I think this does show there is a pretty low marginal return to the "last hour" spent with a child. It seems like halving the amount of time you spend (by adding another child and splitting your time evenly) roughly decreases educational attainment by 0.2 years , which is kind of underwhelming.
I like your point here and I'm a little annoyed at myself that I didn't think about things this way. Taking the numbers from Price's paper, 2nd born kids get 3000 fewer hours of quality time ages 4-13. The lifetime earnings impact is ~2.5% from the other Sandra Black paper I cite. If lifetime earnings are in the range of 2,000,000 that would imply about $17 of kid's future income per additional hour of quality time spent with the parent, i.e. not that big a marginal product of labor. This is all super rough but it does kind of support your reaction.
I thought about this a bit more and I'm not so sure. For parents with college degrees, their kid born today may well have undiscounted lifetime earnings of $4 million (see https://www.ed.gov/sites/ed/files/policy/highered/reg/hearulemaking/2011/collegepayoff.pdf but note they use 2009 dollars and there's economic growth hence $4 million). I think it might make sense to use undiscounted earnings in this calculation for reasons I won't go into here. Moreover the 2.5% income effect of one rung in birth order from that paper I cite strikes me as a lower bound on the impact of parental time, because most other factors seem likely to favor later-born kids (parental experience, learning from older siblings, and also later-born kids are more biologically healthy at birth). So, say the impact of parental time were more like 3.5%. Then we'd be talking about 3.5% of $4m so $140k for 3k hours so about $50/hour. $50/hour is not a bad wage for a parent with a college degree -- that's like a $100k/year salary. And for most people high-quality time with kids is more enjoyable than time spent doing market work. So I'm still a bit on the fence about how "big" these impacts are.
I agree that the 4m number makes more sense than 2m. But I do think we should do some amount of discounting, would be interested in your thoughts there. I would be less interested in (for example) a windfall for my kids that appears when they are 65 vs one that is spread out over their adult lifetimes (assuming same value in real dollars).
One question I had (and I didn't read any of the papers ...) was why they only count hours from age 4 on? I have 3 kids (5,2 and 2months) and they take a lot of time before age 4, some of which is intuitively very formative. And the arrival of a new baby is obviously a time of relative "neglect" (eg more TV, least individual instruction) for older kids.
Super interesting post. Thanks for writing this and sharing.
As a psychologist I can’t quite contain my need to clarify the author’s relative misuse of “good enough parenting” (though I completely understand that the author is referring to it in the way that parents following this mindset are colloquially using it - AND who are also misunderstanding the original meaning). It absolutely does not mean more relaxed parenting in terms of providing less attention or time or some arbitrary measurement of ’good enough.’ It also doesn’t mean less hypervigilance or less effort. What it means is accepting that inevitably we will all make tons of mistakes as parents. But given the combination of a genuine best effort and the opportunity to repair any harms done that sting longer than the moment, the parenting will be ultimately be good enough. The term comes from a psychoanalyst named DW Winnicott. And it’s a beautiful perspective that I think when held well alleviates some of the guilt and worry parents can feel about what they might have done that was more negatively impactful then they anticipated or could mitigate, while also setting up the mindset that you can still repair wounds. Winnicott also argues that those very same ‘harms’ when repaired between parent and child are tremendous learning opportunities that are ultimately more beneficial than having done it perfectly in the first place.
Thanks for clarifying that's useful to know. That does sound different and more appealing to me than the colloquial way I've seen it used in child development discussions.
Hi Nate, and happy holidays! I found this really interesting and although I know (in theory) this isn’t how statistics works, I find it pretty accurate in observing friends and folks around me.
Something this made me think about is the impact of the age gap between me and my (younger) sister — almost 13 years. In effect, my parents spent 25 years dealing with the K-12 system. I’ve found this to be somewhat common among immigrants from China in my age group, where the older child was born under the one child policy, and the second was born much much later after the parents take some years to establish themselves in the US. Personally, this has resulted in a fairly different upbringing that my sister experienced, in terms of both technology/environment and also our parents’ attitudes. But as it relates to your post, I wonder how age gaps affect this finding. It seems to me like on one hand, with a large age gap, parents will not be choosing to read I Am a Bunny to both kids at once, but more likely to be reading to the younger while the older is at soccer or something, plus the younger will typically have more economic resources. On the other hand, the parents are older, and maaaybe parenting less anxiously because they’ve seen the first one go through the same milestones. My guess is that it makes a small difference despite this. But I disclaim that I really don’t know. :)
This is interesting! Footnote 23 in Black et al 25 (https://academic.oup.com/qje/article-abstract/120/2/669/1933962) doesn't find any impact of spacing, and also points out this is hard to study because parents choose spacing for different reasons. For some families as you describe, large spacing relates to careful planning. For other families, large spacing may indicate accidents. I think to really study this well it would help to see a study on complete administrative data in a much larger country than Sweden or Norway e.g. the US or China or something.
