I think the issue here is they're imputing goodness of a population by total or average happiness or some other tricks to make their happiness calculus work.