You're right that the broken window fallacy is a fallacy when the economy is not in a depression, but it can be a bit more situational and complex than that. W

You're right that the broken window fallacy is a fallacy when the economy is not in a depression, but it can be a bit more situational and complex than that. When aggregate demand is "stuck" at a depressed level (high unemployment causes low demand, resulting in more low unemployment), the stimulus of war can actually help pull the economy out of its stuck position, even if the war is destructive. Of course, once it's no longer stuck, the continued destructive stimulus is just destructive. There are better fiscal ways to pull the economy out of a depression than war, because most wars don't pay much in the way of dividends, unlike large scale public works. But even if the stimulus isn't a great ROI in "normal" times, it can have a high ROI in "depressed" times due to its ability to get demand unstuck.