Quantifying the Relationship Between Economic Crises and Political Unrest

It is often assumed that banking crises (a phenomenon wherein a country’s banking services do not have sufficient liquidity to cover all demand) have a relationship with political instability. However, political scientists have had difficulty quantifying this relationship, due to the many factors that play a role in these events. Using economic and political data provided by Dr. Daniel Hansen (CMU, IRP), our paper aims to build on preliminary research in this area, specifically focusing on more advanced statistical modeling techniques. This will allow us to predict revolutions at a high accuracy rate and identify variables which are strongly related to the onset of political instability. 

We construct our models on our entire dataset rather than a subset of specific countries in order to focus on broader trends in the geopolicial climate and also learn more about which factors in particular are influential in predicting unrest. We constructed additional time series models such as the Hidden Markov Model to understand whether this played a role in the onset of revolutions. 

Over the course of this project, we conclude that we can predict political instability using economics data, and we identify a number of variables in our dataset relating to economics, geography, and government structure that are strong predictors of political instability in countries around the globe.