Insurgent Compensation: Evidence from Iraq

File Type: 
File extension: 
Publication Year: 
File size: 
Additional Authors: 
Benjamin W. Bahney, Howard J. Shatz
Citation Information: 
American Economic Review: Papers & Proceedings 103(3): 518-522

Participating in insurgency is almost always physically risky, raising questions about what motivates people to do so. Political economy theories of crime, insurgency, and rebellion posit some opportunity cost constraint so that the net value of participation must be at least as good as the next best option. That net value may con- tain a nonpecuniary component, such as soothing personal grievance or gaining prestige, but it also contains monetary rewards. Sometimes these rewards are implicit: the ability, usually of middle managers, to take advantage of one’s position to skim funds or tax civilians and then pay low-level fighters with a share of the profits. Typically, however, the monetary rewards are explicit: Wages are paid as in other jobs. How those wages vary has implications for counterinsurgency policies and provides evidence regarding the agency problems the compensation scheme is designed to solve. We report initial results from an examination of insurgent compensation paid during the recent war in Iraq. The war provides a unique opportunity to study this phenomenon as the US military collected a large number of insur- gent financial records. Combining this informa- tion with other data allows us to analyze how compensation varied with levels of combat, eco- nomic opportunities, and other conditions that may affect decisions to participate in violence. We analyze data on 3,799 payments to insur- gents in al Qa’ida in Iraq (AQI), one of the main insurgent groups and the most extreme, across three governorates—Iraqi provincial-level sub-national units—in 2006 and 2007. We find no evidence compensation was based on risk or even that it reflected the marginal product of labor. Indeed, there is modest evidence for a negative compensation-risk relationship at the governorate-year level.