Predicting Crisis in the Global Trade Network

The web of international trade relationships is a natural domain in which to apply the discipline of network analysis. Network analysis can reveal insights into the structures of trade relationship and interdependencies between trade partners that are not immediately evident in a straight out statistical analysis of trade data. Unsurprisingly, there is a large body of work from the last decade on this very subject. However, most of this previous work has been focused on analyzing the structure of the network and its evolution through time on a global scale. Relative little work has been to analyze how the trade network might be shaped locally by transformative local events such as natural disasters, violent conflicts, and financial crisis. In particular, an intriguing question arises: Can we build a model that can identify the occurrence of a local crisis simply by examining the trade network? This is the question that we seek to answer. In the following sections, we summarize and discuss some of the previous literature in the study of global trade networks, as well as papers on statistical and network analysis methodology applicable to such a study. Next, we detail some of our own analysis of the global trade network between the years 2005 and 2014, both on a global and a local level. Finally, we describe our machine learned model for predicting local points of crisis using data and statistic gleaned from the global trade web, and discuss our findings and their implications.