 As companies are being asked to by their investors, by their employees, by government, to actually go through their entire wage systems and see if there are any wage differentials for a woman of the same education, the same occupation, obviously working in the same firm, are their differences. And they discover, yeah, there are differences. And then a number of firms having found out there are differences have now taken some action to eliminate them. But the fact that there were differences after you already counted for all these other variables suggested that there was a discriminatory or bias practice. So let's go now to some other areas. Women choose different occupations from men, and that's true around the world. And occupations like teachers pay less than occupations like engineers or lawyers in general. So occupational choice is part of it. And then you can ask questions about well why do women choose the occupations they choose? Are there not enough role models for women in higher paying occupations? Do women feel there are barriers to their advancement in those occupations? Women around the world are much more likely to do part-time work than men. And part-time work for the same kind of job, the same kind of occupation, the same sector, a part-time job does not pay as well as a full-time job. So if you are choosing part-time work, choosing a male occupation, choosing a male sector, but choosing a part-time job, you're more likely to see a woman's compensation wage be lower than a full-time man. Another thing that is remarkable, I think, around the world is that there's evidence of what you would call a motherhood penalty. So if you're in the workforce, let's take advanced industrial economies where actually now the education gap is largely gone. If you look at any part of the job market at the beginning of men and women entering the job market, you're not going to see much of a differential within an occupation or a differential within a sector. Men and women for the same level of education in the same occupation in the same sector working full-time, that wage difference has largely disappeared, same wage. If you then project forward five years, ten years, you see the arrival of children. You see very much a significant growth in what was no wage gap into a growing wage gap. Now that could be because, again, women decide that they're going to move to a part-time status. They're going to basically be removed from the sort of the career pathway so they can have more career acceleration pathways so they can have more time for care. Indeed, in some societies there's evidence of a male premium, so basically if you start out with no differential, you're in a family, the woman then sees a wage penalty over time. Her wage doesn't grow as much as she has a family. For the male, frequently the wage goes a little faster. This is really quite remarkable. A parenthood premium for males, a motherhood cost for females. We have unreported in the world of work the fact that there is a significant amount of work associated with parenthood with running a household and raising children. That work exists and right now, not just in advanced industrial countries but around the world, that work is disproportionately done by women and it's not paid. One policy area we can now see actually begins to distribute the burden more equally across parents is paternity leave. So lots of societies were doing maternity leave and what was happening was women were leaving the workforce, taking the maternity leave. By the time they went back into the workforce, they were already at a wage differential and they stayed there. What if you actually give fathers a paternity leave? What if you actually say to them, you must take this leave, otherwise you're not going to get parenthood leave, there's no discrimination, men and women take this leave. What we've discovered in societies that have that, that actually it's one of a whole set of policies which together, you can't pick out any single one, but which together seem to really address these wage differentials.