Identifying Robust Determinants of the Civil War Risk Using Bayesian Approach

The civil war will have many negative economic and non-economic consequences. In this regard, the main purpose of the current research is to identify the robust determinants of the civil war risk in 127 countries of the world (including Iran) during the years 2001-2020 and using the Bayesian Model Averaging (BMA) approach. For this purpose, 21 variables have been used in 5 groups: 1. Economic variables, 2. The abundance of natural resources, 3. Demographic variables, 4. Social variables, and 5. Political variables. By estimating 100,000 regressions and Bayesian averaging of coefficients, robust determinants of the civil war risk have been identified. Based on the results, the variables the good govrnance index with a 99% probability and negative effect, religious and racial tensions with a 97% probability and positive effect, and human development index with a 96% probability and negative effect have a definite and robust effect on the civil war risk. The index of economic conditions and the total population with a 91% and 76% probability, respectively, have been the fourth and fifth important variables in explaining the civil war risk with a positive effect. Other variables have a weak or insignificant effect on the risk of civil war. Based on this, it can be said that countries with higher quality institutions, higher human development indicators, less religious and racial tensions, a more favourable economic situation and a smaller population will have a lower probability of civil war.