Violence against civilians 1900-87: A four-part analysis using av new dataset to test the relationship between the severity of democide and type of government

Abstract

The study of violence against civilians with its over 200 million dead in the 20th century is understudied if viewed in terms of the death toll compared to the studies of inter-state and civil war. R.J. Rummel’s extensive data collection of democide in the period 1900-87, has unfortunately lain dormant, mostly untouched by other researchers since its publication in 1997. This thesis has made use of this dataset, and converted it into time-series format in order to facilitate multivariate analyses. Instead of focusing on one type of violence against civilians, the thesis splits it into four different analyses on democide; one for democide performed by regimes against its own population; one for democide in a foreign state; one performed by non-state groups; and one category with the total for all of the other three. This division proves fruitful, as the three types of democides yield different results. The analyses show that autocracies are most violent when it comes to regime democide, with no difference between democracies and semi-democracies. Analyses on foreign democide, however, show that democracies are the least violent type of government, with no clear difference between autocracies and semi-democracies. Non-state democide, on the other hand, shows no clear difference between the three types of government. The role of economic development is also an interesting finding. While less-developed countries are more violent when it comes to regime democide, the exact opposite is found on foreign democide.