The timing of ETA terrorist attacks

Terrorist attacks occur in an apparently random manner. However, as terrorist groups are rational, there must be an underlying pattern behind terrorist attacks. This article analyses the duration of peacetime periods between ETA's terrorist attacks, in order to clarify the expectations related to terrorist attacks. We employ parametric and semiparametric hazard model specifications to ascertain the covariates, which explain the ETA terrorist attacks, and the underlying duration dependence. The results are rather robust among different specifications of the hazard. As a result, terrorist attacks seem to increase in summer and decrease with respect to other variables, namely deterrence and political variables.