Working Paper: NBER ID: w16461
Authors: Edward Miguel; Shanker Satyanath
Abstract: Miguel, Satyanath and Sergenti (2004) use rainfall variation as an instrument to show that economic growth is negatively related to civil conflict in sub-Saharan Africa. In the reduced form regression they find that higher rainfall is associated with less conflict. Ciccone (2010) claims that this conclusion is 'erroneous' and argues that higher rainfall levels are actually linked to more conflict. In this paper we show that the results in Ciccone's paper are based on incorrect STATA code, outdated conflict data, a weak first stage regression and a questionable application of the GMM estimator. Leaving aside these data and econometric issues, Ciccone's surprising results do not survive obvious robustness checks. We therefore conclude that Ciccone's main claims are largely incorrect and reconfirm the original result by Miguel, Satyanath and Sergenti (2004), finding that adverse economic growth shocks, driven by falling rainfall, increases the likelihood of civil conflict in sub-Saharan Africa.
Keywords: rainfall shocks; economic growth; civil conflict; Sub-Saharan Africa
JEL Codes: N47; O55; Q54
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| rainfall levels (Q54) | civil conflict (D74) |
| Ciccone's claims about positive relationship between rainfall levels and conflict (Q34) | erroneous (Y60) |
| original findings by Miguel, Satyanath, and Sergenti (2004) (H84) | robust against criticisms by Ciccone (Y30) |
| adverse economic growth shocks due to falling rainfall (Q54) | likelihood of civil conflict (D74) |
| negative rainfall shocks (Q54) | civil conflict (D74) |