Killing Children is Immoral

Ed Glaeser repeats an argument I first heard Steven Landsburg make: children produce significant positive externalities:

When parents decide to have kids, they are creating a massive benefit for their children. As much as parents may love their children, they are unlikely to reap all the benefits those children will offer during their lives. Economists often think that it makes sense to subsidize behavior that generates big "external" benefits for others: parenting seems like a particularly natural example of such behavior.

Greg Mankiw begs to differ:

Ed is implicitly comparing the utility of having been born with the utility of never having been born. But since we do not observe those people who were never born, how can we possibly know their utility?

Mankiw's objection is silly.  If we're going to allow that the unobservability of the unborn's utility is a reason for agnosticism in regards to our population policy, then why not in regards to everything?  We should just as surely be agnostic about whether or not to kill people--for all we know, after death they experience a huge amount of utility, whose unobservability properly renders us unsure about whether or not killing is worthwhile: whether murder is a Pareto improvement. 

Bryan Caplan parries Mankiw:

Actually, this may well be the easiest utility inference in the world. We know that people almost universally prefer existing to not existing because there are so many cheap and easy ways to stop existing.

Here's an argument I've heard before. 

If asparagus plants have little moral worth per se, our question then comes down to this: is it good or bad to create pigs who live for a while and then die? Well, is it good to create people that will eventually die? We usually say yes, if their lives are "worth living" overall. That is, if they get value out of being alive, and are not in a situation like severe torture, where they would rather be dead than alive. So are pigs lives worth living?

We might well agree that wild pigs have lives more worth living, per day at least, just as humans may be happier in the wild instead of fighting traffic to work in a cubical all day. But even these human lives are worth living, and it is my judgment that most farm animal's lives are worth living too. Most farm animals prefer living to dying; they do not want to commit suicide.

If so, the consequence to others of buying that meat in the grocery store, rather than asparagus, is good; you create farm animals whose lives are worth living. And thus the consequence of buying asparagus rather than meat is, by comparison, bad. So if you, like me, think your actions are more moral when you do more good for others, you should agree with me that meat is moral, and veggies are immoral.

From Meat is Moral, by Robin Hanson.

The basic idea is sensible.  Like the run of the mill economist, we're going to judge utility by the principle of revealed preference: we're going to watch what people (and in Hanson's case, animals) actually do.  Caplan's people and Hanson's pigs have a choice between existence and non-existence--the vast majority of the time they choose existence, ergo existence has higher utility than non-existence.  (To be fair, Hanson does allow that some animals' lives may be torture and worse than death, but, he points out, if that's the case then killing animals is merciful). 

What's wrong with the argument?  Nozick, in Anarchy, State, and Utopia, notes it.  I quote Nozick here, though he simply repeats Hanson's argument, because I enjoy his writing style:

One ubiquitous argument, not unconnected with side constraints,  deserves mention:  because people eat animals,  they raise more than otherwise would exist without this practice. To exist for a while is better than never to exist at all. So (the argument concludes) the animals are better off because we have the prac­tice of eating them. Though this is not our object, fortunately it turns out that we really, all along, benefit them! (If tastes changed and people no longer found it enjoyable to eat animals, should those concerned with the welfare of animals steel themselves to an unpleasant task and continue eating them?)

Again, what's wrong?  Nozick points out, persuasively:

[T]he parallel argument about people would not look very convincing. We can imagine that pop­ulation problems lead every couple or group to limit their children to some number fixed in advance. A given couple, having reached the number, proposes to have an additional child and dispose of it at the age of three (or twenty-three) by sacrificing it or using it for some gastronomic purpose.   In justification,   they note  that the child will not exist at all if this is not allowed; and surely it is bet­ter for it to exist for some number of years.

The problem is a problem with utilitarianism qua utilitarianism: it ignores the rights of humans (and perhaps animals).  There are things which we simply should not do to rights-bearing entities, no matter what the balance of utilities dictate. 

Esteemed coblogger and expert unicyclist Brandon Berg repeats Mankiw's argument in the comments of Caplan's post:

Suppose you and your wife decide you don't want to have children. But you could really use some help around the house. So the two of you decide to have a baby and raise him as your slave. By your logic, as long as you treat him well enough that he doesn't commit suicide, this is perfectly legitimate.

And speaking of not having children, it's arguably worse than murder. If you kill a middle-aged person, you're only depriving him of half of his life. But you're depriving every baby you don't have of all of his life.

The utility of not having been born isn't zero--it's undefined because there's no one to look forward to being born or to lament not having the chance. That's why these kinds of comparisons break down.

This argument is not the same as Nozick's.  Neither Brandon (from this narrow snippet of his views) nor Mankiw deny the utilitarian calculus's legitimacy; rather they claim only that it fails to function with undefined quantities.  Berg and Mankiw are incorrect.  Now there may be undefined utilities somewhere in existence, and they may be problematic, but this is not one such case.  The utility of not having been born is not only less than that of what the majority of born beings experience, as Caplan claims, but it is zero (there is no more reasonable means of looking at utility that doesn't come into existence.  Am I holding an undefined amount of uncreated apples right now? No: I'm holding zero apples.) 

The problem is not with the case, the problem is with the utilitarian framework.

UPDATE: The whiff of battered pork seems to be in the air today.