[content warning: beep boop I am a robot; this post uses numbers to analyze the benefits to the world of having children, which may be offputting to non-utilitarians. In addition, people who are distressed by the idea of compulsory parenthood may find that this post is either deeply upsetting or an opportunity to identify as an average utilitarian]
Having children is a surprisingly effective form of altruism.
The average cost of raising a child for a middle-income family is approximately $300,000. However, this number does not include opportunity cost of not investing the money, lost income, and college; when you count those three, some guy at Time Magazine says it actually costs $900,000 for a middle income family.
If you are a total utilitarian, the math here is pretty easy. The life expectancy of an American child is 78; hence, you get 78 QALYs, for a cost per QALY of approximately $11,000.
However, Americans are happier than people in other countries. How much happier? The average happiness worldwide is 5.1 on a one out of ten scale; Americans are at 7.1. Arbitrarily deciding that one year of a 10 life is equivalent to two years of a 5 life, the cost per QALY of having a child for total utilitarians is $5500.
If you are an average utilitarian, the math is a bit more complicated. The world life expectancy at birth is 71. Creating a child who lives seven years longer is equivalent to adding three and a half years to two people’s lives, for an unimpressive $250,000/QALY.
However, if we adjust for Americans’ happiness advantage, creating a single American is equivalent to raising two people’s lifelong happiness to 6.1. Adding an additional American is equivalent to 19.5 QALYs, for a cost per QALY of $39,000
Comparison time! The cost per DALY of a malaria net from the Against Malaria Foundation is very very conservatively $38, and probably substantially less, as that calculation does not include any effects except deaths of children averted. It is unlikely that having a child will be more effective than donating to the Against Malaria Foundation.
However, NICE’s threshold for cost effectiveness of a health intervention is about $30,000 (20,000 pounds) per QALY. Therefore, for total utilitarians, having a child may be considered a cost-effective intervention, although not an optimal intervention.
Note that this is the average amount of money a middle-income family spends on their children. It seems probable that a middle-income family is wasting a lot of their money. You can consider dressing your child in second-hand clothing, cutting their hair yourself, buying them few and inexpensive toys, and making your children share their rooms with their siblings. However, the best thing you can do to reduce your costs is to have more children. The marginal cost of a child is much, much less than the average cost of a child. A single eleven-year-old costs $16,000 a year; her and her sixteen-year-old sibling cost $13,500 each; add a third kid and you’re down to $11,000 each.
What conclusions do I draw from this post?
First: I see a lot of effective altruists who plan on having children say “well, it’s really expensive, but nobody is a perfect utilitarian.” This seems to me to be unnecessarily apologetic. If you imagine a spectrum of Not Perfectly Effective Things, where giving to Oxfam is on one end and lighting a bunch of money on fire is on the other, having children is clearly more toward the Oxfam end than the lighting money on fire end. For a total utilitarian, having a child is equivalent to paying $450/month out of pocket for a medication that will keep someone they love alive– perhaps not what a perfect utilitarian would do, but if someone calls you on it you can go “what the fuck, asshole.” (Average utilitarians may continue to be apologetic.)
Second: if you plan on having children, consider having lots of children.
Third: there are some people who should consider specializing in having children. A lot of the cost of having children is lost income; if you don’t earn very much and really like children, consider becoming a parent who works part-time or stays at home. Being a stay-at-home parent opens up a lot of opportunities for frugality, from home-cooked meals to not hiring babysitters. Instead of paying a premium to live in an area with good public schools or sending your children to private school, you can homeschool (although do run a cost-benefit analysis on that; whether it’s worth it depends a lot on your career and area). It is probably best to marry a fellow effective altruist who is high-earning or does direct work and really really wants kids, and then plan to make all the career sacrifices yourself; that way, you can take credit for all the extra money your partner earns because they don’t have to stay home with sick children. You can get additional utilitarian points by babysitting for other EAs or even running a day care out of your home.
