On determinants of national suicide rates: evidence from Bayesian model averaging

We aim to establish relative importance of socioeconomic, demographic, geographic and other determinants of national suicide rates. To this aim, we apply Bayesian model averaging (BMA) approach to a dataset of 27 potential determinants in a cross-section of 173 countries. Life expectancy at birth, ambient temperature, age dependency ratio and religious affiliation were found to be the most robust protective factors. Life expectancy at age 65 and unemployment rate are the most robust determinants that are positively associated with suicide mortality.