Portugal’s abortion ban

A number of Portuguese reporters have contacted me asking my opinion about the nationwide vote to consider whether women in that country should be allowed to have abortions.

After some thought, I decided not to reply to any of these press inquiries. The reason is that my research in abortion and crime, as discussed in Freakonomics, is virtually irrelevant to any discussion of whether abortion should be legal. It is true that I believe that legalized abortion, by reducing the number of unwanted children, lowers future crime. On the other hand, if you put any value on either a fetus or a woman’s right to choose, any benefits from lower crime get swamped by these other considerations. I could have tried to convey this point to the reporters, but given past interactions with journalists on this topic, that is not the answer they want to hear, and thus, not what makes it into the stories.

Ultimately, legalized abortion is a moral and political issue, not an economic one (even under my incredibly loose definition of what economics encompasses). My opinion simply doesn’t matter.


Good for you. I've always held that (although I don't agree with most abortions) if the issue ever comes to a popular vote, as a male I would conscientiously object to exercise my right to participate. Because the share of the burden of pregnancy is an overwhelming inequality favoring the female, males really have zero say in the matter; in my opinion.



"swamped" means roughly the same as "is vastly outweighed by", and it's standard lingo in the field, devoid of negative or perjorative overtones.

It suggests a numerical comparison, and the idea is that the value of reducing crime is trivial--a small number--compared to the other values at stake, such as a woman's (alleged) right to choose or the fetus's (alleged) right to life--both very large numbers.

An analogy: one reason to keep abortion law the way it is now is that it would save paper. To change the law would require a bit more paper consumption for the Supreme Court, Congress, the media and so on. But surely the value of saving paper is swamped by other considerations, considerations such as a woman's (alleged) right to choose or the fetus's (alleged) right to life.

I'm not sure your saturation metaphor conveys the same idea as that.


"In regression analysis, these other factors (such as impact on crime rate) would factor out due to not affecting the overall outcome trend."
Could you explain how you derived the info needed to state such claims from this conversation? (ie: w/rt the info presented here). Respectfully...



Since I'm Portuguese and voted in the public consultation. I have to say that altough your proven theory about the correlation between abortion laws and crime in the states, is IMO, correct. Here in Portugal that theory is not an issue, due namely to the fact that the crime is less violent than in the US, and also due to the fact that abortion, altough punish by law, is practiced here without the punishment predicted by the law.
Abortion is not a social crime, when I say social is a crime that is not appointed by the others. that are several people that knew someone that did it, but they will never go to the police and pointed them.
So where does the economy enters in this question?
All the "underground" abortions escape the Tax system making some people very rich by doing them, also, if complications arrise in the procedure, the public wealth system as to correct the situation spending more money than expected.

Needless to say abortion, IMO, will always be a question of conscience at least in Portugal..



What if the question from the press was: "Does the vote matter?" "Or what increase in the crime rate might you expect to see?"

I realize these questions might not be perfect. For instance states (or countries in this case) that consider such legislation self-select to some degree (i.e. might already have low crime rates [or not]). Alternatively, your findings with respect to the degree of crime reduction (or increase) might not be generalizable. Point being, your research has policy implications, though the press was asking you a much more subjective question. (That is also not to imply that policy decisions aren't made using sometimes subjective advice/data).


Good call. I am glad to see you are not becoming one of those pundits who will wax philosophic (or economic, if you prefer) on any topic you may have discussed in any capacity.
Abortion is a complex societal issue wrapped up in morals, religion, concepts of Government's role in our lives, etc. What your analysis uncovered is very interesting, but it seems unlikely that any Country would decide this issue based on that alone. Further, you do not even know if you analysis applies to Portugal (banning Soccer may reduce crime more, for all we know ;-)


Kudos to you, Levitt, for recognizing that there's more to morality and legality than economics. Not all economists write as if they recognize this point.


I understand your reluctance to explain your position to reporters, but wouldn't your work actually be anti-abortion from an economics viewpoint? In your book, you state yourself that abortion as a method to reduce crime is a highly inefficient mechanism, from an economist's viewpoint.

I guess this fact is missed by reporters and they don't realize that you are arguing in a different dimension than they are.


It is really nice to see (read) you handle this matter son professionally, congrats!


"Economist: Legalized Abortions Pre-Emptive Capital Punishment"

Come on, it would've made a nice headline.


@ Steven Levitt:

I think you did the right thing by not replying since they would (portuguese journalists or otherwise) spin your response to their own will since here in Portugal this referendum was quite a hot topic with people very much upset on both sides of the issue and they would confuse facts with your opinion.


Hey, at least the book is a bestseller. Was it ever in a headline? (I find your claim preposterous, by the way.)



Hello, I'm Eduardo from Lisbon, Portugal. I just want to say that the article you posted to is tottaly FALSE!! Abortion will be legalized in Portugal until 10 weeks of pregnancy and in autorized health institutions. Although voter turnout was under 50%, prime-minister José Sócrates declared that legalization will be a fact in a short term period.