Fourth: surrogacy is an underexplored way to do good. Rather than costing money, the first-time surrogate earns thirty thousand dollars, which can grow to forty thousand dollars for experienced surrogates– and it still creates 109 QALYs that otherwise would not exist. These children are likely to grow up in wealthy families who really, really want to have them, and are thus likely to be even happier than this analysis suggests.
The total utilitarian may consider having children and giving them up for adoption. However, this is inferior to surrogacy for several reasons. First, it is very difficult to explain to others and, if widely adopted, may present a PR problem for effective altruism as a whole. Second, you are not paid any money for it and have to pay for your own medical fees. Third, adoption (paradoxically) presents more replaceability concerns than surrogacy: since adoption is substantially cheaper than surrogacy, many parents will try to adopt and, if that doesn’t work, hire a surrogate. Parents who hire a surrogate either want a biologically related child or don’t want to wait for an adoption. If you give up your child to someone who would have otherwise hired a surrogate, instead of a juicy 109 QALYs, you get 0.
Thing of Things 
Bringing additional humans into the world increases the strain on finite natural resources, and this effect is much greater in the U.S. than in most other parts of the world. I admit I don’t know how to quantify the effects of resource consumption for utilitarian calculations. But given that overpopulation and scarcity are usually central to discussions of the ethics of having children, it seems strange to ignore these considerations entirely.
LikeLiked by 3 people
Most developed countries are close to zero population growth, and some are even shrinking. I suspect the marginal human is positive. (Besides, total utilitarians would support more people having lower standards of living.)
LikeLike
Estimates for the size of a sustainable human population appear to mostly range between 2 billion and 10 billion, and the meta-analysis here (http://bioscience.oxfordjournals.org/content/54/3/195) suggests that the best point estimate is around 7.7 billion. Meanwhile most estimates of population growth over the next hundred years suggest the total population will reach 10-11 billion. It seems likely that at some point in the next couple hundred years, the population will decrease substantially due to a Malthusian catastrophe. This transition is likely to cause a great deal of suffering. Surely even a total utilitarian would agree that it would be better for the necessary drop in population to be as small as possible.
And even if the population never rises above sustainable carrying capacity, it’s not obvious that total utilitarians should see a larger population as preferable. The drop in happiness due to increased competition for resources could outweigh the benefit of an additional person existing and having experiences.
LikeLiked by 1 person
@unknot
I find it quite unlikely that the human race will remain biological for the next hundred years. Even if it did, I doubt that there are any resource crunches coming up we can’t invent ourselves out of.
My biggest concern in regards to Malthusianism is the possibility that if the human race becomes mechanical in the future we could use up resources by manufacturing people as quickly as we make computers. But if that’s the case, making more people the normal way isn’t going to make a big difference in that regard. It’s less than a drop in the bucket.
LikeLiked by 1 person
Total utilitarian here. I don’t support more people with a lower standard of living, because I think there are a lot of lives out there with negative value. We need to make sure that the lives being created are above zero, and that means paying close attention to quality of life.
LikeLiked by 1 person
The estimates that I quoted attempt to take technological progress into account, and their authors did not reach the conclusion that we can “invent ourselves out of” the associated problems. There’s obviously a lot of uncertainty when dealing with future technology, so they could very well be wrong. But if you have good reason to believe that the estimates are overly pessimistic, I’d love to hear it.
If you believe that in the near future, humankind is going to ascend into a glorious techno-utopian future with a much larger population than the current one, pretty much all bets are off. Other ethical imperatives pale in comparison to making sure that that future happens, making sure it happens as soon as possible, and laying the groundwork so that it will be as nice for those posthumans as possible. If such a future is feasible at all, then I would think that one of the largest threats to its realization would be fossil fuels becoming scarce before we can bootstrap to a sustainable energy source. In which case you should avoid having children to keep fuel consumption down so that existing reserves last as long as possible.
LikeLike
@wireheadwannabe
How do you determine whether a life has negative value? Like, what quality of life is exactly as good as not existing at all? The only standard I can think of that makes any sense is asking people whether they would prefer not to have been born, but this isn’t a very good one; most people will answer in the negative regardless of their quality of life.
