I read a game theory paper a while ago (can't find a link, but maybe someone will remember it) that suggested a stable small population of defectors (ie, psychopaths) does confer a benefit to the overall group in a cooperative public goods game. if the group becomes accustomed to having zero defectors (ie, excessively trusting), the spontaneous emergence of even one defector will wreak havoc. if there are too many, the whole thing falls apart because there is no cooperation. so the small population of defectors provides something like immunity to the whole group.

could just be a consequence of the model they chose, but interesting to think about nonetheless. I'd be very grateful if someone could remember the name of the paper!