Firms and labour markets in times of violence: Evidence from the Mexican Drug War

Hâle Utar 28 March 2021



For the last 13 years, Mexico has been home to a violent conflict known as the Mexican Drug War (Zedillo 2016). Warring factions of criminal organisations fight for control of territories and lucrative drug trafficking routes through Mexico to the US. These groups often use ostentatious killings and heavy weapons, and increasingly target the general public through kidnappings for ransom, collateral killings, and apparently random acts of violence. 

The violence is on such a scale that the Mexican Drug War ranks ahead of the wars in Iraq and Afghanistan as one of the most violent conflicts of the 21st century. To date, it has claimed more than 150,000 lives and over 50,000 missing persons. 

While the gruesome substance of the conflict has been amply covered in the media, along with its context of organised crime, attention to the firm-level economic consequences of the conflict has been surprisingly sanguine.  

Coverage has mainly centred around the understanding that international trade of goods with Mexico has not been substantially affected by the violence and that large international firms continue operations in Mexico undeterred (The Economist 2010). But large, internationally trading companies are not the whole economy, and in a recent paper I take a deep look at the effect that the drug war has on the Mexican manufacturing economy overall (Utar 2020). 

The two policy triggers of the Drug War provide the necessary exogenous variation across Mexican cities and over time to establish a causal link between violence and firm-level outcomes: 

  1. The drug war was precipitated by a sharp policy turn of the newly elected Calderon government in December 2006. The new policy, known as the ‘Kingpin Strategy’, aimed to curb the drug cartels by removing their leaders via military power. While the military deployment was successful in capturing or killing leaders of drug cartels, it had the unintended consequence of leaving power vacuums in organised crime, in which sub-factions of the cartels fight for turf and control (Dell 2015). 
  2. The stakes of the fighting were exacerbated by more forceful drug enforcement in Columbia at the same time, decreasing the supply of cocaine. Consequently, cocaine prices surged and violence intensified (Castillo et al. 2020).

The resulting violence did not affect cities equally (Figure 1). My study shows that within these cities violence does not affect firms and their workers equally either.

Figure 1 The rise in violence due to the Drug War is very different across cities.


Note: The number of homicide occurrences and population information are from the National Institute of Statistics and Geography (INEGI). Populations in the figure legend are the year 2010 numbers. Homicide rates are calculated using annual population figures and are annualized monthly rates of homicides.

The consequences of the conflict on the manufacturing sector are, in fact, dire. Manufacturing firms, on average, experience a significant reduction in capacity utilisation, employment, output, and productivity. The results also reveal strongly heterogonous outcomes both within firms with different types of employees and across firms with different characteristics. 

To get a clearer picture, it is necessary to map out the mechanisms of how the violence acts on the economy. The results tell an interesting and compelling story.

Firm-level employment response to violence

Figure 2 Firm-level employment and wage response to violence

Note: The figure shows the estimates of the firm-level elasticities of blue- and white-collar employment (left) and wages (right) with respect to the metropolitan area level  drug violence. Solid bar frames indicate statistical significance of the magnitudes. See Utar (2020) for details.

Manufacturing employment responds negatively to rising violence in a metropolitan area. But when focusing on different types of jobs within firms, results show stable and almost increasing white-collar employment and, in sharp contrast, a substantial decline in blue-collar jobs. The left panel of Figure 2 shows how firm-level employment responds to increased violence in the metropolitan area where a firm is located, depending on the type of employment. 

Violence may induce a lower labour demand, either because of lower output demand or lower worker productivity. But the right panel of Figure 2 shows lower labour demand cannot be the only driver of employment decline. It also shows that wages move in the opposite direction: violence increases blue-collar wages and decreases white collar-wages. This pattern of employment and wage response signifies a violence-induced labour supply change.

It appears from these within-firm changes that the violence of the drug war deters blue-collar workers from working more than it deters white-collar workers. This suggests Drug War violence creates a negative labour supply shock by disproportionately taking the lowest wage workers (who are likely to live in poorer neighborhoods, most afflicted by violence) out of the labour force.

