I think I’ve come around more to believing the original event studies, also with the new paper on infertile women that shows that they basically have the same outcomes as men.
Do you know of any studies that use spontaneous twins as an instrument? That would be an interesting source of randomness (controlling for age, since likelihood of twins increases with maternal age) to identify the effect of >1 child. It could also potentially identify the effects of maternity leave, since the vast majority of employers only give one leave per pregnancy, not per child.
Thank you for pointing this out. I do have some concerns, however. I am not an economist, so I will skip all the economics jargon.
1) I don't think the reasoning behind "the curve flattens and then woman, because trade-offs change, start to have children" is correct. I know it is an oversimplification, but there is something fundamentally wrong here: that the effects of potential childbearing take only effect at the moment of birth, which is not correct. Woman start to think about pregnancy, and consequently make choices that after their work and economic outcome, way before they become pregnant. And this becomes even more prominent as the woman approaches the end of her reproductive life. So this flattening might just reflect the fact that the woman indeed is considering having a child and (before conceiving) is conditioning her choices to that, even if in the end she doesn't become pregnant. I think most of these studies don't take into account the cost that just being a potential childbarer has on a woman.
2) Similarly, in the case of IVF in Norway. The IVF procedure in itself has a cost for a woman, it is a burden. The hormones, the appointments, the failed procedures, the emotional distress. All of these impact life and work decisions, even if there is no child at the end. I don't think IVF+no child is a good control group.
3) Let's say you take Norway, and accept its 9% gender pay gap. We can assume this gap is related to career choices. In many cases, as it has been shown several times, this happens because woman take over most of the burden (even in developed countries) of taking care of children and elderly, informal work that is not being remunerated. So, if you take the total hours worked by a woman (formal+informal), and the salary she perceives, the gender pay gap becomes even bigger. I think a serious study of this problem MUST include informal work to be realistic, especially when we are looking at gender pay gap and the incidence of it in one of the genders is much bigger. It would be also be interesting to know if the burden of this informal work, which often happens irrespective of having a child or not, contributes to the flattening of the earning curve.
4) I might be wrong, but most of these studies assume that if there is no difference in pay between men and woman, particularly before having a child, there is no gender pay gap. I think this somehow leads to the assumption that a temporarily non-child bearing woman is equivalent to a man. I don't think that is correct. There is a different contrafactual: a hypothetical women who doesn't have to endure gender-related stuff, who would earn perhaps more than men. If woman are earning the same even after having their periods every month, being expected to care for others in their families, fighting with a biological clock, perhaps they would be earning more if they didn't have to go through these gender-bias-related stuff.
I wrote about this last year as well when BHL was still a working paper!
https://open.substack.com/pub/illuminatingfertility/p/how-much-would-mothers-earn-if-they?utm_campaign=post&utm_medium=web
I think I’ve come around more to believing the original event studies, also with the new paper on infertile women that shows that they basically have the same outcomes as men.
There’s an ESR paper that uses the ‘classical’ approach but accounts for public transfers to parents in Denmark: https://academic.oup.com/esr/advance-article/doi/10.1093/esr/jcaf036/8326715
Do you know of any studies that use spontaneous twins as an instrument? That would be an interesting source of randomness (controlling for age, since likelihood of twins increases with maternal age) to identify the effect of >1 child. It could also potentially identify the effects of maternity leave, since the vast majority of employers only give one leave per pregnancy, not per child.
Yes, for example:
https://publica-rest.fraunhofer.de/server/api/core/bitstreams/91bfab51-b051-46f3-91c4-730c21551182/content
and:
https://www.jstor.org/stable/2117765
That said, twin instruments aren't seen as credible anymore as they once were.
Thank you for pointing this out. I do have some concerns, however. I am not an economist, so I will skip all the economics jargon.
1) I don't think the reasoning behind "the curve flattens and then woman, because trade-offs change, start to have children" is correct. I know it is an oversimplification, but there is something fundamentally wrong here: that the effects of potential childbearing take only effect at the moment of birth, which is not correct. Woman start to think about pregnancy, and consequently make choices that after their work and economic outcome, way before they become pregnant. And this becomes even more prominent as the woman approaches the end of her reproductive life. So this flattening might just reflect the fact that the woman indeed is considering having a child and (before conceiving) is conditioning her choices to that, even if in the end she doesn't become pregnant. I think most of these studies don't take into account the cost that just being a potential childbarer has on a woman.
2) Similarly, in the case of IVF in Norway. The IVF procedure in itself has a cost for a woman, it is a burden. The hormones, the appointments, the failed procedures, the emotional distress. All of these impact life and work decisions, even if there is no child at the end. I don't think IVF+no child is a good control group.
3) Let's say you take Norway, and accept its 9% gender pay gap. We can assume this gap is related to career choices. In many cases, as it has been shown several times, this happens because woman take over most of the burden (even in developed countries) of taking care of children and elderly, informal work that is not being remunerated. So, if you take the total hours worked by a woman (formal+informal), and the salary she perceives, the gender pay gap becomes even bigger. I think a serious study of this problem MUST include informal work to be realistic, especially when we are looking at gender pay gap and the incidence of it in one of the genders is much bigger. It would be also be interesting to know if the burden of this informal work, which often happens irrespective of having a child or not, contributes to the flattening of the earning curve.
4) I might be wrong, but most of these studies assume that if there is no difference in pay between men and woman, particularly before having a child, there is no gender pay gap. I think this somehow leads to the assumption that a temporarily non-child bearing woman is equivalent to a man. I don't think that is correct. There is a different contrafactual: a hypothetical women who doesn't have to endure gender-related stuff, who would earn perhaps more than men. If woman are earning the same even after having their periods every month, being expected to care for others in their families, fighting with a biological clock, perhaps they would be earning more if they didn't have to go through these gender-bias-related stuff.