This isn’t intended as a gotcha; these are questions I’ve been struggling with and I’m really interested to hear what other people have to say about them.
LikeLiked by 1 person
I’m very worried about it too. Natural selection inevitably favors those who believe their lives to be worth living, with little regard to whether it’s actually true. We know people are terrible at evaluating their past or future happiness. We may just be screwed. Or maybe we’ll figure out a way to objectively measure happiness and get bailed out.
Moloch, whose name is the mind!
LikeLike
I wouldn’t have minded not being born, and a non-negligible fraction of the time I prefer it.
LikeLike
>If you believe that in the near future, humankind is going to ascend into a glorious techno-utopian future with a much larger population than the current one, pretty much all bets are off.
I consider this a possibility, but I am worried that the specter of Malthus will follow us past the singularity, and that it might actually be exasperated by the ability of posthumans to mass-produce themselves.
>The only standard I can think of that makes any sense is asking people whether they would prefer not to have been born, but this isn’t a very good one; most people will answer in the negative regardless of their quality of life.
One idea that comes to mind is asking people how much of their daily routine they would rather be unconscious for. This removes the “death is scary” element, and focuses purely on the quality of the experience.
Of course, you’d need to eliminate extra variables for this, which would make the question a little harder. For instance, if you ask someone if they’d rather be unconscious than at work, you need to emphasize that this is a hypothetical, so they will not suffer any career setbacks for being asleep instead of working, the work will still get done, they will not feel more rested from being unconscious, and they won’t dream.
I’m not sure how many people you could get to entertain the thought experiment correctly, but it’s a start. I’m lucky enough to prefer the vast variety of my conscious experience. In fact, I love my life so much I often go without sleep because there is so much fun stuff to do!
LikeLike
So, basically temporarily becoming a p-zombie? In the novel Echopraxia, there are people who are modified so that in military situations their conscious mind shuts down and the subconscious takes over in order to fight more effectively. I think that’s an example that fits.
If I’ve procrastinated on homework or the day’s exercise, that reliably makes me feel depressed, especially at an hour or two before bedtime. That seems to me like a part of the day that’s worthless and something I’d rather be unconscious for, but in that case it’s not just a thought experiment—I could just go to bed earlier, there’s nothing stopping me, but I don’t. I dunno, weird.
LikeLike
At least at the moment, shouldn’t there be a preference (slight or strong, I can’t say) for adopting or fostering children, since a large number of children in the system will not have lives of at least average happiness? (And, for that matter, a preference for adopting children from poorer countries?)
Or am I misunderstanding the math for total utilitarianism?
/is not a utilitarian of any sort, so that may well be possible
LikeLiked by 2 people
The total utilitarian reasoning is that if an adopted child has a life worth living, it is better to procreate, to increase utility by (utility of one’s child) rather than (utility of adopted child – utility of child in care).
LikeLike
yes, and because of increased opportunity for specialization and trade each additional human is a greater boon the the whole of humanity than the previous one!
LikeLike
Agreed, this post only seems to count the QALYs the child personally generates. It doesn’t take into account the QALYs the child will likely add by participating in the global economy. It seems like that would be an important factor to consider, although it might be a lot harder to calculate.
Since average income is around $1-2 million over a lifetime I can at least conclude that people add at least that much value to the world around them. But that ignores consumer surplus. And I’m having trouble translating that dollar amount into welfare.
LikeLike
Estimating the relative happiness of Americans for some 78 years is a fairly long term prediction. There are people who predict disaster for America, though I think they also tend to predict disaster for the rest of the world.
That longer lifespan– there’s a lot of reason to think that American end of life care produces misery, which should be added to the computation somehow. On the other hand, policies and technology may change– as stated, 70+ years is a long time.
