Essays on globalization and conflictthe impact of income and environmental shocks in Africa

Abstract

This thesis considers the importance of spatial patterns and the use of geo-localizeddata in panel and repeated cross-section data econometrics by addressing causality issues to obtain further insights into the causes and consequences of con ict in a globalized world. Chapter Two analyzes the link between globalization and the incidence of civil conflict in a panel dataset of 159 countries over the period 1972-2009. Distinctions are drawn between several dimensions of globalization identi ed in political economy literature, i.e. economic, social, and political globalization. I address the potential endogeneity of the globalization variables by introducing country-fixed effects into the analysis. I also use a novel spatial instrumental variable based on the degree of integration of neighboring countries. Chapter Three uses geo-localized information to study the ethnic drivers of food-related income shocks and their e ects on conflict in Africa to explain underlying conflict processes. Thus, I propose the use of a panel database of a full grid of African countries divided into sub-national units of 0.5 per 0.5 degrees of latitude and longitude (10,638 cells) that covers the period 1998-2013. The study contributes to the relevant literature by analyzing several competing theories on the e ects of income shocks on conflict, using geo-localized data which considers the interaction between those income shocks and ethnic diversity. Finally, Chapter Four examines the environmental damage that conflict may cause, such as oil spills in Nigeria and their impact on agricultural production. Thus, I use a consumer-producer household framework to explain how oil-spill pollution might result in changes in the optimal behavior of households. I estimate an agricultural production function using repeated cross-sections of micro-data geo-referenced for farming households and four waves of data from 2009 to 2018 taken from the Nigeria General Household Survey (GHS-Panel). To calculate a proxy for oil spill pollution, I create a function that uses geospatial data with information on around 12,000 oil spills from the Nigerian Oil Spill Monitor.