Association for Public Policy Analysis & Management

Mexico has long had problems with drug cartels, owing largely to the country’s proximity and access to the United States, one of the biggest drug markets in the world. Since 2006, however, the scope of these cartels has grown exponentially. Violence appears to no longer be contained to drug-related activities, as the cartels have expanded into other activities, like extortion, kidnapping, and car theft.

Most of the literature on violence in Mexico has focused on its effect on output, growth, or migration. These are unlikely to be representative of the full effects of the risks associated with the increases in violence. We are interested in understanding how violence has affected the way in which ordinary individuals live their lives. For example, in places where violence is pervasive, citizens may change their habits and their spending patterns. Risk averse residents may no longer frequent restaurants and bars, fearing kidnapping or crossfire. Additionally, citizens may stop wearing jewelry and refrain from purchasing fancy cars, fearful of attracting the wrong kind of attention.

In this paper, we use detailed household surveys and a nationwide database on homicides to examine whether there is a significant relationship between levels of violence and two measures of household response - patterns of conspicuous consumption and behavioral changes in from 2002 to 2012 in Mexico.6 We measure changes in visible consumption using standard measures of conspicuous spending including the among spent on commodities like vehicles, entertainment, gambling, and clothing. We measure changes of behavior using a set of survey questions over whether people have changed how often they eat out, carry valuables, change modes of transportation (or routes) for security reasons.

We find that as violence rises within a municipality, households spend significantly less on visible forms of consumption. We go on to show that the impact of murder on visible consumption is heterogeneous across the country. When we divide the sample by the level of violence, we find no effect of changes in the homicide rate on visible purchases precisely in the areas that are the most dangerous. On the other hand, we document very large and significant negative relationship between increased violence and visible expenditure on households located in municipalities that experienced the largest changes in violence. We find that violence is most strongly associated with changes in conspicuous consumption for middle- and upper-income groups.

We also find that households significantly alter their behavior in response to increased murder rates in predictable ways to lessen their risk. Specifically, we show that households change their routes of transportation when homicides are increasing. This effect is even stronger for female headed households and for low-income households. The latter also changed their method of transportation in response to increased violence. Lastly, we find that female household heads and low-income household heads also appear less likely to carry valuables. In sum, homicides lower visible consumption more in male-headed households and in middle- and upper-income households, while murders change behavior more for female-headed households and the poorest households.