Let's Read A Study: The Impact of Denying a Wanted Abortion on Women and Children

Quite large!

The Impact of Denying a Wanted Abortion on Women and Children is a quasi-experimental study of what happened when women in Colombia were denied abortions.

The design here is a little complicated, so I want to walk you guys through it. In 2006, Colombia legalized abortion:

• If a pregnancy threatens a pregnant person’s life, physical health, mental health, or social health (i.e. ability to form healthy relationships).

• If the fetus has disabilities incompatible with life.

• If the pregnancy is a result of rape, incest, or unwanted insemination, and either the pregnant person reported it to the authorities or the pregnant person is under 14 or the pregnancy is related to the armed conflict in Colombia.

In February 2022, Colombia fully legalized abortions before 24 weeks, while abortions after 24 weeks continued to follow the previous policy. The dataset in this study is from 2006 to December 2022, so this change in rules has a fairly negligible effect on results.

Many medical professionals (charitably) didn’t understand the law or (uncharitably) deliberately denied pregnant people abortions that they were entitled to under the law. The ways that medical professionals denied pregnant people abortions were diverse: refusing to provide an abortion for conscience reasons while illegally refusing to refer the patient to a provider that performs abortions; denying abortions to people whose pregnancies harmed their social or mental health or had more minor harms to their physical health, because those reasons weren’t ‘good enough’; setting themselves up as the police whose job was to figure out whether a rape really happened; denying late-in-pregnancy abortions; denying minors abortions without parental consent; or requesting unnecessary documentation.

Colombia has a really cool legal instrument called a tutela. A tutela is a piece of paperwork that anyone can file if they believe their fundamental Constitutional rights have been violated by a public authority or someone “performing public functions” (such as a hospital). Tutelas are free, can be filed without a lawyer, must be resolved within ten days, and are extremely simple. They can even be filed verbally, without any need to write anything down. 84% of Colombians are familiar with tutelas, and hundreds of thousands are filed each year. 30% of women in Medellín (the second-largest city in Colombia) have filed a tutela, and poor women are actually more likely to file tutelas than rich women. And, because pregnant people have a legal right to an abortion under certain circumstances, pregnant people denied an abortion can file a tutela.

But you can’t just compare pregnant people who had their tutelas accepted by the judge to pregnant people who had their tutelas rejected. After all, many pregnant people had their tutelas rejected because the abortions they wanted were illegal. Those groups are going to be different!

Fortunately for economists, tutelas are assigned to judges at random, which means that the authors could do something quite like a randomized controlled trial. The authors made two comparisons. First, male judges denied 62% of abortion-related tutelas, while female judges denied only 42% of abortion-related tutelas. Therefore, the authors could compare pregnant people whose abortion tutelas were assigned to female judges to pregnant people whose abortion tutelas were assigned to male judges, in order to estimate the effect of having your abortion tutela denied when a female judge would have accepted it. Second, some judges are just more abortion-tutela-denial-y sort of people. The authors compared the outcomes of pregnant people assigned to more abortion-denial-y judges with the outcomes of pregnant people assigned to less abortion-denial-y judges. Both comparisons produced basically the same results.

I want to highlight that this isn’t a study of the effects of fully legalizing abortion. Pregnant people with very legal abortions—such as pregnant people whose lives were in danger—presumably would have had their abortions approved by either set of judges. Conversely, people whose abortions were illegal under Colombia’s abortion laws—such as pregnant people who could raise a child but just don’t want to—would have their abortions denied either way. This is a study of marginal abortions: abortions that some people would think are legal and some people wouldn’t.

The authors looked at a sample of almost 20,000 pregnant people from Medellín. This is a large, robust study that is a pretty representative sample of pregnant people who were denied abortions in Medellín. What did it find?

Denying pregnant people abortions kills them. Pregnant people assigned to a male judge are 1.6 times more likely to die in the next nine months than pregnant people who had an abortion. This isn’t maternal mortality: the increased deaths are among pregnant people who don’t give birth. Instead, the deaths are caused by septicemia and infections—likely consequences of an unsafe illegal abortion.

Nevertheless, pregnant people assigned to a male judge are about twice as likely to give birth in the next nine months as pregnant people allowed to have an abortion. And what happens after that?

Colombia keeps extensive administrative data, which allows the authors to do incredibly long-term followup—the average person in the sample was followed up six years later. Pregnant people assigned to a male judge:¹

• Were twice as likely to have children at all.

• Had, on average, 0.5 more children.

• Were 20% more likely to live with parents or in-laws and 24% more likely to live with any adult relative, suggesting that many of them were stuck living with their extended families in order to survive.

• Were 42% more likely to be single mothers.

• Were 33% more likely to be divorced or separated.

• Were 62% more likely to have a health issue in the previous thirty days (remember, this is an average of six years after the denied abortion).

• Were 43% less likely to have a high school diploma.

• Were 55% less likely to have a job.

  • The employment impact is twice the employment impact typically associated with a woman having a child.

• Were half as likely to be looking for a job.

• Were 22% more likely to be a homemaker.

• Were more than twice as likely to have no activity.

• Were 42% more likely to live in one of the poorest neighborhoods.

• Were 47% more likely to live in moderate or extreme poverty.

• Were 72% more likely to live in multidimensional poverty (which incorporates health, education, and standard of living).

• Were five times more likely to be on welfare.

The effect is large immediately after the denial of an abortion, and persists over time:

In short: denying pregnant people abortions makes them substantially poorer, sicker, and less well-educated, and these results persist for fifteen years or longer after the abortion denial. It causes divorces and separations (some family values). Abortion denial means that pregnant people will have children they otherwise wouldn’t have—a huge limitation of their ability to determine the course of their own lives. Again, these results apply only to the people most in need of an abortion, and likely overestimate the positive effects of abortion for the average person who wants an abortion. But I think the lesson of this study is that—for many pregnant people—abortion rights are absolutely crucial, among the most important kinds of healthcare.

I’m sorry for my non-American readers, who are going to have to put up with me hammering this point home repeatedly until the election. But in the United States, there is exactly one party that wants to force pregnant people to carry their pregnancies to term—even if their health is in danger, even if it makes them suicidally depressed, even if they were impregnated by their rapist, even if they’re children. And criminalizing abortion has real, concrete effects on people’s lives: in broken families, in grinding poverty, in lifelong disability, in children raised by mothers who never wanted them. Please bear this in mind when you vote.

1. The exact numbers are comparing male and female judges, but both comparisons have effects of similar size.