I’m on vacation (my idea of a vacation is to visit a resort called the Research Program in Development Studies/Centre for Health and Well-Being and then a demography conference, so form your own judgments about my credibility), so I hope r.i.c.e. readers will indulge me in an unusual blog post. Readers, you can do me the favor of giving a friendly smile as I unburden myself of a worry that has lately kept me up at night: are we overlooking something important when we overlook existence externalities?
“Existence externality” is the cumbersome name I’ve provisionally given to the fact that more (or fewer) people might be born or caused to exist because of somebody’s decision. For example, providing health insurance coverage to a young family in the U.S. might allow them the security and stability happily to have a baby that they would not otherwise have had. I don’t recall the policy debate including the benefit to this baby of getting to exist (or the children this baby will someday have). Or, educating a young girl in rural Uttar Pradesh may improve her life in many ways, potentially including the consequence of reducing by one the total number of children she has.
In both cases, there may be many consequences of of these policies, some good and some bad. There are also well-known possible good and bad consequences for other people of an extra person existing. In different contexts within the wide variation of human lives, having an extra child has made some women much better off and some women much worse off than they otherwise would be. Here, I’m highlighting a particular consequence that is rarely taken into explicit consideration: a person exists, and gets to live a life, who otherwise would not.
Assuming the extra person gets to live a life worth living (a non-obvious assumption that this literature often makes) it seems obvious to me that this is an additional benefit of the policy. Conversation, however, makes clear that what strikes me as obviously true – that extra happy lives worth living are sometimes one benefit to be considered alongside others – strikes many people as obviously false. This is – perhaps puzzlingly – despite the fact that essentially all such people agree that it is preferable not to shorten (painlessly, instantly, and surprisingly) the life of a person with no social ties. Insofar as my fellow economists evaluate policy on the basis of its effect on average GDP per capita, or the fraction of people who are poor, they are implicitly defining existence externalities out of, well, existence.
Any existence benefit is certainly not automatically decisive. It must be weighed along with other costs and benefits. To be very clear: there is much suffering, especially for low-ranking women, in many of the high-fertility societies that humans have historically actually constructed. I merely ask whether adding an extra life worth living is a benefit that must be considered in the mix. If so, it may have serious policy implications. If there are such things as existence externalities, then quantitatively they may add up to a very big deal.
The branch of philosophy that has considered related questions is called “population ethics.” (You can read a clear statement of some of the big puzzles in population ethics here.) Many people believe that population ethics is not very relevant these days. Although a few decades ago many policymakers were worried that the rapid population growth of the mid 20th century would continue or accelerate, now we know the world has, as David Lam puts it, “survived the population bomb.” For many reasons, fertility has fallen in most places around the world. So, it may no longer seem worthwhile to debate whether, say, a developing country would be correct to restrict couples to one child, in the name of the greater good.
In this blog post (all that was just the introduction, I’m afraid), I’m going to consider five dimensions of the possible relevance of population ethics. I believe that the questions raised by population ethics and existence externalities may be of urgent relevance to real policy decisions that are taken every day. However, I will also suggest some ways in which debates in population ethics could be made even more relevant to important questions. I will not, in this blog post, establish a claim that it is better to exist, or that the world is better when more people exist;* instead, I am arguing that it would be relevant to know whether or not this is the case, especially for certain questions of public policy.
1. Existence externalities would be very important for real policy decisions that happen all the time.
I have many friends who work at the World Bank, a large international bureaucracy that gives loans to poor countries, but also conducts economic analysis of programs and policies and makes policy recommendations. Every day, some World Bank economist is writing a memo to a minister about whether or not a particular policy is a good idea. This happens even more often within government bureaucracies in rich countries and poor countries, both. Indeed, program and policy evaluation is a popular class for policy students and is a big business for consultants. But I do not know of any example where extra lives lived (and worth living) were counted as a benefit.
But it is beyond dispute that many policy decisions influence the number of lives lived, if only because they influence fertility decisions. Consider a policy-maker in a densely populated Asian country (perhaps of a few decades ago) with an infant mortality rate of 150: that means that 15 percent of babies die before their first birthday. Because mothers and families know this, they have many babies so that some will survive. The average mother has three births, and on average 2.55 of them live lives beyond infancy (which we can assume here are complete).
Our policy-maker is deciding whether or not to implement sweeping public health and sanitation reforms which will lower the infant mortality rate to 50: now one in ten more babies born will survive to live a full life. Knowing this, however, mothers freely choose to only have 2 babies instead of 3, and on average 1.8 survive to live their lives.
What are the consequences? Mothers are better off: they suffer through the pain of fewer of their babies dying; they are required to perform less childcare and carry fewer pregnancies; and insofar as they have fewer children, it is only because they themselves chose to, which surely does not make them worse off. Babies who are born are better off: they are more likely to survive infancy. Although taxes were raised to pay for the reforms, the rest of society is almost certainly better-off too: improved early-life health is increasingly well-understood to improve human capital, so the surviving babies will go on to be more capable and productive.
However, fewer people get to live. Assuming that their lives were not so bad as to not be worth living, I propose that this is a cost that our policy-maker should take into consideration. It is not clear that this social cost outweighs the other benefits, which are many and large. But sometimes it might.
Of course, if we want to get serious about including existence externalities in our program and policy evaluations, we have the tricky question of how much an extra life worth living is worth, economically. But economists deal with such issues all the time. Papers upon papers estimate the “value of a statistical life,” the amount of money that the government should be willing to spend to avert an additional death (when building highway safety measures, for example). In a rich countries, I think this is often taken to be a number in the low millions, but I have seen smaller numbers and larger numbers. I seem to recall a death of a statistical person in India being valued at an order around $100,000 in WSP’s Economics of Sanitation computations, for example.
If we like how “value of a statistical life” numbers are computed (and we may not), one starting point would be to use the same numbers for “value of a statistical existence.” This would almost certainly be too large. Part of what society is willing to pay to avoid in a traffic death is the suffering of the victim’s family, friends, coworkers, and projects, none of which would be the same for a life that never happened. On the other hand, exactly because existence is an externality (it is not chosen by those who exist) we cannot use value of a statistical life methods to evaluate it: these ask how much people are willing to pay to prevent probabilistic death. In short, the fact that there is no way to apply economists’ revealed preference methods to one’s own existence, appears to construct a blind spot into any choice-based policy evaluation. Of course, this does not stop policymakers from valuing improvements in newborns’ well-being, even though they cannot reveal the preference.
However, we certainly do not need to know the exact value of a statistical existence. We do not know the exact value of a statistical death, and that does not prevent policy decisions from being made about whether or not to build new bridges. My day job, when I’m not writing overlong blog posts, is to write about whether the Indian state and society should be investing in healthier, larger babies and children; yet, I don’t know exactly how large the benefits of this would be, merely that they are very large.
I want to pause here to emphasize that throughout this blog post, I have been considering the impact of policy decisions on the number of children that women and families freely choose to have. It seems clear to me that a world in which people get to choose their fertility (if they want to so choose) is better in many ways than a world in which people do not have the ability to choose.
Moreover, compulsion can be complicated, especially economic compulsion. If we did decide that the value of a statistical existence were two million dollars, it would not be difficult to imagine very bad consequences of the government deciding to raise enough taxes to offer women a million dollars if they elect to have an extra baby (society makes a million dollars social profit!). Worse still, it strikes me as likely that if humans tried to build a society in which women were institutionally forced to have many babies, then in actual fact the suffering and unfreedom that would result would make for a world obviously much worse than the one we live in. Nothing I’m writing here need recommend such a disaster.
The point is the following: policy decisions impact the set of people who exist. This is inevitable. We can either ignore this fact – perhaps comfortably or even correctly by asserting that existence is of no relevance for social welfare – or take it into consideration.
2. Externalities are of special importance to policy, and existence externalities are no exception.
Societies and economies are complicated, so trying to control too much can lead to big problems. In Seeing Like a State, James Scott offers compelling evidence that some of the worst disasters of recent decades came about when rationalizing, optimizing bureaucracies tried to control everything into too much perfection. In a wide range of cases, we can do little better than to let people make their own decisions.
Economists emphasize an important example to this general rule: externalities. Externalities are situations in which one person’s choices impact a second person’s outcomes, in a way that is beyond the second person’s control but that the first person does not care about. So, if I defecate in the open and make your children sick, that is an externality that hurts you in addition to whatever consequences it has for me. Externalities are a classic case where free markets cannot produce a good outcome because people cannot be trusted to take into consideration the interests of others.
Without much further comment, I will observe that existence is perhaps the ultimate externality. If we agree that potential people have an interest in existing – and I admit that I have only considered this, not argued for it – then we must also agree that this is not an interest that they can advocate. Neighbors can complain; fellow citizens can vote; and even many non-human animals can express pain and suffering. Potential future people are quite external to any decision that we might make about their welfare. So, one reason that existence externalities are relevant for public action, if they exist, is that they are externalities.
3. Keep it real: we can ignore extreme conjectures, if they distract us.
If you clicked the link onto Derek Parfit’s essay above, then you read about his so-called “repugnant conclusion.” In short, he suggests that – if people living extra lives worth living is indeed a benefit – then it could be better to have a universe with a positively enormous number of people living lives that are barely worth living than to have the world we live in (or even a world with 8 billion people all living lives as good as the best life anyone enjoys today). This strikes many people as repugnant: what would be so good about trillions of trillions of people living lives barely worth living?
