Mexico has a massive murder problem. 2012 alone saw more than 26,000 homicides in the country, the fourth most of any state in the world. Why?
Drug violence and the interaction between cartels is a major factor. In a new working paper, Paul Zachary (UCSD) and I argue that uncertainty about local leaders has a great impact on that cartel violence. Cartels benefit from using violence to eliminate rivals. Politicians, however, have a vested interest in limiting that violence. This causes tension between cartels and officials, which cartels often attempt to resolve by bribing the officials to look the other way.
Why does uncertainty matter here? Paul and I investigate a model involving two rival cartels—a status quo and a challenger—and a local politician. The cartels want to capture as much of the drug rents as feasible; the local politician wants to minimize violence, but he is willing to look the other way if he receives a large enough bribe. The game begins with the status quo cartel offering a bribe to the politician to minimize enforcement. If the politician rejects, he chooses an amount of effort to exert to reduce the effectiveness of violence, which undermines the status quo cartel’s ability to maintain its drug rents. After, both cartels choose an amount of costly violence, which determines what percentage of the drug rents each receives.
We find that successful bribes lead to higher levels of violence. This is for two reasons. First, and most obviously, additional enforcement intercepts an additional percentage of violence. But there is an important second-order effect as well. The interception of violence functionally increases the marginal cost of violence for the status quo cartel. Consequently, more enforcement not only quashes violence in action but deters some of its production as well.
The above logic leads us to investigate what might lead to bargaining failure during the bribery stage. We show that the status quo cartel’s knowledge of the politician’s level of corruption is key here. When the cartel knows the politician’s minimally sufficient bribe (which is a function of the level of corruption), it can very easily come to terms. But when the cartel can only guess from a wide range of possibilities, it might ultimately offer a bribe that isn’t big enough for the politician to bite. In expectation, this leads to higher levels of enforcement and less subsequent violence.
Our theoretical argument has a noteworthy empirical prediction. Uncertainty leads to less violence, and some work in IR indicates that there is more uncertainty about newer leaders than older leaders. Thus, in Mexico, we would expect municipalities that the same political party has controlled for longer periods to be more violent than municipalities with greater turnover. (In the absence of our argument, this would be odd: the retrospective voting literature would suggest that less violence should correlate with greater tenure, as voters should be rewarding politicians who keep the streets safe.) The data support our theory. Indeed, we estimate that an increase of one year in tenure is associated with roughly one additional murder within a municipality. Although this might not seem like much, with so many municipalities in Mexico, a countrywide increase of one year in tenure matches up with about 2300 more murders. This number is on par with the 2011 murder totals in France, Germany, United Kingdom, Netherlands, and Belgium combined but still only a fraction of the overall number of murders in Mexico in every given year.
Again, you can download the paper here. This is the abstract:
What role do politicians have in bargaining with violent non-state actors to determine the level of violence in their districts? Although some studies address this question in the context of civil war, it is unclear whether their findings generalize to organizations that do not want to overthrow the state. Unlike political actors, criminal groups monopolize markets by using violence to eliminate rivals. In the context of the Mexican Drug War, we argue that increased time in office increases cartels’ knowledge about local political elites’ willingness to accept bribes. With bribes accepted and levels of police enforcement low, cartels endogenously ratchet up levels of violence because its marginal value is greater under these conditions. We formalize our claims with a model and then test its implications with a novel dataset on violent incidents and political tenure in Mexico. We find that each additional year after an official initially takes office is associated with an additional 2,300 violent deaths countrywide.
It’s a working paper, so we’d love your feedback!