Re-examining Economic Shocks and Civil Conflict

Using annual rainfall variation as an instrumental variable for growth, we show that economic growth is negatively related to civil conflict in Africa. Antonio Ciccone (2011) argues that thanks to rainfall's mean-reverting nature, rainfall levels are preferable to annual changes. We make three points. First, our findings hold using rainfall levels as instruments. Second, Ciccone (2011) does not provide theoretical justification for preferring rainfall levels. Third, the first-stage relationship between rainfall and growth is weaker after 2000, suggesting that alternative instruments are needed when studying recent conflicts. We highlight the accumulating microeconomic evidence that adverse economic shocks lead to political violence.

本研究以年降雨量波动作为经济增长的工具变量（instrumental variable），实证发现非洲地区的经济增长与国内冲突呈显著负相关关系。安东尼奥·西科恩（Antonio Ciccone）2011年的研究指出，由于降雨量具有均值回复特性，相较于降雨量年度变动量，降雨量水平更适合作为工具变量。本文提出三点核心论点：其一，采用降雨量水平作为工具变量时，本文的核心结论依然稳健；其二，西科恩（2011）并未为优先选用降雨量水平提供理论依据；其三，2000年以后，降雨量与经济增长之间的第一阶段回归关系有所减弱，这意味着在研究近期国内冲突时，需要采用替代性工具变量。本文还梳理了不断积累的微观经济学实证证据，表明负面经济冲击会引发政治暴力。