Did the contraceptive pill increase unwanted pregnancies?
No, but there's a surprising amount of nuance here.
I’ve seen a couple of versions of this claim now. The first I came across was in Mary Harrington’s book Feminism Against Progress, where she says: “the existence of contraception so radically changed social norms that many more casual sexual encounters took place. And contraception was only mostly effective - so the absolute numbers of accidental pregnancies went up”. I recently spotted a variant on Twitter, where Sarah Haider said “occasionally [social policies] have a counterintuitive effect, like the pill leading to more single moms, not less”.
Both versions of the claim seem incredible at first glance, so I wanted to check whether either of them could be true. This turns out to be a rather complicated and interesting question!
Although the two versions look related, it is possible that one could be true and the other false, since the rate of single motherhood is a function of multiple factors, of which the unintended pregnancy rate is only one. I’ll discuss each in turn.
The pill caused an increase in unwanted pregnancy: false
Harrington’s assertion is that uptake of the pill increased sexual activity to the extent of driving up unwanted pregnancy. One natural reason to doubt this is that global fertility rates have been on the decline since the 60s (when the pill first became available). It would be a bit odd if rates of unintended pregnancies had nevertheless risen during the same period (though not impossible - maybe the rate of abortion simply rose faster, for example).
And although the post-60s drop in fertility coincides with the pill’s introduction, correlation isn’t causation. It’s interesting to look further back and see that in the US, there’s actually been a steady decline in fertility for much longer; the 1930 - 1970 period was actually kind of anomalous. Clearly fertility has many determinants!
Ok, so overall fertility has declined since the 60s, but what about unintended pregnancy? Directly measuring unintended pregnancy is actually quite hard. Even the definition is a bit unclear: interestingly, much lower contraceptive failure rates (less than 1% for the pill!) have been documented when women are on teratogenic medications, suggesting that some apparent contraceptive failures are really a function of ambivalence about pregnancy. Surveys that ask people whether births were wanted are vulnerable to response and recall bias, plus there just aren’t very many of these. In settings where abortion is readily available, abortion rates are a reasonable proxy though. The data shows that in some countries abortion rates have risen alongside contraceptive prevalence, while in others they have fallen.
Consider your risk of contraceptive failure in a year. This is a measure called the Pearl Index (PI), and is how birth control efficacy is usually reported. It varies between clinical trials and out in the real world, so I’ll stick to real world or “typical use” PI values. The PI of condoms is 0.13 and the PI of the withdrawal method is 0.2. For the pill it is 0.07. That means that if you switch from condoms to the pill, you could have sex for almost twice as many years without exposing yourself to a higher risk of pregnancy. At the population level, if women spent roughly the same number of years trying to avoid pregnancy, but they all switched to a contraceptive that was twice as effective, rates of accidental pregnancy would be halved. If, however, the number of years spent trying to avoid pregnancy doubled, that would totally offset the improvement in contraceptive efficacy.
People in most countries want fewer children today than they did 50 years ago. This means they spend more time avoiding pregnancy. A model that incorporates this can explain the differences between countries. When desired family size remains constant, abortion rates decline as prevalence and efficacy of contraception increase, but if desired family size decreases, uptake of both contraception and abortion can rise simultaneously, depending on the gradients of each trend. This is very different from saying that the use of contraception is itself the cause of the increase in abortion!
The state by state rollout of the pill in the US provides a convenient natural experiment to look at its effect there. A plethora of studies exploit the differences in policy timing between states to show that legalising the pill caused fertility rates to decline faster, marriage to be delayed, and women’s education and wages to increase. When women become more educated and economically empowered, the opportunity cost of having a child gets higher and ideal family size gets smaller. Given the model described above, it is conceivable that such a decrease in desired family size could have caused an increase in unwanted pregnancies (but only in pregnancies that are unwanted relative to this new standard).
In the US, we’re talking about a relatively modest change in desired family size - between 1930 and 2013, Americans’ ideal number of children decreased from 3.5 to 2.5. That adds maybe 3 extra years spent avoiding pregnancy per woman. As already noted, the pill is about twice as effective as condoms, and 3 times as effective as withdrawal. Seems unlikely that with such high efficacy, these few extra years of risk would be enough to increase unwanted pregnancy overall.
Here’s a very rough model. Holding the length of a woman’s reproductive life fixed, suppose the average annual risk of contraceptive failure drops just a few percentage points, while the number of desired children reduces from 3.5 to 2.5. Accidental pregnancies go down in almost all cases. It’s only if you see a much more dramatic shift in desired family size, or if the contraceptive failure rates remain pretty high, that the rate of accidental pregnancy can go up (and these women will still be having fewer pregnancies overall).
For a much more precise model of the relationship between contraception, desired family size, and unwanted pregnancy, check out Bongaarts and Westoff, 2000, The potential role of contraception in reducing abortion.