This resonates with me. My kids are 6 years apart and I feel like my youngest benefits from what I've learned. For example, I didn't even understand what fine motor skills were when my son was little, but watching him struggle with tying his shoes made me lean into to building those skills with my youngest. I also think my son functions like a mini parent proving more opportunities for conversation and engagement. He will even read her the occasional book. Time will tell, but I actually find myself feeling bad that I couldn't give my son all the things we give my daughter when he was little.
Well written and I agree - one key thing I think was left out of the equation, is how much attention and education the 2nd- 11th child (I used that number because our grandfather had 10 younger siblings on a small Ohio farm- and they ALL achieved college and further graduate degrees!) receive from each other AND from simply acutely observing their parents and older siblings. I don't think we need to worry about how much time we spend directly parenting - simply watching parents get work done, organizing their day, cooking, caring for others, etc. is very valuable time for learning. Too much attention makes kids a bit insecure and dependent, perhaps, anyway.
Super interesting, Nate. You assign 100% of the effect to parent inputs — I didn't even think of that!
Ex ante, it's not obvious who should benefit. The firstborn gets more early childhood parent inputs, but the later-borns get more informed and experienced parents. The second born doesn't go to that awful preschool that you thought would be fine!
My prior for birth order effects was similar to age-in-class effects. Older kids in their grade tend to do better — they're bigger, stronger, smarter, more socially skilled, they develop more confidence which could count for a lot. Inside a family, the older sib is like the red-shirted kid — always getting the confidence boost of being the best at all the within-sib stuff.
Curious if you think there's additional evidence that says this is about parent inputs.
That's a different interesting prior -- I can't dismiss that. Here's why I'd bet on parental inputs instead.
1. The evidence on class size and child care quality and mentoring make it seem hard to imagine why thousands of fewer hours of high-quality parental time wouldn't adversely affect kids.
2. The evidence on peer effects typically shows that being around higher-skilled peers is beneficial, which should advantage the later-born siblings. This is consistent with some mild evidence suggesting that mixed-age day cares are better for younger kids but worse for older kids.
3. I suspect red-shirting helps kids partly because teachers/coaches consider them more capable compared to their peers without mentally controlling for age, and then invest more heavily in them on that account even if just implicitly through tracking. So we have reverse intuitions on this.
But I agree the peer comparison channel is very possible -- so your view is toning down my prior. But to me that feels more subtle and second-order than the big decline in high-quality parental time and the normal peer effects. You'd need the impacts of the "look how good I am at stuff compared to my younger siblings therefore I should aspire to really great things" psychological effect to outweigh the more standard peer effects AND you'd have to believe high-quality parental time has little impact.
I think this does show there is a pretty low marginal return to the "last hour" spent with a child. It seems like halving the amount of time you spend (by adding another child and splitting your time evenly) roughly decreases educational attainment by 0.2 years , which is kind of underwhelming.
I like your point here and I'm a little annoyed at myself that I didn't think about things this way. Taking the numbers from Price's paper, 2nd born kids get 3000 fewer hours of quality time ages 4-13. The lifetime earnings impact is ~2.5% from the other Sandra Black paper I cite. If lifetime earnings are in the range of 2,000,000 that would imply about $17 of kid's future income per additional hour of quality time spent with the parent, i.e. not that big a marginal product of labor. This is all super rough but it does kind of support your reaction.
I thought about this a bit more and I'm not so sure. For parents with college degrees, their kid born today may well have undiscounted lifetime earnings of $4 million (see https://www.ed.gov/sites/ed/files/policy/highered/reg/hearulemaking/2011/collegepayoff.pdf but note they use 2009 dollars and there's economic growth hence $4 million). I think it might make sense to use undiscounted earnings in this calculation for reasons I won't go into here. Moreover the 2.5% income effect of one rung in birth order from that paper I cite strikes me as a lower bound on the impact of parental time, because most other factors seem likely to favor later-born kids (parental experience, learning from older siblings, and also later-born kids are more biologically healthy at birth). So, say the impact of parental time were more like 3.5%. Then we'd be talking about 3.5% of $4m so $140k for 3k hours so about $50/hour. $50/hour is not a bad wage for a parent with a college degree -- that's like a $100k/year salary. And for most people high-quality time with kids is more enjoyable than time spent doing market work. So I'm still a bit on the fence about how "big" these impacts are.
I agree that the 4m number makes more sense than 2m. But I do think we should do some amount of discounting, would be interested in your thoughts there. I would be less interested in (for example) a windfall for my kids that appears when they are 65 vs one that is spread out over their adult lifetimes (assuming same value in real dollars).
One question I had (and I didn't read any of the papers ...) was why they only count hours from age 4 on? I have 3 kids (5,2 and 2months) and they take a lot of time before age 4, some of which is intuitively very formative. And the arrival of a new baby is obviously a time of relative "neglect" (eg more TV, least individual instruction) for older kids.
Super interesting post. Thanks for writing this and sharing.
