The Economist has an interesting article discussing regulatory changes regarding child safety seats:
During the Reagan era, only the truly wee—tots aged under three—had normally to be secured in child-safety seats. But states’ governments have, since then, gradually ramped up the requirements. Today, most places in America make children sit in safety seats until their eighth birthdays. That concern for youngsters’ safety has had the unintended consequence, Dr Nickerson and Dr Solomon suggest, of fewer three-child families. . . .
They discovered that tightening those laws had no detectable effects on the rates of births of first and second children, but was accompanied by a drop, on average, of 0.73 percentage points in the number of women giving birth to a third while the first two were young enough to need safety seats. . . .
They estimate that laws requiring children to sit in special seats until they are eight years old saved about 57 lives in 2017 and contrast that number with the 8,000 children who might have been conceived and born in the absence of such rules. There is, they conclude, no “compelling social interest” in requiring child seats for children over four.
This seems weird. Comparing putative lives forgone to actual lives saved is, to put it politely, a strange moral calculation.
The writer for The Economist seems to think the government’s moral calculation is more sensible than the one employed by the authors of this paper, but doesn’t explain why. Do they entirely reject cost/benefit analysis (an approach The Economist generally favors), or do they reject the specific moral calculation in this case? And if so, what’s the numerical equivalence that they think is more accurate? Does one life lost equate to one foregone child? To 10 foregone children? To 100 foregone children? How about a million foregone children? A billion?
It seems to me that the “strange moral calculation” is being made by the government regulator that believes they are better positioned than the child’s parents to solve these difficult moral problems.
I’m completely agnostic on this issue; I don’t have a clue as to how one should weigh 57 lives lost against 8000 children never born. If you are confident that you can do that sort of moral calculation, then please tell me your conclusion. What is the correct equivalence ratio to employ in this sort of cost/benefit analysis?
Because I don’t know the answer, I’d prefer to leave those decisions up to parents, not government regulators.
PS. This has nothing to do with my view on the value of a higher birth rate. I use the same agnostic moral reasoning on issues like abortion, where deregulation cuts in the other direction.
PPS. The decision to have children is implicitly a decision to risk the life of a living mother in order to gain a potential human being.
READER COMMENTS
Euglossine
Jun 26 2021 at 2:33pm
Excellent post. I’ve wondered about this kind of calculus myself in the past, this seems, after having my eyes opened, to be the obvious solution… especially when you consider the millions of people who are required to do something unnecessary without any obvious societal cost other than the money they have to spend.
Flipping the question around though, is there a number at which you accept the idea that we should regulate? For example, we require mirrors on a car, which probably doesn’t cost any lives even though there is an increase in the cost of the car. Or what of something that saves a million lives without costing any potential ones, at least measurably.
Dylan
Jun 26 2021 at 2:38pm
I’m not a fan of this regulation. This is the first I’ve heard that children have to be in child seats until 8(!). However, I think the lives saved vs. foregone children is just not the kind of thing we can compare, and so agree with the Economist that it is a “strange moral calculation,” the kind of thing that inevitably ends up equating masturbation to murder.
Scott Sumner
Jun 26 2021 at 4:52pm
Do you believe that China’s one child policy had important social costs? And were those costs related to the fact that some families were denied the right to have children that they wished to have? I can’t see how we can ignore the cost of regulations that prevent children from being born, even unintentionally.
I do understand how people might have different views on this issue, but isn’t that exactly why freedom is the best public policy in these sorts of cases? Who is to say how potential lives should be valued?
