Analyzing Localized Commodity Returns in the Context of Terrorist Attacks: A Two Phase Approach

Abstract

In this thesis, we investigate the relationship between terrorism and localized commodity returns. We begin with a set of over 100 large-fatality events and implement a two phase methodology on the log returns of the major commodity export of each terrorism-affected country. Phase I isolates extreme events and Phase II tests said events using forecasted confidence intervals from fitted GARCH(1,1) models. Ultimately we demonstrate that there are three types of instances where historical returns are not encompassed by forecasted intervals, which we refer to as exceptions. We provide a political and statistical analysis of these exceptions as well as potential avenues for further investigation.