Transitory Economic Shocks and Civil Conflict

Is civil conflict triggered by transitory economic downturns? Answering this question is complicated by many difficult-to-measure economic, social, political, and institutional factors that may affect both income and the likelihood of civil conflict. In addition, economic downturns may partly be driven by the expectation of future civil conflict. Miguel, Satyanath, and Sergenti (2004) show how both these issues can be addressed. Using a panel-data setup to control for unobservables, they show that civil conflict in Sub-Saharan Africa is more likely to start following years of low rainfall growth and that low rainfall growth is also associated with low income growth. Based on their instrumental-variables estimation results, Miguel, Satyanath, and Sergenti conclude that negative (rainfall-driven) income shocks trigger civil conflict. This is in line with Collier and Hoeffler (1998, 2002, 2004) and Fearon and Laitin (2003), who have argued that economic variables are often more important determinants of civil conflict than political or social grievances.

Estimating the effect of transitory economic shocks on civil conflict involves an interesting issue that has not been analyzed so far. Consider Miguel, Satyanath, and Sergenti’s finding that civil conflict is more likely to start in year t, the lower rainfall growth between t-1 and t-2 (reproduced in Table 2, Panel B, column (7)). This result is consistent with conflict outbreak being most likely when a drought year at t-1 (a negative rainfall shock) follows an average rainfall year at t-2. But the result is also consistent with the view that conflict at t is most likely to start when exceptionally good rainfall conditions at t-2 (a positive rainfall shock) are followed by an average rainfall year at t-1.¹ The general point is that because rainfall shocks are transitory, low interannual rainfall growth between t and t-1 could be due to a negative rainfall shock at t or due to mean reversion following a positive shock at t-1.2 The finding that civil conflict is more likely when rainfall growth is negative is therefore consistent with conflict being more or conflict being less likely following droughts.

It is important to understand whether civil conflict starts following positive or negative economic shocks. If civil conflict follows negative transitory shocks, conflict onset may be related to the temporarily low opportunity cost of fighting during transitory economic downturns. This explanation of the effect of income on civil conflict was suggested by Collier and Hoeffler (1998) and is also put forward by Miguel, Satyanath, and Sergenti (for a theory of the link between transitory shocks and civil conflict, see Chassang and Padró i Miquel, 2006 and 20073). But rent-seeking explanations of civil conflicts appear more plausible than opportunity-cost explanations if conflict is more likely following positive economic shocks.

Download economic journal article: ziddu