Dylan
Jun 27 2021 at 8:36am
Yes, China’s policy had social costs, but as you allude to, these are more in the realm of restricting freedom and I think measuring that cost in the form of # of hypothetical children that would have been born but were not is nonsensical. As for the car seats, measuring it in that way seems even worse. People still have the freedom to have as many children as they want, there is just an additional constraint that adds a marginal cost to the third child and apparently that is enough to cause a fair number of parents to stop at two.* So, as some suggest below, if you want to get utilitarian about it, the proper way wouldn’t be to try to measure the cost of avoided children deaths against unborn children, but instead look at the marginal cost that the child seat law puts on parents and conclude that the value of an additional child is necessarily less than that cost (for those parents that forego the extra child)
For this law, I think it makes a lot more sense to argue that parents already have a strong incentive to do everything they can to make their children safe (indeed, I think you could argue the incentives are for parents to over-invest in child safety). If there is strong evidence that children should be in booster seats until they are 8, we should expect the majority of parents to do that without the need for a law.
Scott Sumner
Jun 27 2021 at 11:59am
“I think measuring that cost in the form of # of hypothetical children that would have been born but were not is nonsensical.”
Why?
Dylan
Jun 28 2021 at 6:07am
Take it to the logical extreme. There are billions of children that could be conceived and born each year that are not. Do those unborn children get weight in the utilitarian calculus? Generally not, because you get to absurd conclusions like no contraception or even abstinence allowed. Or else you get to a place where you are trying to weigh the lives of unborn children against other goods and come up with some rubric like saving the life of one child is worth 100 unborn children, but not 200. Which is the kind of comparisons that utilitarians are rightly mocked for.
Notice, you’ve not actually made an argument here for the moral calculus to compare foregone lives to those of existing lives. The underlying argument is that parents should have the freedom to choose, this is a stronger argument and doesn’t require you to weigh 10 unborn against 1 born.
Ghatanathoah
Jun 28 2021 at 10:07pm
@Dylan
It does sound absurd to say something like the life of one unborn child is worth the life of 100, but not 200. But at some point a line needs to be drawn. And utilitarian philosophers are not the only ones who need to draw it. Regulators and economists need to draw it as well in scenarios like this.
Even parents need to draw it, no parent ever has to sacrifice 1 child to have 100, but they might find themselves in a scenario where they have to take a 1 in x chance of losing one child to have another. For example, imagine two parents are driving to a fertility clinic to help them have their next child, and can’t get a babysitter. Should they drive there with their existing kids and leave them in the waiting room, even though there’s a 1 in x chance that the kids will die in a car accident on the way there?
Most people are willing to tolerate some risk of killing someone in order to have a child. Besides the risk to the mother, there are other risks, like the aforementioned risk of getting into a car accident on the way to a fertility clinic. On the other hand, there is also a level of risk no one would tolerate. If a parent kills one of their children they could not defend their actions by promising to make another one. If there was some occult ritual that had a 50% chance of killing an existing child and a 100% chance of allowing conception of a new one, no one would allow it.
So we have two extremes. Killing a child and replacing them with a single new child is definitely wrong. And I don’t think it’s just wrong from a deontological perspective, I think it is bad from a consequentialist perspective too. On the other hand, tiny risks to have a child are acceptable. Where is the line drawn? Wherever you draw it will sound silly. Is 200 acceptable, but not 100? Should there be error bars? But that’s okay, it actually makes sense that drawing a line would be a hard decision that would cause anguish. The place you draw a line is the place where you are deeply ambivalent, the place where it does feel absurd to say that one side is bad, but the other is good. I have no idea where to draw that line, maybe it is different depending on the specific circumstances of a family, or of society. But it will definitely feel and sound weird to do it.
Of course, precisely because it is such a hard and difficult choice, Scott is right that pluralism is the correct answer from a utilitarian perspective. Even if we could get everyone to agree on a number, regulating everyone’s behavior to meet it would have all sorts of negative second order consequences. And that’s not even looking at other second order consequences, like the children that the foregone children might have when they grow up, the positive and negative externalities children generate, etc.
Frank
Jun 26 2021 at 4:36pm
Utilitarianism is unclear about the sum of whose utilities should be maximized. For example, do we include baby seals?
For the utility of the children of those who are definitely included in the utility maximization , there may be an appealing solution:
The regulation saves eight lives, which makes eight sets of parents happy. Without regulation, there would be 8000 happy sets of parents. Thus the regulation is bad.
