Can Events Predict Violent Intra-State Crises?


  1. Tikuisis, P.
Corporate Authors
Defence R&D Canada - Toronto, Toronto ONT (CAN)
Events data were analyzed for their predictive potential as indicators of occurrences of politically violent intra-state crises marked by adverse regime changes, revolutionary wars, and ethnic wars. An average of 22 randomly-selected destabilizing and stabilizing events per month were analyzed for 25 crises between 1990 and 2003 inclusive selected from those identified by the Political Instability Task Force (PITF) as onsets of state failure. Events covered a 12-month period preceding a crisis plus the month of crisis for a total of 7147 events. Subject matter experts analyzed these events according to various CIFP-defined state power and state performance characteristics, and scored them according to an impact assessment involving casuality, centrality, and escalation. Sixty-eight percent of crises did not indicate trends in any event characteristic leading up to the crises. Of those crises where trends were significant, there was no consistency in the nature, frequency, intensity, and direction of the event characteristics that did change. This study concluded that randomly-selected events analyzed using the CIFP construct of state power and state performance factors do not readily reveal imminent politically violent intra-state crises identified as state failure onsets by PITF during the 12 months leading up to the crisis. Recommendations for extracting a deeper disaggregation of events data for prediction purposes are discussed.

Il y a un résumé en français ici.

Political violence;Failed state;Early warning;Crisis forecast;Political Instability Task Force;PITF;Data analysis;Humanitarian crisis
Report Number
DRDC-TORONTO-TR-2010-189 — Technical Report
Date of publication
01 Jan 2011
Number of Pages
Electronic Document(PDF)

Permanent link

Document 1 of 1

Date modified: