Method for Forecadting Ladge Protests (the сase of South America)

The technique for predicting major conflicts tested at South American countries in 1996–2020 is based on the assumption of an increase in social tension in recent years before the event. At the same time, an increase in social tension is considered not as a direct cause of protests, but as a marker indicating their higher probability in the presence of trigger events. Time series of homicides and suicides are used as indicators of social tension. To assess the dynamics of social tension, the concept of “conflict level” is introduced – a five-year growth rate of the indicator approximated using linear regression. Exceeding a certain threshold of the conflict level is interpreted as a sign of a high probability of mass protests. The optimal threshold selection is based on the condition of maximum sensitivity with a positive value of the conflict level. Correlation analysis of the conflict dynamics level showed that the countries of South America are divided into two groups. The first group is Chile and the countries that build “socialism of the 21st century”; the second – the main members of MERCOSUR. The retrospective forecast demonstrated a fairly high sensitivity of the method (81%) confirming its applicability for early social-political crises warning. The analysis of unpredicted mass protests causes allowed us to put forward a hypothesis explaining why they could occur at a low level of conflict.