Might be worthwhile to point out that even if it should turn out to be true, this doesn't imply that male CEOs were paying their female employees unfairly before they had daughter. It would just show that they paid women more afterwards than before (which could be discriminating against men, for example). Perhaps they even just paid men less, or earned less themselves (perhaps daughters require more time spent with family or whatever), thus dragging down the averages. I have not read the actual PDF yet, though, only the newspaper summary.