Last semester I took a political science seminar on international conflict. I wanted to learn more about international conflict. I blogged about it here, here, and here. Unfortunately, the class was oriented primarily around mainstream (i.e., empirical and quantitative) political science, so I didn’t learn as much as I would have liked. Everything I learned must be qualified by the fact that questionable methods, epistemology, and datasets were used for all of the research we studied.
Another downside to the class was that I had to do an empirical, quantitative research paper. So I did the paper on the democratic peace thesis. I “tested” the thesis at the system level of analysis, meaning that I tested for the effects of the percentage of democracies in the international system on three types of war: inter-state war, intra-state war, and extra-state war. The idea was that if democracies are not supposed to go to war with each other, then a higher percentage of democracies in the international system should be correlated with a lower incidence of war.
I found virtually zero support for the democratic peace thesis and even found some support against it. In a number of the models, the percentage of democracy in the international system was positively and significantly correlated with the incidence of intra-state (i.e., civil and secessionist) wars.
Of course, war here is measured according to the coding rules established by the Correlates of War Project as conflict that results in at least 1,000 battle deaths. My dataset ranged from 1816-1997.
I’m not at all surprised by the results. Joanne Gowa, in Ballots and Bullets, argued that the democratic peace was an artifact of the Cold War; it appeared to be true only because Western, capitalist, democratic nations had a shared security interest against the Soviet Union. My professor, David Sobek, though he argues that Gowa’s book suffers from methodological deficiencies, improved on Gowa’s methods in an as yet unpublished paper and was surprised to find her results confirmed.
Again, while I am wary about making any definitive claims based on empirical, quantitative evidence regarding social phenomena, the evidence against the democratic peace thesis is continuing to grow. And, more importantly, this empirical evidence is supported by strong theoretical arguments.
My own paper can be found here.