This study tested conventional wisdom about how insurgencies end against the evidence from 89 insurgencies. It compares a quantitative and qualitative analysis of 89 insurgency case studies with lessons from insurgency and counterinsurgency (COIN) literature. While no two insurgencies are the same, the authors find that modern insurgencies last about ten years and that a government's chances of winning may increase slightly over time. Insurgencies are suited to hierarchical organization and rural terrain, and sanctuary is vital to insurgents. Insurgent use of terrorism often backfires, and withdrawal of state sponsorship can cripple an insurgency, typically leading to its defeat. Inconsistent support to either side generally presages defeat for that side, although weak insurgencies can still win. Anocracies (pseudodemocracies) rarely succeed against insurgencies. Historically derived force ratios are neither accurate nor predictive, and civil defense forces are very useful for both sides. Key indicators of possible trends and tipping points in an insurgency include changes in desertions, defections, and the flow of information to the COIN effort. The more parties in an insurgency, the more likely it is to have a complex and protracted ending. There are no COIN shortcuts.

"Prepared for the Marine Corps Intelligence Activity"--T.p.

Introduction -- Classifying outcomes and selecting cases -- Assessments of insurgency endings: time and external factors -- Assessments of insurgency endings: internal factors -- Assessments of insurgency endings: other factors -- Conclusions -- Appendix A: case studies, methodology -- Appendix B: supplemental findings -- Appendix C: multivariate regression analysis -- Appendix D: insurgencies not examined for this publication -- Appendix E: categories use for the spring 2006 survey -- Appendix F: unavoidable ambiguities -- Appendix G: questions use for the autumn 2006 survey.