Figure 3 How employment responds to violence depends on the type of workforce 

Note: Firm-level employment responses to violence depending on wage level (top) and female workforce intensity (bottom). Solid bar frames indicate statistical significance of the magnitudes. See Utar (2020) for details.

Although blue-collar production workers are among the occupations most exposed to violence, the blue-collar labour supply does not decrease due to a direct decimation of the labour force by death or emigration of workers to safer locations. Rather, the strongly heterogeneous impact of the violence across firms (Figure 3) indicates that the drug war affects blue-collar workers’ labour market decisions, especially those of lower-paid female workers. The figure below shows that the employment reduction in response to the local violence is much stronger in firms with lower wages and a female-intensive workforce. 

These results show that firms are affected obliquely through the labour market in an interesting mechanism opposite to most other economic shocks (that hit the firms first and, by consequence, the labour market). In the case of the violence shock, results suggest that it is the other way around: the violence deters workers from working and increases the reservation wage, below which the risk of working outweighs the benefit. 

Plants’ output response confirms the labour market channel but also reveals an additional channel: Local demand

When focusing on plant-level output, results show that output reductions resulting from local violence are also strongest in plants with a low-wage female workforce, confirming the heterogeneous labour market response channel. But the starkest contrast in output response is between plants operating in the local market versus exporting, importing, and geographically diversified plants. 

I find that firms’ export revenues are not disproportionately affected by the increased violence due to the Drug War, and neither is their probability of exporting.

Figure 4 shows the plant-level output response to the metropolitan area level of violence, depending on whether plants were exporting, importing, or geographically diversified before the shock.

The figure shows that the negative output effect of the war is mostly borne on non-diversified, smaller, local firms. This makes sense; violence is a local shock, so it affects the most local firms more.

And in a reverse confirmation of the media reporting, it is the small- and mid-size, domestically trading manufacturers that are hit hard by the violence. As it is the seed and growth layer of the economy that is suppressed by the biased attrition of the violence. This effect will likely mark the Mexican economy for a long time, even once the drug war is put to an end.

Figure 4 Heterogeneous output response to violence

Note: In this figure, each bar height shows the estimate of the output elasticity with respect to drug war violence among plants with specific characteristics. Solid bar frames indicate statistical significance. See Utar (2020) for details.


In all likelihood, the understanding of the economic effects of violence derived from studying the Mexican Drug War is more broadly applicable. Many cities in developing countries suffer from violence, and drug trafficking often plays a central role. The results in Utar (2020) illustrate how labour market developments affect firms, and point to important distributional and inequality consequences of violence. As the disproportionate impact of the Mexican Drug War was borne on plants that tend to be less productive, the aggregate output implications may be limited. But, at the same time, firms start small and local, and the most productive ones grow bigger and become international. By affecting those plants that have the potential to become big and diversified, organised crime-related violence is an important obstacle in the development of domestic industrial capability.


Dell, M (2015), “Trafficking Networks and the Mexican Drug War”, American Economic Review 105(6): 1738-1779.

Castillo, J C, D Mejia and P Restrepo (2020), “Scarcity without Leviathan: The Violent Effects of Cocaine Supply Shortages in the Mexican Drug War”, The Review of Economics and Statistics 102(2): 269-286.

Utar, H (2020), “Firms and Labor in Times of Violence: Evidence from the Mexican Drug War”, CESIfo Working paper 7345.

The Economist (2010), “Signs of life; Trade with Mexico”, 26 June. 

Zedillo, E (2016), ”Re-thinking the ’war on drugs’: Insights from the US and Mexico”,, 22 April. 



Topics:  Labour markets Politics and economics

Tags:  US, Mexico, Latin America, Narcotics, Conflict, Labour Markets, development, crime, Mexican Drug War

Associate Professor, Department of Economics, Grinnell College


CEPR Policy Research