We assess the consequences of the 27 year long Angolan civil war on child health,
household expenditures, educational attainment, and fertility by exploiting geographical
variation in the intensity of violence experienced by different communities. We
generate exogenous variation in the intensity of conflict using the distance separating
each community from the main rebel group’s headquarters as our exclusion restriction.
Even after controlling for other geographical characteristics that might also
directly affect our response variables, conflict intensity is found to be a decreasing and
convex function of the distance to UNITA headquarters in Jamba, Cuando Cubango
province, as predicted by a simple rent-seeking model of civil war. Instrumental variables
estimates, based on household survey data collected in 2000, indicate that, in the
short-run, conflict intensity worsens child health, does not significantly affect household
expenditures, increases school enrollment and decreases fertility, as would be predicted
by a Neoclassical unitary household model. We find significant differences between
the short- and long-run impacts of conflict intensity on our response variables, and we
uncover non-linearities in the impact of conflict intensity at different levels of violence
using a nonparametric two-stage least squares procedure