That's not the stance though. The stance is that administrative costs isn't what matters. Impact is what matters. Sometimes impact requires more administration, sometimes it doesn't. But impact is something that we can measure, so why use a bad heuristic instead?
