Quote Originally Posted by Steely Glint View Post
That's an odd choice. Why not defense spending? Do you include reservists or just active personnel?
Because I don't think defense spending captures the concept of militarization. I defer to Mullins (1987): expenditure data has the following disadvantages: replacement cycles cloud trends, defense budgets are calculated differently across time and space, inflation rates aren’t fully accounted for, and these don’t take into account arms transfers; military personnel is a better measure because: soldiers aren’t subject to radical technological change, and cross-country comparisons are much easier.

It's active personnel. I do actually have military expenditures in the current model as a kind of control (might not keep it), and its effects are very ambiguous. Plus there's a vast arms races literature, and I don't really want to get into that area.