That looks right to me! Those results are consistent with our results -- it just depends on how you think about expectations. I wonder what the ratio of F 20-60 / M 30-60 looks like (which would match my framing above better)?

Plus, in contemporary polygynous populations, men marry women 5-10 years younger on average, so maybe the bands should be shifted from each other. But, cutting the female group off at 45 also means that you treat men who do marry a wife of a similar age as unmarried once they age past 45 (2/2)

I think that code should work! One question is whether you think that all men 20-60 and all women 20-45 would expect to be married, or whether they just expect to be married at some point in their lives. People who divorce or get widowed can marry single people (1/2)

Thanks for adding this! I'm curious, are there any papers reporting an adverse causal effect on child health that you're methodologically confident in?

Happy to talk through the evolutionary implications of the work! We focus a lot on the political science implications in the write up. @rebeccasear.bsky.social might have more thoughts

But the essence of our argument is actually that sex ratios don't matter a ton, at least in the populations represented by our census data. For example, controlling for sex ratios doesn't explain away the negative association between polygyny and unmarried men (4/4)

I think your point raises the issue of reverse causality, which we talk about! Ember (1974): conflict could promote polygyny by skewing sex ratios. I'm just not sure it makes sense to characterise human sex differentials in mortality as mostly due to competition (3/4)

www.jstor.org/stable/3773112

For example, here's a classic bit of demography on 1950s-60s African marriage markets that comes to that conclusion using a counterfactual approach that's different in spirit from ours (2/4)

bsky.app/profile/hgga...

The results show that it's also due to age gaps at marriage and men marrying 2nd+ wives at older ages (like we see in the real world). And it's definitely possible to get the result that there are no/few excess men without assuming (big) mortality differentials (1/4)

Looking forward to it! Sorry for the maybe frighteningly quick reply haha, your posts were at the top of my feed 🙃

We don't model pairing preferences, partly because tractable assumptions about preferences do weird things, like your model shows with all the unmarried women. But you can see how we try to think through preferences in supplement S2.1: we argue that they should ↑ the sustainability, not ↓ (3/3)

If you want a cool throwback to some really old school demography, Etienne van de Walle (1968) made a very similar point to us (pics attached but let me know if you want the PDF, it's hard to find). Goldman & Pebley is another classic source here (OA link: publishing.cdlib.org/ucpressebook...) (2/3)

This is interesting, thanks! Like you said further down in your thread, I'm wondering how the two different approaches lead to such different results. Building on refs. 35-40 in the paper, we model "availability ratios" of male-female age pairings, rather than whole marriage market (1/3)

@patrickwallis.bsky.social is a good place to start for the historical perspective! Since economic history is apparently in vogue now 🙃

blogs.lse.ac.uk/lsereviewofb...

One of the polygyny → bad outcomes papers jumps through hoops to justify China as polygynous, inferring from TFR in 1950 (bad idea) that ~7% of men practised it and saying that polygyny since time immemorial shaped contemporary institutions anyways (doubtful, I guess?) (2/2)

doi.org/10.1017/ssh....

That's so interesting (and typical, I guess). The meme of men thinking about the Roman Empire comes to mind (1/2)

Thanks!

The demography is tricky for sure! And thanks! We took some direct inspiration from your work trying to debunk "if parent marry, child won't suffer poverty" arguments (referenced in the conclusion)