Mexican drug violence endangers development by increasing poverty and disrupting economic activity

Conflict, Crime and Violence
Poverty - Inequality - Aid Effectiveness

Since the beginning of the new millennium, Mexican drug cartels turned against each other, killing more than 6,680 people between 2001 and 2005 (Ríos and Shrik, 2011). Concerned about this violence, Felipe Calderón, the then recently elected President, declared war against cartels in December 2006. He actively prosecuted cartels with military force, intensifying and spreading the killings to areas that had not experienced this violence before. The so called, “war on drugs”, claimed the life of another 63,000 people just between 2006 and 2012 (Molzahn et al., 2013).

As the killings increased, civilians and local businesses also became subjected to extortions and other thefts. But how extensive have the externalities of the drug violence been? How about areas where cartels are still operating “peacefully” without killing each other, do these cartels foster the local economy?

In the working paper Gutiérrez-Romero and Oviedo (2014) we estimate the impact that drug cartels and separately drug related homicides have had on development in Mexico. Specifically, we assess the impact on poverty, inequality and human capital. We also explore some of the mechanisms through which drug trafficking and drug related homicides might be affecting these outcomes. We do so by assessing the changes in internal migration and economic activity. We focus on the industries where it is possible to identify from census records where their production is taking place at small-area-level (municipality).1 So, we analyse manufacturing, one of the biggest industries of the country accounting for 35% of Mexico’s Gross Domestic Product, as well as real estate and wholesale trade.

To identify the areas where cartels have been active, with and without related homicides, we surveyed official records; national and international media reports; and specialised blogs. We also use the official statistics on drug related homicides available since December 2006.

We evaluate the impact of cartels and drug related homicides using the difference-in-difference kernel matching estimator. Specifically, we estimate the change in outcomes before (2000-2005) and after cartels settled in areas for the first time (2006 or afterwards). We compare that change in outcomes to the ones experienced in areas that did not have cartels or drug related homicides over the same periods. We match these areas, “treatment” and “controls” based on their characteristics and probability of experiencing cartels and drug related homicides.

In figure (1) we illustrate the 70 municipalities where according to our estimates cartels started operating for the first time in 2006 or after without experiencing drug related homicides. In that figure we also illustrate the areas we use as controls to build a counterfactual of what would have happened to the areas where cartels moved in had these areas remained free of cartels.

We find that inequality, as measured by the Gini coefficient, declined nearly by two points in areas where cartels were active without incidents of drug related homicides.2 A fall in remunerations in manufacturing is likely to explain this decline in inequality, since we did not find changes in poverty rates nor a relative increase in the income of the poor in these areas.

To assess whether the impact of drug violence differed according to the intensity of drug related homicides, we divided the affected areas into four subgroups. The first subgroup consists of municipalities in the tenth decile according to their rate of drug related homicides per 100,000 inhabitants. This group has a considerably higher average drug related homicide rate (282.2) than the rest (22.4). We split the remaining 90% of areas affected by drug homicides into tertiles. Figure (2) shows the 688 municipalities that experienced drug related homicides for the first time in 2006 or after and for which we found similar areas free of drug homicides we could compare to.

Drug violence increases poverty

We find that drug related homicides increased the percentage of people living in food poverty by 3.1 percentage points in the areas with highest drug related homicides between 2006 and 2010. The percentage of people living in patrimony poverty, those who cannot afford food, health, education, clothing, housing and public transport needs, also increased by 2.6 percentage points in the areas with the lowest levels of drug related homicides.

Two factors might explain the increase in poverty. First, the manufacturing industry in areas with the highest rates of drug related homicides experienced a reduction in production, profits, number of establishments, workers and remunerations. Real estate’s production (sales) in these areas also declined. Second, in the least violent areas remunerations in manufacturing declined, and people migrated from the more violent places. Most of these immigrants were mainly of low earning income.

The areas affected by drug related homicides also experienced a small but statistically significant increase (0.3 percentage points) in the number of children aged 6 to 14 dropping out of school early, despite not experiencing a decline in the number of schools or teachers per school age population.

All these results refer exclusively to the areas that experienced cartels or drug related homicides for the first time in 2006 or afterwards. Focusing on this period offers the advantage of capturing the short-term impacts of cartels expanding into new areas. In our working paper, we also show that areas that experienced drug related homicides in an earlier period, during 2001-2005, also experienced an immediate increase in poverty and reduction in the number of workers employed in manufacturing. Both these impacts worsened even further in these areas during the 2006-2010 period when drug related homicides intensified.


The war among drug cartels endangers Mexico’s development by increasing the number of people living in poverty, children dropping out of school, displacing population and disrupting local economies. In our analysis we controlled for poverty-relief subsidies that people received from the government and remittances from abroad. The fact that despite these transfers poverty is still on the rise suggests that these areas need urgent complementary policies to ensure that these negative impacts do not persist over time.

We failed to find positive effects in areas where cartels are working “peacefully”, without killing one another. This suggests that on average the local population is not benefiting from such drug trafficking, or at least not enough to reduce poverty or unemployment rates in these areas.

1. We omit other industries, such as construction or finance, as the census records refer indistinguishably to either the municipality where the headquarter company is based at, not where necessarily these indicators are taking place.

2. The poverty and inequality statistics analysed were estimated at municipality level by CONEVAL. Other outcomes analysed are drawn from the economic and population census provided by INEGI.


Gutiérrez-Romero, R. and Oviedo (2014) The good, the bad and the ugly: The socio-economic impact of drug cartels and their violence in Mexico, 10/2014 Working Paper, Universidad Autonoma de Barcelona, Department of Applied Economics.

Molzahn, C., Rodriguez Ferreira, O., Shirk, D. 2013. Drug violence in Mexico: Data and analysis through 2012. Trans-Border Institute, Kroc School of Peace Studies at the University of San Diego, Special Report: February 2013.

Ríos, V. and Shrik, D. 2011. Drug violence in Mexico: Data and analysis through 2010. Trans-Border Institute. Kroc School of Peace Studies at the University of San Diego, Special Report: February 2011.


Share this