Scott Sumner
Jun 26 2021 at 4:54pm
Or maybe the way to think about this is that if we don’t know the answer, then err on the side of letting people choose how to live their lives?
Euglossine
Jun 26 2021 at 5:48pm
You overlook the fact that those 8, 000 sets of parents were only marginally happier by having those children, in their own estimation. I say this because they were deterred by the need for another car seat. But you also overlook the perhaps millions of people who had to go through the hassle of having another car seat without any benefit whatsoever. I still say, though, that this is a question we cannot simply brush under the rug by saying we should leave it up to individual choice unless we are prepared to say that for all such questions like this. Which, I suppose, would be a libertarian Utopia but is not an actual option in today’s world. So to be persuasive, Scott needs to say how he would draw the line or that he is arguing for no regulations with any costs.
Scott Sumner
Jun 26 2021 at 9:37pm
I don’t think that’s the right way to think about the problem. The burden of proof is on those who favor regulations. They must make the case for restricting people’s freedom. It’s up to them to make a persuasive argument. The Economist didn’t even come close to doing so. They suggesting the authors had made a weak argument, without telling us exactly what was so weak about the argument. What’s wrong with their moral calculation?
And no, I did not ignore the loss of utility to the parents.
There may be other areas were the moral calculation clearly favors regulation. Say with a carbon tax. This is not one of those clear cut cases.
Jim Ancona
Jun 27 2021 at 8:38pm
They weren’t “deterred by the cost of another car seat”. Child safety seats must be placed in the back seat, and you can’t put three of them across the back seat of most cars. So they were deterred by the cost of replacing their car with a minivan or three-row SUV. That’s why the effect kicked in at three children.
BK
Jun 26 2021 at 6:45pm
It is hard to weigh up the extreme trauma of involuntary sudden loss of a child in a car crash (even if you know there is risk to driving a car, it seems fairly evident from studies that people do not calculate/comprehend such low level risks well) against a deliberate decision which may have some wistful sense of loss in a hypothetical sense. Not to mention alleviating the guilt of a second driver in the crash who may feel culpability for the death of a child.
Mostly I’m of the opinion that these are different classes of problem, and any comparison between the two is like trying to subtract a string from an integer, or perhaps trying to buy credibility. But to indulge and take the step the economist doesn’t and ponder a basis for the calculus, boiling things down to dollars. How much would a parent pay to avoid the killing of their child if the threat was explicit? I don’t know what kidnapping pays these days but I’m going to say between 200K and 2M. Split the difference and put it at $1.1M. A car seat costs $300. This means a child not killed is “worth” 3667 third children not born. 57×3667=209000. More than an order of magnitude above what the market has resolved to. Seems like the bargain the government is implicitly making adds up.
BK
Jun 26 2021 at 7:08pm
Actually it occurs to me now that I have woken up more that the incremental cost of a seat isn’t the problem, it’s the incremental cost of a minivan or similar over a normal sedan. Seems the break even point in my rough math would then be $8000 rather than $300.
Scott Sumner
Jun 26 2021 at 9:39pm
Aren’t you ignoring the forgone utility of the child that is not born?
john hare
Jun 27 2021 at 4:52am
I think he was adding in more costs than just the utility of the unborn, bolstering your point as it were. How many millions of hours are parents spending per year to comply with a law of uncertain effectiveness? And how many billions of dollars per year buying the car seats and larger vehicles to accommodate them. And the gasoline for the larger etc.
Frank
Jun 26 2021 at 7:59pm
“How much would a parent pay to avoid the killing of their child if the threat was explicit?”
That’s very close to the cost-benefit [utilitarian] answer, I think. The more precise question is: How much would a parent pay to avoid the killing of their child, knowing the [low] probability of death, and knowing the cost of the abatement technology, the seat. Add over people. A very small number comes out, given the low probability, and/or the cost of the seat.