As a psychologist I can’t quite contain my need to clarify the author’s relative misuse of “good enough parenting” (though I completely understand that the author is referring to it in the way that parents following this mindset are colloquially using it - AND who are also misunderstanding the original meaning). It absolutely does not mean more relaxed parenting in terms of providing less attention or time or some arbitrary measurement of ’good enough.’ It also doesn’t mean less hypervigilance or less effort. What it means is accepting that inevitably we will all make tons of mistakes as parents. But given the combination of a genuine best effort and the opportunity to repair any harms done that sting longer than the moment, the parenting will be ultimately be good enough. The term comes from a psychoanalyst named DW Winnicott. And it’s a beautiful perspective that I think when held well alleviates some of the guilt and worry parents can feel about what they might have done that was more negatively impactful then they anticipated or could mitigate, while also setting up the mindset that you can still repair wounds. Winnicott also argues that those very same ‘harms’ when repaired between parent and child are tremendous learning opportunities that are ultimately more beneficial than having done it perfectly in the first place.
Thanks for clarifying that's useful to know. That does sound different and more appealing to me than the colloquial way I've seen it used in child development discussions.
Hi Nate, and happy holidays! I found this really interesting and although I know (in theory) this isn’t how statistics works, I find it pretty accurate in observing friends and folks around me.
Something this made me think about is the impact of the age gap between me and my (younger) sister — almost 13 years. In effect, my parents spent 25 years dealing with the K-12 system. I’ve found this to be somewhat common among immigrants from China in my age group, where the older child was born under the one child policy, and the second was born much much later after the parents take some years to establish themselves in the US. Personally, this has resulted in a fairly different upbringing that my sister experienced, in terms of both technology/environment and also our parents’ attitudes. But as it relates to your post, I wonder how age gaps affect this finding. It seems to me like on one hand, with a large age gap, parents will not be choosing to read I Am a Bunny to both kids at once, but more likely to be reading to the younger while the older is at soccer or something, plus the younger will typically have more economic resources. On the other hand, the parents are older, and maaaybe parenting less anxiously because they’ve seen the first one go through the same milestones. My guess is that it makes a small difference despite this. But I disclaim that I really don’t know. :)
This is interesting! Footnote 23 in Black et al 25 (https://academic.oup.com/qje/article-abstract/120/2/669/1933962) doesn't find any impact of spacing, and also points out this is hard to study because parents choose spacing for different reasons. For some families as you describe, large spacing relates to careful planning. For other families, large spacing may indicate accidents. I think to really study this well it would help to see a study on complete administrative data in a much larger country than Sweden or Norway e.g. the US or China or something.
This resonates with me. My kids are 6 years apart and I feel like my youngest benefits from what I've learned. For example, I didn't even understand what fine motor skills were when my son was little, but watching him struggle with tying his shoes made me lean into to building those skills with my youngest. I also think my son functions like a mini parent proving more opportunities for conversation and engagement. He will even read her the occasional book. Time will tell, but I actually find myself feeling bad that I couldn't give my son all the things we give my daughter when he was little.
Thanks this is very interesting please see my reply above to Anqi's comment!
Well written and I agree - one key thing I think was left out of the equation, is how much attention and education the 2nd- 11th child (I used that number because our grandfather had 10 younger siblings on a small Ohio farm- and they ALL achieved college and further graduate degrees!) receive from each other AND from simply acutely observing their parents and older siblings. I don't think we need to worry about how much time we spend directly parenting - simply watching parents get work done, organizing their day, cooking, caring for others, etc. is very valuable time for learning. Too much attention makes kids a bit insecure and dependent, perhaps, anyway.
Super interesting, Nate. You assign 100% of the effect to parent inputs — I didn't even think of that!
Ex ante, it's not obvious who should benefit. The firstborn gets more early childhood parent inputs, but the later-borns get more informed and experienced parents. The second born doesn't go to that awful preschool that you thought would be fine!
My prior for birth order effects was similar to age-in-class effects. Older kids in their grade tend to do better — they're bigger, stronger, smarter, more socially skilled, they develop more confidence which could count for a lot. Inside a family, the older sib is like the red-shirted kid — always getting the confidence boost of being the best at all the within-sib stuff.
Curious if you think there's additional evidence that says this is about parent inputs.
That's a different interesting prior -- I can't dismiss that. Here's why I'd bet on parental inputs instead.
1. The evidence on class size and child care quality and mentoring make it seem hard to imagine why thousands of fewer hours of high-quality parental time wouldn't adversely affect kids.
2. The evidence on peer effects typically shows that being around higher-skilled peers is beneficial, which should advantage the later-born siblings. This is consistent with some mild evidence suggesting that mixed-age day cares are better for younger kids but worse for older kids.
3. I suspect red-shirting helps kids partly because teachers/coaches consider them more capable compared to their peers without mentally controlling for age, and then invest more heavily in them on that account even if just implicitly through tracking. So we have reverse intuitions on this.
But I agree the peer comparison channel is very possible -- so your view is toning down my prior. But to me that feels more subtle and second-order than the big decline in high-quality parental time and the normal peer effects. You'd need the impacts of the "look how good I am at stuff compared to my younger siblings therefore I should aspire to really great things" psychological effect to outweigh the more standard peer effects AND you'd have to believe high-quality parental time has little impact.