Habituation to atrocity: low-level violence against civilians as a predictor of high-level attacks

‘Habituation to atrocity’ is characterized as an actor’s increased willingness to carry out high-level violence against civilians (VAC) owing to the choice of low-level attacks in an earlier period. We theoretically analyse habituation to atrocity using a rational choice model in which a government, rebel organization or militia group allocates resources to fighting, attacking civilians and identity formation to achieve territorial control. Based upon concepts available in the rational addiction literature, the model generates a demand function for VAC in which substantial additional demand arises owing to the ‘bad habit’ generated by previous atrocities. The model guides our empirical inquiry into VAC for a sample of forty-nine African countries over the period 1997 to 2014. We find that the number of past low-level civilian attacks (even sometimes those involving zero fatalities) significantly affects the number of high-level attacks in the present. We also find that previous low-level civilian attacks sometimes better predict high-level attacks than civil conflict. Our work suggests that regional and global datasets on ‘small’ VAC incidents can serve as valuable early warning indicators of more severe atrocities.