You're looking at Less Wrong's discussion board. This includes all posts, including those that haven't been promoted to the front page yet. For more information, see About Less Wrong.

If interventions changing population size are cheap, they may be the best option independent of your population ethics

6 ericyu3 13 August 2014 03:03AM

In this post I'll explain why you might want to assist altruistic interventions that change the size of the world population regardless of how valuable you think additional lives are. The argument relies on a combination of 2 population-changing interventions that combine to produce the effect of a non-population-changing intervention, but at a lower cost.

Suppose you can donate to the following 3 interventions:

  • "Growth": increase one future person's income from $500/yr to $5,000/yr for $10,000
  • "Plus": cause one more person to be born in a middle-income country (income ~$5,000/yr) for $6,000
  • "Minus": cause one less person to be born in a poor country (income ~$500/yr) for $1,000
Assume that the interventions are independent, and that donating multiples of the cost produces multiples of the effect without diminishing returns.

The cost estimates are completely made up; the point of this post is to explain what happens if the total cost of Plus and Minus is less than the cost of Growth. The cost of Plus is probably least well-known, since it's the least popular of the 3. Also, in the real world, you would probably want to spread the impact of $10,000 across at least several people instead of increasing one person's income by 10x, but I think the post makes more sense this way. If you know a more reasonable estimate for the costs, please post them!

If you donate to Plus and Minus, the total effect is the same as the effect of Growth in many ways - in the future, there is one more person with income $5,000, one less person with income $500, and the size of the world population remains the same. In my last post, I asked about whether consequentialists actually view the two outcomes as equivalent, and people seemed to think yes, so it's reasonable to say that Plus+Minus is just as beneficial as Growth. But Plus+Minus only costs $7,000 while Growth costs $10,000, so regardless of your population ethics, you should prefer donating to Plus+Minus.

But unless your population ethics are "fine-tuned" to make Plus and Minus equally cost-effective, one of them will be clearly better (more cost-effective) than the other. If you think Minus is better than Plus, then Minus is better than Plus+Minus, which is better than Growth, so you should donate exclusively to Minus. The same argument applies if you think Plus is better than Minus. If you donate to only one of Plus and Minus, you will change the size of the world population. So this seems to show that if population-changing interventions are cheap, you should act to change population size regardless of what you think about population ethics. Even if you are very uncertain what the value of a new life is, you can still use your best guess to decide between Plus and Minus as long as you are risk-neutral about how much good you do. 

Numerical example: suppose that Growth yields 100 "points" of benefit, where "point" is an arbitrary unit. Then regardless of population ethics, Plus+Minus yields 100 points as well. How these points are distributed between Plus and Minus depends on your population ethics, however. If you are a total utilitarian, you might say that Minus is worth -20 points and Plus is worth 120 points, and if you're a negative utilitarian, you might say that Minus is worth 150 points and Plus -50 points. If you're an average utilitarian, you might say that Minus is worth 70 and Plus is worth 30. But these all sum up to 100, and they would all choose Plus or Minus over Growth: Plus for the total utilitarian and Minus for the others.

What might be wrong with this reasoning? I can think of a few things:
  1. Plus+Minus is more costly than Growth in reality (quite likely)
  2. Growth and Plus+Minus are actually not equivalent, since Growth actually helps a particular person (again, see my last post)
I'm really curious about what the costs of economic-growth and population interventions are. I'd guess that population interventions would be competitive with unconditional cash transfer programs like GiveDirectly, but I don't know that much about their effectiveness, and I don't know whether there are economic interventions that are more cost-effective than cash transfers. Here are some population interventions that can be done or funded by individuals:
  • Education about contraception
  • Having children yourself (cost varies from person to person)
  • Paying others to have children
  • Subsidizing contraception
  • Subsidizing surrogacy (there are replaceability issues here, but I couldn't find any estimates of supply/demand elasticity)
  • Being a surrogate yourself (doesn't cost you any money, but can be unpleasant, so the cost varies from person to person)
Have people made estimates of how cost-effective these are? The Plus+Minus vs. Growth hypothetical doesn't work if Growth is actually cheaper, so I want to know if I'm thinking too much about something irrelevant!

 

Comments (19)

Economics/demographics question: If a child unexpectedly dies, how much does this shrink the next generation?

1 ericyu3 07 August 2014 06:53PM

The answer seems obvious - the next generation will have one fewer person (in expectation) - but it's not that simple, and it's been bugging me for about a day now.

Suppose you are an average 15-year-old, and your parents are too old to have any more children (they won't have more children to "replace" you). The ~2 children you would have had obviously won't be born. Naïvely that means the next generation will be smaller by 2, but this disagrees with the obvious answer (smaller by 1).

Where this reasoning goes wrong is in assuming that everyone else will still have the same number of children. The sex ratio will shift so that the surviving members of your sex have n more children, and the size of the next generation will decrease by 2 minus n. If n is 1, we get the intuitive answer that there'll be 1 less person.

But there's no reason why n has to be 1 for both sexes! If both a boy and a girl die, the sex ratio is unaffected and the next generation will be 1 smaller, so n has to average to 1, but n may or may not be the same between sexes. Have there been any studies estimating the value of "n" for each sex?

(I posted this because it's relevant to population ethics, but I'm not entirely sure whether it belongs here, so I also posted it to Reddit. Should questions like this go in Discussion or in an open thread?)

Comments (12)
Powered by Reddit Powered by Reddit