This article examines the causal effect that coca leaf production has on the dynamics of the Colombian civil conflict. The project provides credible evidence on the mechanisms that explain how economic conditions due to unforeseeable environmental changes affect ongoing conflict and prevent to stabilize areas prone to conflict. In particular, it will shed light on the effects of particularly positive unexpected economic conditions for coca farmers on illegal group violent actions: how do these groups react to positive unexpected environmental conditions that increase coca production (i.e. effects on attacks to populations, effects on clashes between them and the government)? How are farmers affected by these actions undertaken by the illegal groups (i.e. effects on forced migration and casualties)?

My project also helps clarify the dilemma about poor economic conditions and security. Policies have been formulated based on the premise that better economic conditions improve security. My preliminary results show that under certain conditions, democratically elected representatives’ security is much worse in coca-producing areas when economic conditions are better.

I exploit the random month-to-month variation in precipitation by county in Colombia to calculate if wetter months are associated with higher or lower agricultural yields, and then, use the estimated variation in yield that is due to changes in precipitation to estimate changes in the Colombian civil conflict.

If rainfall affects coca leaf yield and illegal armed groups maximize income from coca, we can expect rainfall changes to affect the dynamic of the conflict. My hypothesis is that these groups will try to control higher-productivity areas in order to obtain higher income from coca. This, in turn, should imply that the drug-producing illegal armed groups will try to expel the government from the higher productivity areas (mayors and councilors), as well as former supporters of their enemies (farmers). However, they will try to retain as many loyal farmers as possible to maximize coca leaf production. Therefore, the effect of coca leaf production on government representatives should be negative (higher productivity should imply more assassinations of government representatives) but the effect on farmers’ security is unknown.

To conduct this analysis, I compiled, on the one hand, a novel database on coca farming production and technology, that spans from 2004 to 2010. I match this information with NASA information on precipitation and a rich set of civil conflict variables. All variables are available at the county level with a monthly frequency, except NASA data which frequency is daily.

This is the first study, to my knowledge, to prove a link between microeconomic estimates of productivity at this geographical level and violence against civilians, combatants and democratically elected representatives.

My results show that changes in productivity caused by weather have an effect in the nature of conflict. When harvests are good, illegal groups fight to take control of coca-producing areas, but violence is targeted toward government forces or elected politicians. Farmers are better-off when coca production is higher as evidenced by large reductions in forced migration during these periods of time.