Burning a pile of money is far from the least utilitarian thing you can do– it’s only destroying a trivial amount of wealth. I have no idea what would tend to be the least utilitarian thing you can do would be, especially since we’re looking for high probability, and I leave the topic open for discussion. The worst things that come to mind easily relate to really bad ideas, and any particular idea has a low probability of being adopted.
LikeLiked by 1 person
Burning money is equivalent to donating it to the government, so how utilitarian you think it is depends on how effectively the government would use your donation.
LikeLike
I think it’s closer to reverse counterfeiting. The counterfeiter devalues all the money in existence by making more, the money burner increases the value of all the money in existence. So really, by burning money you’re donating it to everyone.
I suppose what would be the optimally anti-effective altruist use for money would be something like buying mosquito nets and then burning them. Or buying a mosquito net factory and shutting it down.
LikeLike
Is it me, or does the pessimally utilitarian action involve a lot more atrocity & torture than just burning some nets?
LikeLike
Ghatanathoah:
That’s without taking into account monetary offset, ie the response of the central bank. Since the central bank manages the money supply in order to hit some nominal target, money burning directly translates into dollars that the central bank can print without causing any inflation, compared to the counterfactual where you spent the money instead of burning it. Central bank profits are turned over to the treasury, so the ultimate result is that you donated to the government, not to everyone. You could argue that the government could respond by adjusting taxes, and then you’d really be giving the money to everyone, but in the current system the central bank really does adjust the money supply according to market conditions all the time, and the tax rate is adjusted much more infrequently, and for reasons only loosely connected to the budget surplus or deficit, so it’s much more realistic to say the donation is incident on the government than on the people.
LikeLiked by 1 person
@Lambert
I was assuming violence was off the table when evaluating the worst possible IEA action, the same way violently overthrowing evil dictatorships is off the table when considering the best possible EA action.
LikeLike
Also, you generally discount QALY gains at the same rate one discounts money (or else you get wacky perverse incentives). NICE, which Ozy cited in their post, discounts at a rate of 3%/year IIRC, so those 7 years at the backend are only worth about an eight of a current QALY apiece.
LikeLike
I think you underestimate the costs of childcare. My toddler almost exclusively wears clothing that her older friends have outgrown and I cut her hair myself, but I’m still staring down almost $2K a month for preschool because there’s only so much I can stimulate a child at home.
However, I agree that being a stay-at-home parent often is the cheaper option if only because of childcare, and that the lost income can be divided by multiple children.
LikeLike
Preschool and childcare are definitely budget killers. The breakeven point for a parent to stay home can easily be about 30-40K just on that part.
LikeLike
May I ask what happened to your tumblog? Did you change URLs, or delete?
LikeLiked by 1 person
Isn’t this just the argument for tiling the universe with rats that prefer to be tiled across the universe?
LikeLiked by 1 person
QALY maximization makes sense as a standard for altruism for a fixed population, which is usually an acceptable simplification for most problems. But it’s not at all clear that it should be the standard when the population size being variable is central to the issue. It suggests that letting someone live for 50 years, killing them painlessly, and replacing them with another person who lives for an additional 50 years would be equivalent to one person living for 100 comparably healthy years, which is a highly counterintuitive result.
LikeLike
Agreed. I think the value of a life grows faster than linear with the number of years. It would be nice if someone came up with a way to make this quantitative.
LikeLike
I’m surprised that you find the whole “two people living 50 years is better than one person living 100 years” result to be counterintuitive. This is the intuitive result for me, because most people have a finite amount of things they want to do in their lives, which means that an additional 50 years after the first 50 doesn’t mean as much (since people have already accomplished most of their major life goals by age 50), whereas a first 50 years means a ton to a new person.
I know there’s a word/phrase/term for what I’m talking about here, but I can’t remember what it is. I think it involves “discounting,” though. Anyway, I definitely think having two people live 50 years is better than one person living 100 years because if we assume that all three postulated people have roughly the same number of life goals, more life goals will get accomplished by the two who live to 50 than the one who lives to 100, which generates more utility.
