This is a post that I promised to do a while ago… Unfortunately, writer things happened, and it got stuck in the pipeline five months ago. Better to publish the unfinished version than to never publish, so here we are.
The total fertility rate (number of children per women given current lifespan and birth trends) is declining in almost every single country, besides a few post-soviet countries like Kazakhstan. Due to the existence of a general factor of socioeconomic development, low fertility rates coincide with higher levels of income, intelligence, atheism, education, and contraception usage nationally. This is uncontroversial and farily well known. However, it is not clear which of these variables related to economic development cause low fertility rates.
Currently, TFR rates sit at about 1-1.6 in Europe, 0.7-1.2 in East Asia, 1.5-2.5 in Latin America, 4 in sub-saharan Africa, 1.7-3 in the Arab world, and 1.4-1.7 in Anglo countries. Replacement fertility is 2.1, so without immigration, Europe, East Asia, and the Anglo world should expect their populations to fall.
I don’t consider human extinction to be an achievable goal for any country. This is because fertility has a heritability and shared environmental component that adds up to about 0.2, so families who have lots of children will have children with a lot of children. However, if the TFR is low enough, population collapse does begin to look like a plausible scenario. Currently, South Korea has a TFR of 0.76 and a population of 52 million. If we multiply last years births by South Korea’s life expectancy, we can expect South Korea’s population to be 20 million in 2107. Assuming a constant TFR and life expectancy, this projects to roughly 8 million in 2190.
There is no shortage of theories as to why these decreases are ocurring: a decline in the fuck rate, women being lower quality, women entering the workforce, contraception, housing getting more expensive, men being too risk averse, the flynn effect, lack of gender equality, genetic degradation, lowering testosterone, growing infertility, less religious people, people being too wealthy, and people not being wealthy enough.
From an intuitive perspective, I could see there being some truth to most of these ideas — though it’s also important to distinguish which ones are the dominant causes and which ones are background causes. The phrases “genes and IQ correlate” and “parental SES and IQ correlate” are analogous, though one of these variables clearly correlates with IQ more than the other. It’s also important to distinguish between causes of low fertility rates and paths to increasing them — if low T is causing a reduction in fertility rates, it would be pretty easy to increase it. On the other hand, if it is women having careers, then we’re probably fucked.
The most reasonable course of action is to determine what the main causes of the world’s fertility decline are before determining how to solve it, given that finding a reversible cause also inadvertently results in the discovery of a solution. Looking back at the chart of TFR declines by country, these declines are observed in almost every country except for Mongolia, Kazakhstan, and a few others.
In most countries, these declines happened between the 50s and 90s, with a consistent small decline occuring after this. This leads me to believe the main causes must be changes that occurred during this timeframe. This gives us several plausible candidates for the cause of the global decline:
Women having careers
Contraceptives
Increased population density
Urbanization
Irreligiosity
Globalization
Immigration
Now, to review them:
Female careerism/educational attainment
Generally, people like to delay childbirth while they are in university or early in their career, as they would expect the monetary and time cost of a child to be easier to bear when they are more wealthy. Theoretically speaking, this should have a net negative effect on the fertility of women for several reasons:
Women’s ability to concieve is more dependent on age than men’s. From the ages of 25 to 35, the probability of biological infertility in women linearly increases, and then after 35 it begins to spike until the age of 60, when almost all women can no longer concieve. Men’s fertility is also causally affected by age, mostly through the genetic deterioration of the sperm and a decrease in ejeculate volume.

There is a preference for the man to earn more than the woman. This can be noted by the fact that there is sharp decline in the proportion of partners where the wife earns the same amount as their husband. If this is the case, women who have successful careers are less likely to find a partner than those who do not.

Women who have high wages may be discouraged from having children, as they will incur a larger economic penalty from taking care of the child.
There is some evidence that the introduction of women into the workplace has stalled fertility. Within individuals, educational attainment correlates with lower fertility. This relationship only holds within women after you take into consideration factors such as IQ, religiosity, and gender attitudes.