Lisbon, Portugal


Wow! Portugal will authorize abortion! How times have changed. Only took 30 to 40 years after other western European countries.

I guess I'll have to update my 1972 backpacker view of Portugal. I went to Spain last year. Boy, has that place ever changed. I saw women in miniskirts inside of the cathedrals.



"On the other hand, if you put any value on either a fetus or a woman's right to choose, any benefits from lower crime get swamped by these other considerations."

What does "swamped" mean here? If my son or daughter does not end up getting raped, mugged, killed, robbed, the harm prevented looks huge to me compared to the other considerations. Swamped is vague and arguably negative - perhaps unduly negative. I supposed I could agree with something more neutral, like irreversibly saturated. Once the saturation occurs, and its is irreversible, then the economist can properly say this looks like a philosophical/ethical/moral issue more than a strictly social science issue.


majikthise, "saturation" is a term any engineer can get behind, especially in favor of an overloaded term like "swamped". Saturation means that one factor overwhelms the others, so that the others are just background noise. I think that applies well to the meaning that was intended in the blog entry.
Swamped has more of the pejorative common language meaning. But, I think it is clear that both mean that one is such a high percentage (emotional or otherwise) as to make the other not a factor. In regression analysis, these other factors (such as impact on crime rate) would factor out due to not affecting the overall outcome trend.


Snubgodtoh, you are right that what I said gave a false impression. I meant that impact on crime rate (as one factor) may not affect outcome of a vote, because it would be overwhelmed by other factors (such as emotional/religious beliefs, attitude about women's rights, attitude about government intrusion, etc). I did not mean it would not affect the outcome in terms of crime rate, rather the outcome in terms of a decision process regarding legality. So, if doing sampling/polling looking at voting trends (whether public or congress), regression analysis would not find that Levitt's analysis and conclusions would be a factor in such outcomes.
I hope that clarifies my intent for that statement.

John Hart Ely

Excuse me if I do not join the other sheep in congratulating Levitt for avoiding a difficult but relevant question.

First of all, just what is the distinction between a "moral and political issue" and an "economic one"? It seems to me that normative economics is about maximizing those measurable correlates (efficiency/wealth/whatever) to utility, which is what normative ethics is concerned with maximizing.

In any case, Levitt seems to ignore this point soon after he makes it, hanging his hat not on the incommensurability of the two factors, but rather on the idea that any effect on crime is "swamped" by the other factors. Now, this is certainly possible, but it is a controversial position! Where is its defense? Certainly there is a quantity of crime reduction that would swamp the other considerations, no?--eg, if legalizing abortion ended all crime in the world we would think that a pretty strong MORAL reason in favor of doing so.

Furthermore, if Levitt really believes that his abortion/crime research is irrelevant for making policy decisions, WHY DO IT IN THE FIRST PLACE? I guess you could say that counter-intuitive/controversial results sell books or whatever, but from the perspective of a scholar, why spend time trying to tease out a casual relationship when you admit that such a relationship, even if it existed, would be irrelevant?

My view: Abortion legalization's effect on crime rate is OBVIOUSLY relevant (as are all its effects on social welfare). The other effects of legalization may or may not “swamp” its effect on crime rate, but this is a fantastically difficult empirical question--we should not arbitrarily assume one answer is correct by decreeing that the various effects are in different and incommensurable categories.



Mr. Ely, I think you may have missed the point. Production of information does not mean immediate course of action. For example, if you determined that killing every 3rd person in the US would reduce crime to 10% of current, that is a great statistical outcome, but may not be justified in light of other information.
Therefore, analysis and information is fed into decision and deliberation mechanisms. Whether that information sways the outcome depends on many factors. This is clearly no different with hard-science analysis (such as CO/CO2 effects on global warming, health risk/mitigation assessments, etc). We always have to compare the simpler input->output analysis with broader social issues as well as economic costs (e.g. stopping all CO/CO2 production right now would reverse the Global warming trend, but the economic costs would be unacceptable to society).
In the present case, there are a lot of studies on the factors influencing views on abortion; none currently include down-stream crime statistics that I can find. So, it is a reasonable 1st order assumption that other factors are swamping/saturating crime-factors in the "abortion debate" with regard to the general public (rather than say among researchers).


John Hart Ely


YOU have missed MY point. I am simply arguing that Levitt must JUSTIFY saying that any reduction in crime is swamped by the other results of legalizing abortion. He must say something LIKE, “legalizing abortion will reduce crime by 5%, the welfare gains of this are approx. X, it will also give approx. Y welfare gain to women, the cost of losing these fetuses is approx. Z, so given that X is so small relative to Y and Z, reduction in crime is swamped by other effects.”

He can't just say “legalizing abortion would reduce crime by some percentage and from my crystal ball I know that this percentage generates a welfare gain far below that which would be relevant to the discussion.”

Yes, clearly the general public doesn't think that crime is a relevant factor in the abortion debate. The general public is also basically irrational.