LikeLike
The way I see it is – for whose sake are we doing all this? It’s possible to make existent people worse off in various ways, such as by decreasing their QALYs, but preventing someone from coming into existence isn’t a harm, because one has to exist in order to have interests. So when there’s a tradeoff between creating new people and making existing people better off, the latter seems obviously better. So if Scenario A is one person living for 50 years and then being replaced by another person who lives for an additional 50 years, and Scenario B is the first person living for 100 years and the second person never existing, Scenario B is better.
My intuition when it comes to population ethics is to maximize utility per person (not total utility) subject to the constraint of not killing people for population-ethical reasons (to avoid scenarios in which it’d be right to kill sad people against their will).
LikeLike
You can take average utilitarianism as your foundation with no additional constraints, and still not be motivated to kill sad people against their will. A world in which you can be executed for being too sad would be a nightmarish dystopia in which people are terrified of being sad and hide their negative emotions for self-preservation. Which would be pretty bad for everyone’s utility.
LikeLiked by 1 person
That’s one way of justifying the constraint. Another is to treat death as a hugely negative-utility event.
LikeLike
@blacktrance
I agree that making existing people’s lives better matters more, and I also don’t think non-existent people “matter” for the purposes of figuring out how to distribute utils. I was just thinking about this as though none of the three postulated people (the two who would live to 50 or the one who would live to 100) existed. I figured it was a trade off between creating the two or creating the one, rather than being a trade off between extending one already-living person’s life or creating two new people.
If none of them existed and we were trying to figure out which to create, I’d go with the two 50-year-life-expectancy people, but if the one guy who wants to live to 100 already exists I’d go with him.
LikeLike
If you ground utilitarianism in contractualism, then from behind the Veil of Ignorance you have a choice between being born as someone who’ll live for a hundred years vs a higher chance of being born as someone who’ll live for 50 years. If you turn out to not have been born to begin with, you aren’t behind the Veil, either, so it doesn’t matter, and the expected value for you is higher in the 100-year scenario.
LikeLike
Ozy, I’m glad you opened discourse on this important question! The calculations you present are interesting. However, for me the importance of having children always hinged on a different consideration, namely the opportunity to impart my values and worldview to these children.
The claim I’ve seen circulating in EA circles is “it’s easier to convince new people to become EA than raising EA children”, but something about it seems off-key. Namely, what I observe in the real world is that most people hold worldviews which were strongly influenced by their parents. This suggests that a movement that relies entirely on “horizontal” transmission of its ideas will find it challenging to survive in the long run.
Moreover, it seems to me that stable tightly knit communities usually revolve around families, and that this has to do with some fundamental properties of human psychology / social behavior. Therefore, if we believe there is value in building a tightly knit EA community (and I believe there is) then EA families must be a part of it.
LikeLike
Better yet, move to Switzerland, then have lots of kids. Since kids there have significantly longer expected lifespan and higher happiness rating.
LikeLike
Pingback: On The Allegedly Repugnant Conclusion | Thing of Things
i don’t think this argument stands with negative utilitarianism,though?
LikeLike
I’m not sure you can count all 78 years of life expectancy as QALYs – doesn’t that assume they have full health until death? Presumably many people spend some of their life at less than full health.
Data on what we, as pretty frugal people, spent the first year ($20k including opportunity costs): https://thewholesky.wordpress.com/2014/12/23/how-much-does-it-cost-to-have-a-baby/
LikeLiked by 1 person
One other thing I just realized is strange – “The cost per DALY of a malaria net from the Against Malaria Foundation is very very conservatively $38, and probably substantially less, as that calculation does not include any effects except deaths of children averted.”
The link to the source isn’t working, and I know the numbers were different when you wrote this post, but I think that cost is low even for 2015. GiveWell’s estimate for AMF preventing the death of a child under 5 (extending a life by 30 years) is in the neighborhood of $7,000, which is more like $233/DALY. Still much lower than most of us would pay for a year of healthy life!
http://www.givewell.org/how-we-work/our-criteria/cost-effectiveness/cost-effectiveness-models
LikeLike