Introducing the International System(s) Dataset (ISD), 1816-2011

We argue that the Correlates of War dataset on sovereign state membership has two weaknesses: a requirement that states maintain diplomatic relations with Britain and France, and a size inconsistency that disqualifies many mid-sized states in the pre-1920 period. As a consequence, entire state systems are excluded from the data, and the total number of states during the nineteenth century is undercounted. After reviewing two other approaches to identifying states, we offer an alternative set of criteria that identifies 100 completely new cases, and a total of 363 states between 1816 and 2011. These modifications reveal several previously overlooked patterns. Most importantly, the global trend in the number of states over time is concave. From a high of 134 in 1816, states declined precipitously in the mid-nineteenth century through the processes of accession, conquest, and unification. This pattern of state consolidation bottomed out in 1912, and states have proliferated since 1945. However, the pattern of state death and state birth varied by region in the nineteenth century. Whereas the state systems of South Asia and Southeast Asia experienced a steady reduction in the number of states, Africa underwent a more dynamic process of state formation, consolidation, and death.