| dc.description.abstract | This dissertation seeks to identify causes and consequences of some of the most complex social phenomena, such as civil conflict and climate change. In the first essay I draw on existing theories of labor coercion (Acemoglu & Wolitzky, 2011, Dippel, Greif and Trefler, 2015) to examine how poor labor market institutions, as those present in places where cocaine production takes place in Colombia, prevent low-income farmers to grasp the returns of positive productivity shocks generated by good weather, and instead, witness increasing coca-profiting group confrontations in high productivity areas.
I employed an Instrumental Variables approach together with Fixed Effects estimators to calculate the effect of exogenous variation in productivity on the dynamics of the conflict, to find that citizens security improves in high-productivity period and worsens in low-yield months.
The second essay is a research project with Alberto Abadie, Maurice Kugler and Juan Vargas, where we examine the causal effect of Plan Colombia, the largest US aid package ever received by a country in the western hemisphere, on citizens security (measured by civilians and military killings) and illegal crop acreages in Colombia. To infer the causal effect of the policy on the illegal crop and violence outcomes, we rely on GMM estimators and high-frequency variations in violence. We show that the marginal effect of spraying of one acre of coca reduces the cultivated area by about 11 percent of an acre.
Since aerial spraying may shift coca crops to neighboring municipalities, this results should be interpreted as a local effect. In addition, since the same coca fields are often sprayed multiple times, this figure constitutes a lower bound of the mean eradicating effect of aerial spraying. Our results also suggest that guerrilla-led violence increases both in the short and the long term. We interpret this result as evidence that the guerrilla tries to hold on violently to the control of an asset that is of first order importance for their survival.
In the third essay I seek to understand household adaptation and labor market impacts of extreme weather events in developing countries. This project focuses its attention on labor supply in the developing world – the primary source of household income throughout the world. Also, household allocation of adult and child labor in response to precipitation represents an avenue for exploring potential adaptations that may minimize or worsen the welfare effects from extreme weather events. My econometric results provide evidence of reductions in labor income mainly through an increase in adult unemployment. Individuals try to smooth the loss of labor income by restorting to “forced entrepreneurhip” or self-employment and by sending youth to work. The worst estimate of the loss in real wages per hour is 8% in the rainy season, but this coefficient is most likely an under-estimation of the effect of floods on real wages per hour, as individuals may have been adapting to ENSO and the unavailability of labor market data from the most affected municipalities during the floods of 2010. Finally, estimates of the causal effects of floods are non-linear. While an additional 95th percentile flood raises the probability of unemployment by 0.0026 percentage points, the effect doubles with one additional 99th percentile flood. | en_US |