Besides the adjustments for covariates, the relationship between educational attainment and fertility remains within identical twins sampled from the Virginia Twin Registry, indicating that the relationship is causal to some degree.
The mechanism through which this occurrs is rather interesting. Highly educated women have the same amount of births as less educated ones, but much less unwanted or mistimed births.
There is other data on educational reforms that also suggests a causal relationship. Many times, regions implement educational reforms concerning the amount of schooling a child must complete at different times, which can allow researchers to estimate the causal effect of education on fertility. There are several studies like this, such as this one in Argentina and this one in the USA. The latter of these estimates that 3-4 years of education result in one less child being born.
One thing that has puzzled researchers is that female education and childbirth correlate very weakly in Nordic countries, and that the nordic TFR has been relatively high (~1.9) until very recently. This was mainly due to a decrease in the fertility of uneducated people - the fertility of other people was mostly stagnant.
This has resulted in a very large amount of what I call “nordic fertility literature”, which has tried to connect these to some unique features they have. One theory is that gender equality has a non-linear effect on fertility rates — first women have careers, which delays their fertility and forces them to choose between their work and their life. Then, after some amount of gender equality, men begin to engage in more childcare activities to soften the blow of childcare on a woman’s career.
In their defense, there is a decent amount of literature which suggests that there is some truth to this theory. David Hugh-Jones reviews this literature in this post —finding that the cross-cultural literature does validate this theory. For example, countries with higher levels of female labour force participation used to have less births, and the relationship eventually reversed. Looking at the chart more closely, it looks like the TFR of countries with high levels of female labour force participation had stable fertility rates, while those with medium or low levels saw larger decreases.
In addition, countries with lower levels of male-female disagreement in the desire to have a child also tend to have higher levels of male childcare, though the number of countries used to calculate the correlation is rather low.
I have some bad news for those who believed this theory: the fertility rates are reducing in scandinavia. This cannot be attributed to COVID, as the pandemic seems to have been causing a baby boom, not a baby bust. In addition, the birth rate in Sweden is continuing to decline past the pandemic.
There are also other reasons to be skeptical of this theory besides the recent decreases. Within countries, religious groups tend to be more fertile than non-religious ones. For instance, catholics have a TFR of 2.3, while atheists have a TFR of 1.6. In the previous study on IQ and childbirth I cited, the path coefficient from gender attitudes to childbirth was about 0 in every single racial/sexual group except for White women, where conservative gender attitudes were linked to slightly higher rates of childbirth. He also notes that there is evidence that unrestricted unilateral divorce laws cause lower TFR — this is true in both the USA and Europe, corresponding to a decrease of about 0.2.
As for how strong the effect is - based on data from the NLSY79 I gathered, every year a women spends in educational institutions is associated with -0.103 less children, with the average woman in this cohort having 1.84 children. The univariate model predicts that if all women had no educational attainment, they would have 3.22 children on average. Thus, 1.36 units of the TFR decline can be directly attributed to increases in female education within individuals.
Based on Meisenberg’s structural equation model within White women, 50% of this is a direct effect, and the other 50% is due to the effect education has on postponing marriage and increasing income. None of the observed corelation can be attributed to covariates such as gender attitudes, IQ, or religiosity.
###########################################################################
Religion
Religiosity is declining in most countries, and religious people tend to have more children There are various reasons to think this is a causal as well as non-causal relationship:
Religions tend to encourage people to have children - for instance, LDS church leaders encourage their followers to have children.
Religous people tend to have conservative gender attitudes, which may facilitate child rearing.
Encouragement is clearly a factor, religious people from more pro-fertility religions (e.g. Orthodox Jews, Mormons) have more children than individuals of comparable religions. The effect is somewhat pronounced - Orthodox Jews have 1.5 more children than conservative Jews, while mormons have 1.2 more children than other religious groups.
