Atheists Just Don't Have Many Kids

And it's not because they are Democrats

One thing I love about studying religion is that it impacts every aspect of social life. How people vote, where they live, what kind of jobs they pursue, are influenced in some way by their religious beliefs and behaviors.

One choice that is clearly shaped by religion is when (and if) adults have children and how large they want their families to be.

In the United States, the fertility rate in 2008 was 2.06 children per woman. In 2023, it’s projected to be 1.78 children per woman. Forty-nine out of fifty states had a lower fertility rate in 2020 compared to 2010 (North Dakota was the only one to buck the trend.)

Obviously, there are a ton of factors that lead to a drop in fertility. Economics is usually considered to be a leading culprit for a drop in fertility. The Great Recession is supposed to lead to an enrollment cliff in higher education in the next five years because people decided to delay pregnancy.

But here’s another explanation that may be playing a noticeable role in the drop in American fertility: the increasing secularization of the United States.

Judaism, Christianity, and Islam all encourage their adherents to marry and have children. But lots of Americans don’t adhere to those faiths anymore. I wrote an entire book (actually two of them) about the rising number of Americans who reject religion entirely.

Does this actually matter, though? Do we see in the data a difference in parenting rates for atheists compared to Latter-day Saints, for instance? The answer is: yes, without a shadow of a doubt. Atheists are the least likely to be parents compared to any other religious group and it’s reasonable to infer that some of the drop in American fertility is due to rising secularization.

The Cooperative Election Study asks folks, “Are you the parent or guardian of any children under the age of 18?” Now, this question wording makes it impossible to understand what share of atheists will never have children but it still can give us some insight into how few of them are currently parents.

I compared to the most fertile religious group (Latter-day Saints) to the least fertile (atheists). The differences in parenting rates are gigantic. For instance, about 60% of 30-year-old Latter-day Saints are currently parents to young children. For atheists, it’s just below 25%.

Peak likelihood of being a parent tends to be between 30 and 45 years old. The absolute peak for Latter-day Saints is late thirties. 72% of them are parents. Atheists tend to have their children a bit later with their peak being in their early forties. But even then, just 40% are parents.

Let me put this plainly: there is no age in which the majority of atheists are parents to children under the age of 18. For the LDS, a majority of them are parents to children between the ages of 27 and 49.

Obviously, this comparison is incredibly crude. Latter-day Saints tend to be Republicans, while atheists tend to be Democrats. And maybe LDS are just this weird outlier when it comes to fertility. I wanted to simplify this a bit, so first I restricted the sample to just those between the ages of 30 and 45, that’s when people are the most likely to have young children at home. Then I calculated the share who were parents among Republicans and Democrats at every level of worship attendance.

There are two trends that clearly emerge here. The first is that people who attend religious services more often are much more likely to be parents than those who never or seldom attend. Only about 40% of never attenders are parents, compared to about 65% of those who attend more than once per week. Religious people are more likely to be parents - that’s unmistakable in this data.

But the other trend that emerges is that Democrats are less likely to be parents compared to Republicans. Thirty-six percent of never attending Democrats are parents, compared to 45% of Republicans. That gap sometimes disappears, but it never reverses. There is no attendance level where Democrats are more likely to be parents compared to Republicans.

The differences in parenting rates are just mind blowing to contemplate, though. Among people in their peak parenting age, a Republican who attends religious services weekly is nearly twice as likely to be a parent compared to a never attending Democrat (69% vs 36%). Religious participation leads to a higher likelihood of having children and this tends to be a bit more prominent among Republicans than Democrats.

However, this is just one piece of the puzzle. How about the total number of children in their households? For those in sample who were not parents, I coded that as having zero children and compared several of the larger religious traditions. Again, this broken down by partisanship.

The same general trend is here: Republicans have more children than Democrats. This is true when comparing Protestants, Catholics and yes even atheists/agnostics. There aren’t a ton of Republican atheists/agnostics out there (about 400 in this sample), but they have nearly twice as many children as their Democrat counterparts.

The other clear trend is that religious people have more children than non-religious people. The average Protestant in the same has about 1.1 children, Catholics are just a bit higher at 1.2 children. For atheist/agnostics, the overall average is about .7 children per adult. A Protestant Republican has over twice as many children as a Democrat atheist/agnostic. Obviously, it’s difficult to calculate the overall fertility rate of atheist/agnostics using this data, but replacement level fertility is 2.1 children per women - this is nowhere close to that level.

I had to test this one more way. I had a nagging suspicion that this is just a weird combination of partisanship, income, and education. Rich, educated, left-leaning folks are just less likely to have children. Atheists check a lot of those boxes. So, to test that hypothesis I put together a regression. It’s an interaction between religious tradition, income, and political partisanship with some basic controls for race and education, too. Again, this is just folks who would be the most likely to be parents to minor children - those between the ages of 30 and 45.

The upshot here is crystal clear - religious tradition (or lack thereof) has a huge bearing on the decision to become a parent. For instance, among Democrats who make $75,000 per year, here’s the share who are parents by religious tradition:

• Evangelicals and Catholics: ~65%

• Non-evangelical Protestants: ~55%

• Atheists/agnostics: ~37%

Among those making $150K to $200K:

• Evangelicals and Catholics: ~80%

• Non-evangelical Protestants: ~65%

• Atheists/agnostics: ~48%

For Catholics and evangelical Democrats, parenting rates rise above 50% at household income levels around $25,000 year. For atheist/agnostic, a majority are parents when incomes move closer to $250,000 per year.

And this same general pattern is there for Republicans, too. It’s key to note that Catholics and Protestants (evangelical and non-evangelical) have parenting rates that are noticeably higher than atheist/agnostics at income levels of $50,000 per year and above. A Republican atheist/agnostic who makes between $150K-$200K per year is about 15 percentage points less likely to be a parent than a Christian at the same income level.

There’s no other way to put this: secular folks (atheists/agnostics) are less likely to choose to be parents and they have (on average) smaller families than religious people. This is not a function of income, education or partisanship. Republican atheist/agnostics are less likely to have children than Democrat Protestants and Catholics at basically every income level, too.

I don’t want to make too much of this but look at the change in the atheist/agnostic share of the adult population in recent years. In 2013, it was 10%. In 2016, it was 12%. In November of 2022, it was still 12%. One way for any group to grow is through offspring. Atheist and agnostics are fighting an uphill battle in this area. It’s harder for a “religion” to grow if it has to achieve that growth largely through conversion. If there’s a reason to believe that the rise of the nones will plateau - just look at the number of atheists and agnostics who are parents.

The code to reproduce this analysis is available here.