Trade and War

In this thesis the relationship between trade and war is researched, in particular the possi­ble pacifying effect of trade. Most data is obtained through the Correlates of War project. A panel dataset of state-dyads is constructed which contains information on militarized interstate disputes, international trade, contiguity, capabilities, interstate alliances, dis­tance, GDP, trade system membership, religion and language. Various case studies will be discussed alongside a panel-wide analysis. Causality is inferred through a series of (panel data) vector autoregressive models and Granger Causality tests. The relationships are quantified using a pro bit model on conflict and an ordinary least squares regression on real bilateral trade. Evidence for the pacifying effect of trade is found, for a 1 % decrease in the probability of conflict real bilateral trade has to increase by 81.16% on average. This effect can be enhanced through a higher intensity of bilateral trade. The reliability of the models that are used is questionable however. Endogeinity and multicollinearity impede the probit model, whilst heteroscedasticity and serial autocorrelation impede the panel data vector autoregressive model. Relevant and genuine exogenous instruments for real bilateral trade could possibly solve these issues.