Along with an excellent team of 5 research assistants, I have spent the summer doing all kinds of empirical work: historical digging into colonial articulations of nationalism, quantitative data on security force casualties, case studies of state-armed group interactions, and codings of armed orders, among various other things (like Sri Lankan election turnout from the 1980s). This is all related in some way to my current book project on armed politics in South Asia.
One of my RA’s organized data on police fatalities in Balochistan (from the Balochistan Police website – I’ve been amazed by what you can find if you go looking, if you know where to look).
This is obviously limited in numerous ways: starts in 1979 (after the 1973-77 conflict), is only the Police (not Army or FC), and does not include either militant or civilian casualties. Plus violence alone doesn’t tell us about many important parts of state-armed group interaction. But it does have the virtue of not being reliant on press coverage, which can be very problematic, and it shows us where and when provincial forces have been most involved in some form of direct conflict.
Here’s a graph of fatalities over time; the spike in the mid-2000s is unsurprising given the return of war to Balochistan, but it’s remarkable how much more violent the last decade has been than the previous two decades:
We need to clean up various aspects of the data moving forward (especially to nail it down spatially), but it will be among the first of the datasets that the project will be releasing.