Abstract Many countries have embarked on decentralization as a way to mitigate separation tendencies. Even though decentralization could satisfy some separation tendencies, the central government might be unable or unwilling to award the localities the level of decentralization they require to stay in the union. This mismatch may give rise to domestic violence and internal conflict. We propose a simple two-stage model that combines the decentralization and domestic terrorism literature in a single structural model, which can be readily tested using existing data in the disciplines. As long as the actual decentralization level meets the unobserved desired value by the localities, there will be a high likelihood of observing a zero incidence of domestic terrorism. Otherwise, the incidents happen regularly with mitigating effects from decentralization. Our model also accounts for the left-censoring of the terrorism data at zero. We examine our model using 5-year interval panel data of more than 60 countries based on the information for 1970–2019. Our results indicate a strong negative and statistically significant relationship between fiscal decentralization and domestic terrorism events. One percent increase in fiscal decentralization is associated with a two to three percent decline in domestic terrorism incidents. The coefficients of the selection model and domestic terrorism model generally conform to the literature.
Read full abstract