I’m assuming here that the introduction of the pill meant a decrease in the average annual risk of unwanted pregnancy, but Harrington is perhaps implying that such large numbers of women went from being abstinent to using the pill that the average annual risk of unwanted pregnancy actually rose. I’ll just note how implausible this is: in 1960, before the introduction of the pill, 70% of all American adults were married. The median age at first marriage was just 20. One of the most reliably documented social effects of the pill is an increase in the age at first marriage, and even though fewer single people are abstinent today (10% of US women wait for marriage vs 50% in 1960), single people still have less sex than coupled people. As marriage rates have declined, overall rates of sexual activity have declined also.
In the US natural experiment, states that introduced the pill sooner saw lower rates of unwanted and mis-timed pregnancy. In fact, perhaps as much as 40% of the decline in marital fertility between 1955 and 1965 was due to the pill averting unwanted pregnancy (married women had earlier access than single women).
It seems like a win for common sense - the introduction of the pill did not cause an increase in the rate of unintended pregnancy.
The pill caused an increase in single motherhood: maybe
I’m going to keep focusing on the US, since most of the best data is here. In this case the trend does at least go in the right direction: the percentage of families headed by single parents has grown since the 60s.
There are two primary ways people become single parents. The first is that a single woman becomes pregnant. Even if unintended pregnancy rates have gone down, if more women are single and sexually active for longer, a higher proportion of contraceptive failures will occur outside of relationships. And since it is the case that age at first marriage has increased and overall marriage rates have declined, women are single and sexually active for longer these days. As mentioned above, the pill does seem to have had a direct causal effect here. The proposed mechanism is that when women have control over their fertility, it becomes possible for them to invest in longer programs of higher education and careers, deferring or even forgoing marriage.
One thing that counts against this narrative is that the majority of the rise in single motherhood within the US is accounted for by trends among women who spend less time in higher education. For example, a 2004 study finds that education is strongly anticorrelated with single motherhood, and that “less than 5% of all unmarried births are to college-educated women”.
When you disaggregate the data on single motherhood by educational level, it’s really striking: the increase after 1970 is driven exclusively by those in the bottom 3 quartiles. These are the women who are least likely to be using hormonal contraception.
Percentage of mothers who are single
It’s also unclear what proportion of births to single mothers are actually the result of contraceptive failure; another way single motherhood would increase is if more single women actually want to get pregnant these days (e.g. because welfare provisions for single parents have improved and social sanctions lessened). Promises I Can Keep: Why Poor Women Put Motherhood Before Marriage is an absolutely fascinating read. Very young poor women, teenagers many of them, are having babies because they want to. With few educational or career aspirations, the romance of a baby is very alluring. These girls have knowledge of and access to contraception, but they choose not to use it, or to use it irregularly. “Children offer a tangible source of meaning, while other avenues for gaining social esteem and personal satisfaction appear vague and tenuous”. Meanwhile, many of the poor young men fathering children are not good marriage prospects: criminality, incarceration, intimate partner violence, infidelity, and addiction are common. There just aren’t any powerful social or financial reasons for these couples to stay together.
Marriage rates are actually the lowest among lower socioeconomic classes. These groups are also the least likely to be using contraception, so it doesn’t look like the pill is having a direct effect here. But it is possible that it has had a more diffuse effect, by changing social norms around sex and marriage for everyone. It’s absolutely true that marriage rates have declined for all groups, and that as a result, unmarried mothers are more common, even though everyone is having fewer babies.
The other primary way that people become single parents is through relationship breakdown. In 1969, the first no fault divorce law was passed in the US. As state divorce laws liberalised, rates spiked, and remain higher among less educated women, consistent with the differences in rates of single motherhood. In America today, married people with Bachelor’s degrees have about a 30% chance of getting divorced, but high-school graduates have a 50% chance, and those who didn’t graduate high school have a 60% chance. So divorce seems like part of the picture as well.
Percentage of marriages ending in divorce within 10 years
It is hard to get good data on what proportion of single mothers are single due to relationship breakdown, as opposed to falling pregnant outside of a relationship. One suggestive dataset is from the General Social Survey (GSS), which asks 16 year olds not living with both parents why they are not living with both parents. This isn’t perfect as it includes children living with two parents where one is a step-parent, and it doesn’t explicitly ask about being born to a single mother. But it looks as though both parental separation and “other reasons” for not living with two biological parents are more common today than they were in 1970, with the latter increasing at a slightly higher rate.
There is no simple story that can explain the rise in single parenthood. Economists and sociologists have suggested a whole host of factors, including welfare policies, female wages and employment rates, male wages and employment rates, sex ratios, divorce rates, marriage trends, religiosity, and a range of other social norms. Is modern contraception one of these factors? Insofar as it has shifted cultural norms around marriage, sex, and gender roles, it seems likely. But pinning down the exact mechanism and magnitude of its effect is hard.
The self interest calculus (“Good for me is actually good in itself
!”) of trad cons has been amply outlined in work by @Rob Kurzban and Jason Weeden.
https://pleeps.org/2015/01/12/sexual-politics-and-self-interest/