Date of Graduation

Spring 5-23-2026

Document Type

Thesis

Degree Name

Master of Science in International and Development Economics (MSIDEC)

College/School

College of Arts and Sciences

Department/Program

International and Development Economics

First Advisor

Andrew Hobbs

Abstract

This paper examines whether crop yields affect violent conflict at the subnational level across seven Sub-Saharan African countries: Burkina Faso, Ethiopia, Mali, Malawi, Niger, Nigeria, and Tanzania, over the period 2010 to 2024. A central identification challenge is that observed yields are endogenous to conflict: armed violence disrupts planting, cultivation, and harvest. To address this, I train a machine learning model on observed maize yield and satellite and environmental features to generate predicted yields for 1,338 second-level administrative districts annually. These predictions serve as the key regressor in a two-way fixed-effects panel with district and year fixed effects. I find that a 10% increase in maize yield is associated with an approximately 30.5% reduction in the per-capita conflict rate in a 3-month post-harvest window. The paper’s most novel contribution is a systematic reversal of this relationship across agro-ecological zones: in Non-Sahel countries (Ethiopia, Malawi, Nigeria, Tanzania) the yield effect is negative, consistent with an opportunity-cost mechanism, while in Sahel countries (Burkina Faso, Mali, Niger) higher yields are associated with more per-capita conflict, consistent with a resource-predation channel in which armed groups appropriate agricultural surplus. These findings demonstrate that aggregate estimates mask a fundamental heterogeneity in the direction of the yield conflict relationship and carry direct implications for the design of agricultural interventions in conflict-affected settings.

Download

Share

COinS