Friday, May 28, 2010

Conservatives are outbreeding Liberals

Journalist Steve Sailer made the important point that white fertility was substantially higher in Red States than Blue States, something he calls "The Baby Gap". I would like to update and expand on his insight.

First, here is the correlation between total fertility per women in 2008 and McCain's vote share in 2008. The association is strong. In the latest available figure the total fertility rate of non-Hispanic white women is 1.86. You need 2.05 for replacement levels. The total fertility rate of Utah is 2.7, compared to 1.7 in Vermont.


On occasion patterns that are true in the aggregate are not true at the individual level. For example, rich states vote much more Democrat, but voting for President Obama in 2008 was actually negatively associated with income. The reason for the paradox is the *within* Red states in particular and to some extent also within blue state, income is associated with voting Republican.

For this reason I have evaluated the hypothesis using micro-level data. The source is National Annenberg Election Survey, which is essentially a super-poll of tens of thousands of Americans, with lots of questions.

I am looking at women 40-45 to get a good proxy for lifetime fertility. At this age mostly the children have not left home yet, and above this age the probability of having more children is extremely low.

Overall, conservative non-hispanic white women have 26% more children than liberal white women, a sizable difference.

The very conservative have 45% more kids on average than the very liberal.


So what does this imply? Political preferences are strongly related with parental views. This is probably not genetic. The likelihood to vote has been shown to be to some extent genetic, but not who you vote for. The link is cultural. Still how the ideology is transmitted is not important, just that it is.


Let us assume that if you have conservative parents, you become conservative with 70% probability and liberal with 30%, and vice-versa. With this number if you start off equally divided between liberal and conservative (there are more of the later, but many moderates are really liberal who don't want to admit it) in 1 generation the right will be 10% larger than the left.

Of course at the same time because of immigration and very high Hispanic (especially Mexican) birth rates, the Hispanic share is rising. Hispanics and African American voters are far to the left of middle, especially on fiscal issues.


The higher Conservative birth rate slows the demographic ascent of the current, liberal leaning Democrat party. However liberal white voters, who are ideologically very committed, will be replaced by equally left leaning but less ideological minority voters.

Based on these 2 long term demographic trends whites will become increasingly republican and the Democrat party increasingly minority.

Of course these long term trends are dominated by short term events. In 2004 Kerry won the overall young vote (18-29) by 9 points, while Bush won the white young vote by 11 points. In 2008 President Obama won the young vote by 34 points, and the white young vote by 10 points (54-44%).

20 comments:

  1. Thanks.

    How big are sample sizes?

    It would be interesting if you have the data to look at number of children and marriage and see what's more closely related to political views. They're so closely related that it can be hard to tease out whether its marriage or children that is the big factor.

    ReplyDelete
  2. Sample size is 72000.

    I will look at marriage next.

    ReplyDelete
  3. This comment has been removed by the author.

    ReplyDelete
  4. Off Topic: You might find this article of interest. In fact I first assumed it was one by you.

    Spiked are formed of an odd group of ex-revolutionaries that appear to have done a complete about face.

    ReplyDelete
  5. So, half of 72000 are women, maybe 10% of voters are 40-45, so that's 3600. Sounds reliable, but maybe you could just double the age range to 40-50 to double the sample size.

    ReplyDelete
  6. This comment has been removed by the author.

    ReplyDelete
  7. Yeah the sample of women 40-45 whose ideology and number of children is known is 5500.

    If you look at women 40-50 the conservative have 21.5% more children than the liberal.

    (it makes sense that the difference is smaller, since conservative women have children earlier in life, and many of them would have moved out by the time the mom is 50).

    For women 35-45 conservatives have 27% more children than liberals.

    70% of conservative women this age are married, compared to 56% of liberal women. The numbers are near identical if we take women 40-50.

    Married conservative have 19% more children than married liberals (35-45), and 13% more if aged 40-50.

    In a regression of conservatism, both the number of children and marriage are strongly statistically significant for both men and women, for all 3 age ranges. I will do this a little more advanced in a future post.

    ReplyDelete
  8. In addition to TFR you might want to see what effect generation length has on the demographic balance. If conservative women have 2 children before the age of 30 and liberal women have 2 children between the ages of 30 and 45, and their daughters replicate this pattern, then over the span of a century those conservative women will be seeing the arrival of their great, great-grandchildren while the liberal women will be seeing the arrival of their grandchildren.

    ReplyDelete
  9. As a side note, I wonder if it makes sense to say something like "26% more children." That's always been an issue with things that happen discretely. The very conservative are 50% more likely to have two or more children than are the very liberal. It also acknowledges the marginal nature of the decision to have children: do we want one more?

    ReplyDelete
  10. Excellent work. But Is that first graph measuring overall vote and overall TFR? I don't think anyone alleges high-fertility blacks will vote Republican. You've got measure the white-TFR to white-vote-share. [Your original source link is broken].

    You need 2.05 for replacement levels

    Where did you hear that? I've only ever heard 2.1 for a first-world state.

    ReplyDelete
  11. What are the statistics on political conversion rates? Children don't always stay with their parents on politics in adulthood or even as teenagers. What if some factors make conversion one way more likely than the other?
    And what about mixed households? Or what if one or both parents are independent, moderate, libertarian, or some other political label that is not liberal or conservative?

    ReplyDelete
  12. Nice post!! keep it up!!Female infertility is an extremely sensitive issue and is increasing in India due to numerous reasons. These are hormonal problems, scarred ovaries, premature menopause, and follicle problems etc. its quite difficult to select best female fertility supplement in Availability of many fertility treatments.

    ReplyDelete
  13. :)We don't kill our children,maybe thats why our birthrate is higher...

    ReplyDelete
  14. The excellency of your post compelled me to give a comment. Your post is awesome. Keep up the good work. We are always here to support you.
    Funny Pictures

    ReplyDelete
  15. I love surfing the internet because of all things I get to profit from it. I always find out something brand-new! Thanks for uploading! I wish you have even more information posts in the near future! http://miguelholmes.blog.fc2.com/

    ReplyDelete
  16. talks about it I'm curious to find out what blog system you have been utilizing? that guy I'm having some minor security problems with my latest blog and I would like to find something more secure. Do you have any solutions? talking to

    ReplyDelete
  17. I would imagine this is a push though. Liberals tend to be better educated which is why they choose to have fewer children. And the more educated the more likely you are to vote. So it makes sense that more and more female voters are liberal which is shown election cycle after election cycle.

    ReplyDelete
  18. What about religion? Protestants have less children and are mixed politically. Jews are similar and they vote liberal. Muslims vote liberal and have an extremely high birth rate. Catholics are mixed and also have high birth rates. Also, changing party attitudes are not infrequent. Most Southern Democrats of the 60 and 70s now vote Republican.

    ReplyDelete

Google Analytics Alternative