Let's not get it twisted: if the headline were something to the effect of "the Colombian government paid its soldiers to murder whomever they could credibly claim were rebels" and the econometric analysis were the same, this wouldn't necessarily be so dunk-worthy (that said AEA decided to be dicks and paywall the paper so I haven't had a chance to read it yet). Without a penalty for killing false positives, of course that'll result in an awful incentive structure with a horrible outcome!
Of course since we're talking about a paper coauthored by Acemoglu - an MIT economist who famously described the European settlers' rape and pillage of the Americas and Africa as a "reversal of fortune", euphemistically and underwhelmingly describing colonization and slavery as "changes in institutions" - interpreting the spike in the state-murder outcome in Colombia as the result of mere incompetence producing "bad incentives" is hardly surprising.