Compare to the willingness to pay of others to avoid having the seats installed by not having children. They clearly each wish to pay little, but it seems there’s lots of them.
Highly probably, a bad regulation.
Looks like individualism supports utilitarianism in this case! 🙂
OneEyedMan
Jun 26 2021 at 10:12pm
I once heard Amartya Sen speaking about John Rawls’s Vail of ignorance, that we should imagine what rules we would want if we didn’t know what place in society we would get, which Rawls’s combines with the MaxiMin principal, that we should want rules of justice that maximize the welfare of the people with the lowest welfare.
He was talking about the appeal of Rawls’s approach, but remarked that it raises some thorny issues when it comes to rules that influence who is born.
Because, if you count people who are not born as having standing for purposes of MaxiMin and the veil of ignorance, it is pretty clear that they are served very poorly by rules that make them not exist! But, if we agree that those that those that are never born are the worst off, we can instead get mandatory fertility policies (make sure those unfortunate unborn become the born to make them better off) that also don’t jive with our moral intuitions.
He hinted that there are solutions or at least possible solutions to this conundrum, but he sadly never got into them and the lecture moved on to another subject.
nobody.really
Jun 29 2021 at 12:57am
Dunno.
The moral of these entire post is that arguments from “non-existence” often lead to counter-intuitive/absurd results.
But regarding Rawls: Everyone would be presented with a series of visions of a possible future, without their knowing what their position would be in that future–including, presumably, without knowing if they’d even BE in that future. Everyone would then choose (vote?) for the future they found most palatable, and later learn what position they’d have (or not have). So far, so good.
The question then arises: Does the MaxiMin strategy–maximizing the welfare of the least-well-off person–require maximizing procreation? First, I understand that MaxiMin was not a goal; it was merely a strategy that Rawls identified for coping with the Veil of Ignorance. If we could demonstrate that the strategy would inevitably produce absurd results, I expect people would choose a different strategy.
But second, I expect that people would pick the possible world they find most appealing–in full knowledge that if they pick a world with fewer people, they increase their own chance of never being born. That is, we’d pick a world that people would find maximally appealing BEFORE they know what roll they’d have in that world. To my mind, this would fulfill the MaxiMin strategy.
Consider a common tragedy-of-the-commons scenario: a lifeboat. What would be the fairest way to allocate the scarce space on the Titanic’s lifeboats? Let’s say randomly. Here’s the harder question: HOW MANY people should we let on each boat? The more people we admit to a boat, the greater the likelihood the boat will capsize, killing all the passengers. But the fewer people we admit, the more will be left behind on the Titanic to drown. Under these circumstances, I’d pick some number that balanced (1) the goal of increasing the number of lives escaping the Titanic with (2) the goal of reducing the risk of capsizing the lifeboat. I’d seek to maximize the welfare of everyone on the boat–measured from the point BEFORE I pick who would get on the boats. That’s the way to implement the MaxiMin strategy under these circumstances.
Oneeyedman
Jul 6 2021 at 4:33pm
Evidence from labor markets and other estimates of willingness to pay for safety suggest people place enormous value on existence. In fact, by many accounts the average person seems “willing” to pay more than the average person’s income to avoid death (and pay more than a year of income for a year of life). From the perspective of a person existing under the veil, nonexistence is very similar to death and relative to the long run fertility maximalist world, the likelihood of non existence is probably very high. So in my estimate basic utilitarian and Rawlsian approaches do point to fertility maximization unless you can justify more compellingly why the potential people should not count.
Or consider discount rates. Lots of serious people argue social discount rates should be zero because the welfare of people who don’t exist yet should count as much as people who exist today. But that argument applies equally well to the welfare of the people who could exist in the present under alternative policies.
robc
Jun 26 2021 at 11:32pm
I remember hearing, probably in an econtalk episode, that if it was actually tested, a standard seatbelt would meet the legal qualification for over age 3. But no one has paid to have them tested.
