This paper leverages a large-scale dam construction project in Southeastern Turkey to test the hypothesis that climactic shocks and their impacts on rural incomes play a causal role in civil conflict. I use original 5km and 10km gridded datasets on irrigation and Kurdish separatism from 1985-2019, exploiting exogenous topographical variation in the distribution of irrigation schemes. I find that conflict incidence and insurgent recruitment decline significantly in areas following the introduction of irrigation. A district-level analysis suggests the decoupling of agricultural income from rainfall as a likely mechanism. Clashes are more frequent following a poor harvest, and yields for all major crops except irrigated cotton are highly sensitive to rainfall. However, there is substantial treatment heterogeneity related to land inequality. Consistent results were derived from both cross-sectional Instrumental Variables approaches and spatial panel models, and were robust to the inclusion of a wide array of highly detailed political economic control variables, alternative measures of irrigation and conflict, and fixed effects.
Authors
Ollie Ballinger
Oxford Department for International Development