Three essays on rural development

This dissertation combines research on three topics related to rural development in low-income countries. It is motivated by my experience as a farmer and extension agent in Colombia and the conviction that efforts made to address these challenges can significantly improve the lives of rural communities. The first chapter evaluates the impact of partial conflict resolution on legal and illegal economic activity. The second chapter, based on the joint work with Mary Paula Arends-Kuenning and Anina Hewey, tests the potential of the teacher-student-parent channel of information exchange built around vocational training programs for the diffusion of agricultural technology. The third chapter, based on joint work with Sandy Dall´Erba, estimates the effect of climate variability on future coffee productivity in Colombia. The first chapter links the occurrence of positive or negative outcomes after a conflict resolution process to the configuration of armed group presence that existed in a locality before the peace talks started. Specifically, I look at two possible configurations: (i) local bilateral conflicts where the legitimate state is engaged in a contest with a single armed group in the process of demobilization and (ii) local multilateral conflicts where the legitimate state opposes two armed groups, one which is demobilizing and one that continues its activity unabated. Identification is based on a difference-in-differences design that leverages the heterogeneity in the configuration of armed group presence and estimates the effect of the configuration on economic activity, intensity of violence, and coca leaf cultivation. The results show unequivocally positive outcomes after the demobilization of the armed group in local bilateral conflicts evidenced by a decrease in the number of violent incidents, and an increase in economic activity, and no evidence of an increase in coca cultivation. In contrast, I find evidence of negative outcomes in local multilateral conflicts where coca cultivation increased and no significant increase in economic activity is observed. I argue that these opposing results relate to the shift (or absence thereof) of the local monopoly of violence. In the former case, the legitimate state regains the monopoly of violence and is able to control illegal activity more effectively. In the latter case, the state still fights for control of the monopoly of violence with the remnant armed group, which in turn is able to consolidate the illegal activity around itself after the competitor armed group is removed from the scene. The second chapter consists of a randomized control trial (RCT) to assess the potential of using the teacher-student-parent channel of information exchange for the diffusion of agricultural technologies to underserved rural communities. We posit this channel as a complement to the increasing literature on social learning and the farmer-promoter models that looks at increasing farmers’ exposure to new technologies on the extensive margin (by reaching more farmers through the extended networks of the promoters) and the intensive margin (by increasing the frequency of the interactions between farmers and promoters). The RCT was conducted among high school students - and their parents – to whom the Fabretto Foundation was gradually offering the vocational training program known as Tutorial Learning System (SATec). By comparing the results of knowledge-based tests and survey data we were able to show that the knowledge of agricultural and accounting techniques improved among students and parents. More importantly, we find evidence that those techniques were being adopted by the parents who were also more likely to access credit markets to finance these new endeavors. The last chapter builds a bridge between the crop physiology literature and the econometric estimation of the effects of climate change on agriculture. We build a simplified agronomic model that links changes in temperature and precipitation to the productivity of coffee. The insights from this model are used to estimate a dynamic panel model that accounts for the perennial nature of coffee production. After showing that this model outperforms conventional econometric models, we use it to predict future coffee productivity in Colombia using up-to-date forecasts of future weather for 2041-2060. The results show that national coffee productivity won’t be affected on average, but that careful attention must be paid to the heterogeneity of the predictions at the local level. In particular, we show that low-altitude municipalities will likely experience a decrease in productivity, whereas high altitude municipalities will increase productivity