murder rate in the United States in 2013.
Espinosa and Rubin addressed such concerns by comparing actual homicides in cities and towns where the army had fought drug cartels with predictions of homicide rates had there been no armed conflict. Since the researchers couldn't observe something that never happened, they used murder rates in demographically, economically, and politically similar towns around Mexico where there'd been no fighting as a proxy. That's better than using nearby towns' murder rates, they argue, since the effects of military intervention could have spilled over there.
Overall, Espinosa and Rubin estimated that murder rates in Mexican states where the military had fought cartels went up by 11 people per 100,000, an increase about two and a half times the total murder rate in the United States in 2013. Much of that increase was concentrated in the area around Ciudad Juárez, where military action indirectly pushed the murder rate up by about 86.5 deaths per 100,000 people per year. Even if the murder rate had been zero before military intervention, that increase would put it up there with heart disease and diabetes as a leading cause of death in Mexico.
While the drug war did curb the bloodshed in some places—Espinosa and Rubin estimated that military action cut the homicide rate in Apatzingán by 21 murders per 100,000 people—it wasn't enough to offset the killings in other, more populous regions. Even excluding Juárez, where the increase in drug-war related homicide rates was more than double anywhere else in Mexico, Espinosa and Rubin estimated Calderon's drug war increased the homicide rate by 6.5 murders per 100,000 people per year.
Though the results confirm others' expectations, they also illustrate the importance and challenge of determining what's a consequence of a policy choice—in this case, military action against drug cartels—and what's just a chance correlation, the authors suggest.
Quick Studies is an award-winning series that sheds light on new research and discoveries that change the way we look at the world.