The previous chapter showed that different levels and forms of social organization can reduce levels of civil war violence suffered by civilians. Intertwined with this relationship are the questions of where social organizations come from and where they are likely to arise. This chapter addresses these questions and does so for three main reasons. First, knowing where social organizations arise in a developing country is itself an interesting question, with implications for social and economic development. I find that geographic, demographic, and social factors all play roles in the formation and spread of local junta councils (qualitatively, I also consider promotion efforts by government personnel and private actors).

Second, understanding the causes of juntas helps deal with the threat to valid inference of reverse causality – that junta councils might only exist and survive in historically peaceful places. Introducing the indicator of the violence suffered during the La Violencia conflict of the 1950s shows that juntas were actually more likely to be formed later in the 1960s and 1970s in areas that suffered relatively more violence in the prior conflict, easing concerns of spurious correlation. These statistical tests along with a close reading of Colombian history between the period of La Violencia and when juntas are measured – both up to 1960 and from the 1960–1985 interim period – show that conflict conditions were low and relatively stable through most of the country. The tendency for juntas to be formed in historically conflicted areas was therefore not likely disrupted during this period.

Third, with a better understanding of junta origins I am also able to unite the statistical analysis with the case studies to better rule out confounding explanations. An ideal research design to study this question would be an experiment that randomly assigns juntas to communities.