The Perils of Third-Party Counterinsurgency Campaigns

Abstract

What explains the paucity of victory among states fighting counterinsurgency campaigns overseas?

Guerrilla strategies are regarded by many in the West as universally successful. But while rebels fighting as insurgents do have somewhat higher odds of success than those adopting other strategies, states fighting what many would refer to as traditional civil wars still defeat insurgents well over 70% of the time. However, states fighting insurgents overseas have rather less luck. Here the trend is markedly reversed and creates something of a paradox: while many third-party states enjoy significant resource advantages over their indigenous counterparts, they are very rarely able to translate that material superiority into political victories in counterinsurgency campaigns. In the empirical analysis offered below, I find only six instances of outright third-party victories over insurgent opponents.

How then do states defeat insurgent opponents? And, more specifically, how do third parties fighting overseas conduct counterinsurgency campaigns? Relying on a dense review of the secondary literature and developing novel data for regression analysis, I argue the answer ultimately lies in the intersection of selection effects and intelligence: third parties only face the most resolved insurgents and typically in unfavorable information environments. This then limits both their war-fighting and war-termination strategies. The availability of intelligence is in many ways what distinguishes states fighting indigenous insurgencies from those third parties or colonial powers fighting overseas. Ultimately, I argue that third parties win when they're able to overcome these intelligence challenges before public support runs out. This typically requires rather substantial military reforms and complex deal-making with local leaders. Unfortunately, the nature of selection effects in these cases gives rise to a population of insurgencies whereby these conditions are very unlikely to be met.