Research

Working Papers

International Migration Amid Insurgent Violence (Job Market Paper)

How does exposure to insurgent violence influence an individual’s decision to leave their home country? To answer this question, I combine anonymous mobile call detail records with data on the location and timing of violent events in Afghanistan to identify individual exposures to insurgent actions. I propose a novel method to detect international migration events that combines inferences from users' spatiotemporal movement patterns and episodes of sample attrition. Estimating a migration choice model reveals that a unit increase in the rate of monthly exposures to insurgent violence raises the odds of international migration by 6.4%. I show evidence that this effect is due more to extreme, high-casualty events than to the gradual accumulation of exposures to common, low-intensity events. In contrast, internal migration is unaffected across all violence types and intensities. I explore how wealth and employment composition might explain the differential effects. These findings provide policy-relevant insights into how potential migrants respond to insurgent violence.

The Impact of Mobile Internet on Immigration: Evidence from Mexico's 3G Rollout*

This paper examines the effects of Mexico's rollout of mobile internet infrastructure on immigration to the United States. It asks: does the arrival of mobile internet (3G) in a Mexican municipio fundamentally change individuals' migration choice? There are several decisions on which one might expect an effect--the binary migration choice itself, destination choice, optimal border crossing point, and aid uptake during travel. I measure the relative strength of two different mechanisms affecting migration behavior. On one hand, 3G access introduces an important information channel for shaping expectations of life in the destination country and avoiding risk en route. On the other hand, 3G is a valuable amenity whose addition to the home community may increase the utility of staying. The paper also examines a relevant side effect of the aforementioned information channel, namely the possibility of price convergence for the services of border-crossing guides.

Cell Phone Coverage and Violence: Evidence from a Radio Propagation Model in Afghanistan (with Robert Gonzalez and Lin Yang)

We provide evidence that mobile phone coverage lowers insurgent violence using two separate empirical designs in Afghanistan. First, we propose a novel regression discontinuity (RD) design for multiple, unknown cutoffs that uses signal strength as the forcing variable. We estimate signal strength for each cell of a 1X1 kilometer grid using a high-resolution radio wave propagation model. Second, we employ a differences-in-differences design leveraging the staggered rollout of cell towers for the largest operator in Afghanistan. We explore insurgent coordination and information sharing as possible mechanisms. We find weak evidence of spatial and temporal changes in the clustering of attacks following the installation of a tower in a district. However, we find evidence of coverage having larger effects near populated areas, primary roads, during morning hours, and a sharp increase in call traffic preceding certain attacks, all suggesting that information sharing is a key mechanism.