Another piece of evidence pointing to causality is Israel’s incredibly high TFR of 2.9, which has refused to drop despite a decently high national IQ (92), high GDP per capita (PPP adjusted - 55k), and high average years of education (13). Unsurprisingly, this high fertility rate seems to be driven by Haredi Jews and Religious Jews.

Besides this outlier, there is an interesting case study in which the Patriarch of the Orthodox church, Ilia the 2nd, promised that he would personally baptize every 3rd baby born to a married couple in 2007. This promise resulted in an increase of children born to married couples and families with two children.
This appears to have single-handedly raised the fertility rate of Georgia from 1.65 in 2007 to 2.2 in 2015, though it has recently fallen to 1.87 in 2022.
While the evidence in favor of religiosity causally influencing fertility seems to be true, decreases in religiosity do not appear to be largely responsible for the fertility declines we observe in the world today. Using the USA as a case study, fertility rates fell before religiosity - the fertility decline started in the mid 60s, while the decline in religiosity started sometime between the 80s and 00s.
#########################################################################
Wealth
There is a perception that individuals have not gotten wealthier in the United States and Europe ever since the 70s, many of which who cite that wage increases have not tracked productivity increases. This is largely a myth — some of the decoupling can be explained by the fact that inflation is calculated differently for wages and productivity and some of it is due to the fact some compensation is ignored if only wages are used. This doesn’t even take into account that what can be bought with this money — a laptop from the current year is a radio, communication device, printing press, book, and entertainment device in one package.
Regardless, it is somewhat difficult to reconcline the existence of increases in wealth with decreases in fertility. Based on what we know about wealth shocks, it appears that individuals who’s houses appreciate in value tend to have more children - an appreciation of 100,000 is associated was an increase in probability of having a child by 17%. This is also the case for winning the lottery (study 1, replication)
There are also studies that attempt to incentivize fertility using monetary incentives - as far as I can tell, most of these have been failures, though I can’t be bothered to dive into the literature.
###########################
It has been many months since writing I began writing this, and I unfortunately don’t have the patience to comb through every single fertility theory. To summarize, I think the following theories are empirically validated:
Religion/cultural incentives have causal effects on fertility.
Wealth has a small positive effect on fertility.
Education, particularly in women, delays fertility and causes them to have less children.
These theories might have something to them, but need to be researched in detail:
Male involvement in domestic chores increases fertility.
Cultural and social animosity between men and women reduces fertility.
The decrease in the sexual market value of men and women reduces fertility (e.g. people getting more mentally ill, fat, narcissistic).
Side note
As countries industrialize different services get cheaper and more expensive. Products from highly competitive markets such as cars, clothing, and toys become much cheaper, products from somewhat competitive markets such as food and housing rise at the same rate as wages, while products from noncompetitive markets such as education and childcare increase.
When considering whether to have a child, the most relevant costs to consider are housing, childcare, clothing, and medical care. Out of these 4 goods, 2 of them have increased massively in price, 1 has stagnated, and 1 has decreased. In comparison, many of the goods required to live a luxurious life have decreased. Because of this, the cost of living trade off between quality of life and number of children has become larger with time. I tested this hypothesis using regression (details in appendix. The coefficients went in the expected direction, but the p-values are abysmal (0.09, 0.06).
housing cost by country from here
cost of living by country from here
Cost of living did negatively correlate (r = -0.38) with fertility, but cost of living also correlated with gdp per capita at 0.7, so it doesn’t really mean anything.
Hereditarianism may promote fertility. You can ignore childcare stuff if you are hereditarian. Send baby to kibbutz or nursery inc. and move on...
"In addition, these associations continued to hold when controlling for gender inequality, female education, contraception, religiousness, social norms, GDP, economic inequality, HDI, amount of white collar jobs, pathohens, climate, precipitation, migration, kinship, and sex ratio."
Also temporal autocorrelation:
Let's say there's a trend over time where a country is slowly declining in fertility and people are also slowly packing themselves into a couple of cities; if there's a particular year where there's a sudden spike in such density which ignores the otherwise-linear trend, this corresponds to a sudden drop in fertility for that year.