Violence, Insurgency, and the Collective Action Problem By STATHIS N. KALYVAS and MATTHEW ADAM KOCHER Lea Štranjgar žUsing data from Vietnam war and the Greek Civil war authors revisit the collective action paradigm žThe collective action paradigm has two gorunds: ž1. the free-riding incentive generated by the public goods dimension of insurgency ž2. the risk of individual participation in insurgent collective action žThe collective action problem (Olson, 1960s) žIt pays attention to violence žFocuses on the costs associated with insurgent participation žObstacles in launching and sustaining insurgencies à ‘’first movers’’ or ‘’late joiners’’ žCosts of nonparticipation and free riding often equal or exceed those of participation žLinear, concave and convex functions of participations žAuthors claim: insurgent participation is much less dangerous relative to nonparticipation than is posited by the collective action paradigm and than generally thought žWar differs from other types of violence: different targeting – higher individual risk žRisk also varies across specific types of warfare (conventional, irregular war) žIdentification problem à selective and indisciriminate violence žVietnam war – rate of victimisation of participants in violent collective action žAnalysis of the U.S. Phoenix Program data žAttempt of the U.S. and South Vietnamese governments to target the Vietcong – TO AVOID CIVILIAN VICTIMS ž‘’whom we killed was far more important than how many we killed’’ žSystematic record of intended victims ž ž73, 697 individuals (members of Vietcong) as targets and 21% of those killed žUnknown method of selecting victims à those most likely to be innocent were most likely to be victimized žTwo groups: ž1. those on the list: ž confirmed - 5,88% killed + captured ž unconfirmed – 52,53% killed + captured ž2. those off the list ž žHOW? žInoccents were easier to find žVietcong agents were able to escape capture or assassination žFull party members were less likely to be captured but more likely to be killed žConfirmation took time: everyone entered the base as ‘’unconfirmed’’ ž‘’Better to kill mistakenly than release mistakenly’’ ž žExtensive archival sources à exact toll of violence in Argolid (region in southern Greece) žProcommunist resistance army ELAS vs. right-wing collaborationalist militias supported by the German occupation troops ž žTotal rural civilian population – 45, 140 žEstimate of military-age men -13,542 ž žEstimate of local rebel combatant – 500 žEstimate of local militia combatants - 300 žMale victims of militiamen/Germans - 318 žLocal rebels killed in action – 20 ž žCivilian victims of rebel violence – 353 ž žA civilian man of military age was more likely to be killed by one of the two sides than was an actual combatant à it was safer to be combatant than civilian žFree riding in civil wars is not attractive option which term implies žSometimes the real puzzle in civil wars is nonparticipation rather than collective action žRecruitment increases when the state is totally absent OR under conditions of extreme violence