Real average data is necessary but not sufficient. Policy should also consider the real outliers it may affect. So while you could argue that normalizing parental leave is unnecessary because children are better off with their mothers, we should do that while recognizing that it is not true for all family units. E.g. if there is no mother, or as in the examples I provided for the person I was replying to, there is a mother but breastfeeding is not possible.
Beyond that, if you adopt a policy that considers these outliers and normalize parental leave, there will be parents where breastfeeding is possible but the fathers still take the parental leave.
Beyond that, if you adopt a policy that considers these outliers and normalize parental leave, there will be parents where breastfeeding is possible but the fathers still take the parental leave.