One answer is that we humans are pretty bad at thinking seriously about trillions of trillions, so we shouldn’t really trust what seems true to us in these cases. But another answer – and certainly not one that I am the first to offer – is to observe that while population ethics may be urgently relevant, population ethics for a species of trillions of trillions is not.
Economists write about “social welfare functions” which map from the distribution of well-being across people into a ranking of how good a society is. It may be that in social welfare functions that are not wrong, the mapping from quantity of people into overall social welfare is complicated and not linear or not additive in this range. I don’t know: this is only something that humans have been thinking seriously about for a few decades, and we have no experience with trillions.
When a statistician is trying to fit a model to data, the answer is almost always an approximation. This approximation can be increasingly bad far away from the range of the data that the model was designed to fit. Similarly, we might reasonably settle on ethical rules of thumb (and be stuck with automatic intuitions) that are very bad approximations in situations that are “far” from the decisions that humans face in any choice sets actually available to us. But why would this be a practical problem?
[I think there’s something related to say here about whether the difficulty and, to some, apparent absurdity of sacrificing almost all of our resources to help others is an argument against an obligation to make more modest sacrifices, but this is already getting too long.]
My point here is only that I don’t know, or need to. There are many decisions on manageable, human scales that shape the set of people who exist and that must be influenced by policy decisions that either must be made (or omitted to be made, with other consequences). Existence externalities would matter for these decisions.
4. It does not diminish the policy relevance of existence externalities that some people do not believe that they need to help people who already exist, either.
There’s not much more to be said about this. If you don’t agree with me that we sometimes have obligations to help strangers merely because we can improve their well-being or prevent bad things from happening to them, then you might not agree that these obligations could extend to potential future persons. But some people do agree with me; this is a conversation for people who agree that something like social welfare exists and is important.
5. Population ethics could be even more relevant if it had more to say about lives that are not worth living.
When we hear about the repugnant conclusion, perhaps we find it repugnant because we think about some of the lives and human experiences that exist in our time and have existed over the past few centuries, and we realize that some of these experiences were very, very bad. A life that cycled relentlessly among famine, death of many of one’s children, patriarchy, racism, casteism, fear of separation from one’s family, civil war, mass incarceration, torture, rape, drone strikes, toil, mockery, boredom, and disease may well not be worth living – but unfortunately does not require advanced imagination. Such suffering is fact.
So when Parift invites us to compare our world to one with many, many, many people living lives barely worth living, he may well be proposing an alternative where the worst off are much better off than the world in which we live today. I propose that for millions who have lived and died over the past few hundred years, this “repugnant” alternative would be an improvement. On the flight from India yesterday, Diane watched a recent movie about slavery in the U.S. South in which a character asked to be killed. Although this case is fiction, it is not fanciful: is it clear that this woman was wrong that her life was not worth living?
If I am right that the “worth living” threshold cuts through the distribution of past and present human experiences – rather than running below it – then Parfit’s conclusion may not be repugnant, but instead all too much pie-in-the-sky. “Of course,” a low caste daughter-in-law in rural Bihar of a few decades ago might reflect quietly after another baby dies, “it would be an improvement to switch to a world in which everybody lives a life worth living!” Perhaps population ethics could be a better guide for policy if we thought more about what this commonly invoked “life worth living” is. Perhaps many women in high-fertility, impoverished, patriarchal societies have had many children, have not themselves had lives worth living; or, for all I know, essentially all of them have. It’s probably not for me to say. But isn’t this relevant?
* If I were going to try to establish that it is better, all else equal, for policy-makers to choose a world where additional people exist, all of whom live lives worth living, I would probably go for an absurd contradiction around eggs [there is an unknown but finite and presumably approximately knowable number of fertilizable human eggs in existence] and babies and unfolding personhood, but there’s really nothing interesting or new there. Moreover, if biological facts happened to be different such that I could not make these boundary-discontinuity absurd comparisons, I suspect I would still believe that creating an extra happy person without making anyone worse off is good.
If we believe that people must be persuaded that all this is important for policy choices, experience shows that it won’t get us anywhere to be able to tell people their policy views imply something absurd. Alas. What we really need to persuade a lot of people is a simple forward-going affirmative proof.
Although maybe I’m wrong about that, too pessimistic. These days you hear people say useful things like “denying that gay people’s interests count equally is like being racist.” Is there a blank that can be filled in: “denying that probabilistic future people’s interests count [and all future people are probabilistic] is like ____”
Empirical question: do people people who have decided to have a child but have not yet conceived the child often make decisions as though that child had ethical interests which must be weighed? I think so.
I make no claim that any of these thoughts are original, and I am certain that some of them are not. I am grateful for discussion with Diane Coffey over the past week or so as I babbled about these issues, although she has no blame for this drivel, and for United Airlines for creating such a congenial environment in which to host a visiting research scholar.