This is our sixth blog post for our Job Market Paper Series blog for 2024-2025.
Felipe Parra-Escobar is a Ph.D. candidate in the Department of Agricultural and Applied Economics at the University of Wisconsin-Madison. His research interests lie in the field of applied microeconomics, with a focus on development economics, labor economics, and industrial organization. You can find his JMP here.
Antipersonnel landmines are widely used in conflicts worldwide, currently deployed by armed actors in at least 13 countries and contaminating areas in a total of 60 countries. Although intended to target combatants, landmines disproportionately harm civilians—85% of registered casualties in 2022 were non-combatants (International Campaign to Ban Landmines, 2023). However, the impact of these devices extends beyond physical harm. Individuals living in mine-contaminated areas face a persistent fear of encountering these devices in daily activities, which disrupts economic activity and distorts household decision-making.
In my job market paper, coauthored with Sakina Shibuya, we study how landmine-related events impact household livelihoods, focusing on labor market decisions and healthcare-seeking behavior. Our analysis takes place in Colombia, where non-state armed actors have, over the past three decades, installed landmines across agricultural land, along walking paths, and near roads. These devices predominantly affect rural households, which may decide to reduce agricultural labor and limit travel to lower the risk of encountering landmines. Specifically, we explore how households adjust labor activities in their own fields and on other households’ farms. We then focus on healthcare-seeking behavior, a potentially risky activity given that health centers are often located in municipal capitals in rural Colombia, requiring individuals to travel through areas where they may encounter landmines.
Identification of landmine events close to households’ homes
We identify whether landmine events occurred near households’ residences both shortly before and several months prior to being surveyed. We do so by combining data on the exact locations and dates of landmine events in Colombia since 1990 with detailed spatial data on households from the Colombian Longitudinal Survey, covering four rural regions from 2010 to 2016. During this period, preceding the 2016 peace agreement, non-state armed actors installed landmines continuously, allowing us to examine household responses to recent events. Unlike prior studies, which primarily focused on long-term impacts and relied on district-level measures of landmine exposure, our approach enables us to capture variation within districts, uncovering individual effects often overlooked by aggregated measures.
The primary challenge that we face when estimating the effect of landmine events on households’ behavior is the correlation between their occurrence and characteristics of the places where households live. Non-state armed actors typically install landmines to protect their strongholds or attack official forces, leading to a higher likelihood of landmine events in contested territories or areas experiencing intense conflict. To address this issue, we exploit a unique feature of landmines: their exact locations are known only to those who install them. This makes landmine events sudden and unpredictable for both civilians and military personnel, given the area’s landmine contamination level and conflict intensity. We argue that, conditional on these factors, landmine events can be treated as plausibly random. Since landmine contamination is not directly observed, we use a fixed-effects model to control for baseline landmine contamination levels around households, as well as conflict dynamics that vary across municipalities based on their distinct characteristics.
Households avoid risky activities after landmine events
We find that households tend to avoid risky activities after recent, nearby landmine events. Individuals are 16% less likely to work in long-term agricultural jobs (non-jornalero) on other households’ farms when a landmine event has occurred within 5 km of their residence in the six months preceding the planting season (figure 1, first panel). This effect diminishes over time, with individuals more likely to take on this type of work if exposed to landmine events in the past one to three years. Likewise, households exposed within the six-month timeframe are 12% less likely to carry out agricultural tasks on their own fields and more likely to hire workers from other households for these tasks (figure 1, fourth panel). Specifically, we find that households are 21% more likely to hire agricultural labor if exposed to a landmine event shortly before the planting season.
This behavior extends to other activities that require individuals to leave their homes, such as seeking healthcare. We find that adults are 12% less likely to visit a medical professional for preventative reasons if a landmine event occurred six months before the planting season. Instead, they seem to turn to alternative medicine, with exposed adults being 60% more likely to visit alternative medicine providers after a landmine event. Households may reduce visits to formal medical providers to avoid the risks of travelling to health centers, which are often located in municipal capitals far from their homes. By doing so, they not only lower their risk of encountering landmines but may also reduce medical and transportation costs during times of reduced labor income.
Liquidity-constrained households engage in risky labor out of necessity
Our results so far show that households avoid activities that may increase their risk of encountering landmines after an event. However, we find that individuals are more likely to engage in agricultural day labor (jornalero) in other households’ farms following a landmine event (figure 1, third panel). Though this behavior may seem counterintuitive, as we expect individuals to avoid any activities that increase landmine exposure, households facing liquidity constraints may turn to riskier work to offset income losses. In contrast, households without liquidity constraints can afford income reductions without resorting to other high-risk activities. To examine this hypothesis, we estimate heterogeneous effects by land ownership, a factor strongly correlated with access to credit markets in rural Colombia.
We find that non-landowners—who are more likely to face liquidity constraints—engage in risky labor following recent landmine events. These individuals are more likely to work as agricultural day laborers if a landmine event occurred within six months prior to the planting season, while reducing the hours they allocate to long-term agricultural jobs and non-agricultural work. As a result, their overall labor income does not change following recent exposure to landmines. In contrast, landowners—who are less likely to face liquidity constraints—simply work less in long-term agricultural jobs, leading to a decline in their labor income. We observe a similar behavior with agricultural work on households’ plots: while landowners are less likely to work in agricultural tasks on their own fields, non-landowners continue working on the plots they have access to.
What are the policy implications?
We draw two main implications from our study. First, our findings reveal how landmines distort economic activity and household behavior, potentially leading to negative welfare impacts for individuals living in conflict zones. Demining campaigns could therefore provide significant benefits to communities in these areas, as shown in other studies (Chiovelli et al., 2024; Prem et al., 2024), even if households return to conduct the activities they were doing before being exposed. Second, our results highlight the unequal risk-bearing capacity of households based on their wealth. While wealthier households can afford to avoid risky activities after landmine events, liquidity-constrained households are more likely to engage in high-risk activities to sustain their livelihoods. Recognizing these differences in coping mechanisms is essential for designing policies that support individuals in conflict-affected areas.
Feature image captured by the Felipe Parra-Escobar during exploratory fieldwork in Colombia in the summer of 2023.
