André De Palma, Federico Perali, Nathalie Picard, Roberto Ricciuti, Alexandrina Scorbureanu
August 7, 2013
This study presents a novel approach to crisis prevention based on data on premonitory political and religious events and the international media coverage of publicly sensitive circumstances. We implement our method to the Israel-Palestine conflict. First we identify two main political scenarios associated with “good” and “bad” political times of low or high levels of political unrest using a hierarchical clustering technique. Then we construct a political alert index to predict the probability of occurrence of good and bad times. Bad times are positively and significantly associated with the number of Israeli victims at the checkpoints, the number of homeless or injured Palestinians and with the number of demolitions. The number of Palestinian prisoners and injured Israelis negatively affect the probability of occurrence of a bad time. Media coverage is positively and significantly associated with the transition to bad times. Our results show that our statistical tool can be a reliable method for early warning of social crisis and can be effectively replicated to other social crisis situations.