If so, the 57 lives may be an overestimate ( or the difference between the typical child seat and the minimum legal seat).
robc
Jun 26 2021 at 11:33pm
Thinking more, I am pretty sure it was Steven Levitt on econtalk.
Billy Kaubashine
Jun 27 2021 at 10:30am
The difficulty of estimating or counting the number of jobs that are NOT created due to higher taxes, increased regulation, or higher mandated wages accounts for decision flaws in many other policy arenas.
When the “good” is obvious, but the harm is not the trade-off is difficult to evaluate.
bill
Jun 27 2021 at 7:42pm
It seems that the 57 saved lives is estimated using the assumption that if the child seats weren’t required for 8 year olds, then zero parents would use the child seats. But if it were voluntary, maybe 66% (or some other number greater than zero) would still use the seats. In that case, only 19 kids die.
Dylan
Jun 28 2021 at 6:21am
I’ve not read the paper, but my guess is they have tried to build the estimate off the fact that different states put in additional booster seat regulations at different times and they are not uniform in place (it looks like only some states require booster seats until a child turns 8). Again guessing, but figure they would have then looked to build a synthetic control to try and tease out what would have happened in places without the regulation, and that includes some parents using booster seats voluntarily for a longer period than the law requires.
In an earlier post in this thread I put in an *, and then forgot to add in the explanation, but what I meant to say up there is that I’m generally skeptical of findings like these. There’s so much going on and I think it is impossible to know if your synthetic control is really a control or if there is some confounding variable that’s not getting accounted for. But, I did want to point out that I’d be greatly surprised if they didn’t at least attempt to account for your objection.
Matthias
Jun 28 2021 at 11:53am
I have no clue what the right moral answer is.
But the right answer for the regulator is clear: err in favour of minimising damage that can be directly attributed to your actions.
It’s somewhat similar to the incentives in favour of keeping airline security theatre going.
nobody.really
Jun 29 2021 at 1:24am
It seems clear to me that children generate externalities–that is, parents choose to make children, but the consequences of that choice may be borne by third parties. This is especially obvious when those kids become criminals or drains on the public purse.
It is less clear whether a LACK of children generates an externality, too. Clearly, many people had anticipate that the US would have more kids than it currently does, and those frustrated assumptions will have consequences for third parties. But arguably those third parties did not have any property rights in those assumptions–so, as a technical matter, the Birth Dearth may not qualify as an externality.
If we regard problems arising from kids as causing more market failure than problems arising from a LACK of kids, then it may make sense to impose regulations designed to reduce the possible problems arising from kids, even at the expense of increasing the LACK of kids. (Oddly phrased, I know.)
Tangentially related: While I defend a person’s discretion to abort a pregnancy, I also think society has an interest in defending any fetus that is NOT being aborted. This leads to some curious conclusions: A woman may terminate her fetus without legal sanctions, but could face sanctions for engaging in conduct that posed an unreasonable threat to the fetus’s development. (A man would be similarly liable, presuming his conduct posed a similar threat to the fetus.) The point of this policy is not to defend the fetus per se, but to defend society’s interest in minimizing the risk that the fetus would be born with expensive medical conditions or other impairments.
may terminate a pregnancy, but should not harm
Yaakov
Jun 29 2021 at 4:57am
I would like to join the motion that the unborn are not the only downside of the regulation. I would list three groups of costs:
1) people spending money on the required safety seat would not spend that money on other safety items such as fire-extinguishers and newer and safer cars. We do not know, but it may be that more people will die due to the lack of these items than are saved by the government safety requirements
2) People fleeing police or otherwise doing foolish things due to disobeying the law. I do not know how many casualties are caused by these people, but there are real tragedies caused by these people every year.
3) the cost of travel not carried out because people do not have the required safety measure. Grandparents, for example, may nt be asked to drive children when necessary, and instead overworked exhausted parents may take the children for one drive too many.
I realize this is not the issue of the post, but it is what I have been thinking of recently.
Comments are closed.