= 27d567bc2d48d9f40e9ccc20ea370b15)
If interventions changing population size are cheap, they may be the best option independent of your population ethics
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:
- Plus+Minus is more costly than Growth in reality (quite likely)
- 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!
= 783df68a0f980790206b9ea87794c5b6)
Loading…= b715d02e85f7b3b4ae65ca3d3e450868)
Subscribe to RSS Feed
= f037147d6e6c911a85753b9abdedda8d){kind=logo}
Comments (19)
Alternatively, we could allow more immigration from poor countries to rich ones.
That would be an example of Growth.
I think this would almost certainly be more cost-effective than Plus+Minus if you were a government, but I'm not sure how easy or hard it would be for an individual to influence their government's immigration policy.
You wouldn't necessarily have to effect policy; you could just start smuggling people in.
I realize this is a made-up scenario, but Westerners tend to assume it's irrational for those in poor countries to maximize their number of offspring. There's a number of factors (labor value tradeoffs, insurance against periods of inability to sustain sufficient income esp at end-of-life) that militate otherwise. Minus could result in a huge amounts of unintended disutility, particularly as populations age.
As pointed out above, the best option is to move the poor into situations where institutions are stronger, producing higher incomes. However, this is very difficult on any large scale. For instance, moving the entire population of Africa to Norway (ignoring space constraints) probably wouldn't increase overall incomes because local institutions would be overwhelmed by new voter preferences and cultural norms. Unfortunately, it's also extremely difficult to develop these institutions in poor countries (the incentives tend to point the wrong way, see Acemoglu on the "iron law of oligarchy").
So the best donation bet might be "meta-Growth" -- attempts to increase the value of Growth by researching more effective methods of intervention to increase Growth.
Say that the universe has Clippy the paperclip maximiser + 9 of his friends and Roger the ruberband maximiser with also 9 friends. Say also that the world is metal rich which makes paperclips easier. You have the options to:
*Increase the production efficiency of paperclips *increase the production efficiency of rubberbands *Decrease the number of rubberband maximisers by 1 *Increase the number of paperclip maximisers by 1
For plus+minus to win out you have to show that an individual would be "better off" converting than increasing efficiency. The upper two options raise the utility value of the world upwards within a single utility function/evaluator. The conversion must somehow make a utility conversion mapping. While I have assumed that paperclipping is easier I have not assumed paperclipping is more moral than rubberbanding. Yet the recommendation seems to be to either work for the paperclippers or try to convert everyone. The rubberbanders got utility-monstered. It's also dubious that converting people will only select the political direction of the world and doesn't impact the ability to purse that direction.
Thus suprisingly treating all values identically ended up favouring one of them over all others. You could have thought that the values would have a similar distribution as in the beginning. It also seems that a person that "wants to do most good for the world possible" is rather doing the thing that creates the world that owes it's existence to the benefactor the most. Thus easily accomplished values will have priority. This deviates from my understanding what it is to do good.
I think the ability to judge the values of others should not be hidden in an implict assumption that all values are equally duty generating. But being insensitive or overtly harsh seems also problematic. It should be recognised as a problem of choice rather than have theories makes such choices for us in an accidental manner.
Interesting, this casts some light on the repugnant conclusion for me. A naive utilitarianism will favor creating lots of minds that have easily satisfied preferences, so that more of them can be created given a resource constraint. We can improve on this by noting that we value complex minds enjoying complex things. If a complex mind has more worth, then how do I evaluate a dyson sphere sized brain relative to my own utility?
We know that the kind of mind we value to have is complex which is a different thing than valuing it because it is complex. It doesn't strike me as intuitive that I would value a person that is maximally twisted up.
When i check my intuitions I seem to value simple minds less, and more complex minds more, robustly across the range of complexity in minds we observe. It does feel weird to try to imagine stretching this scale to include things more complex than me, but it feels weirder to make current humans the cutoff if that makes sense.
When I check for which minds I seem to appreciate among the minds we observe it seems those minds that have larger surface area are worth more. Extrapolating this is weird and it is unlikely that the human mind is the apex of surface area possible. But I am pretty sure that having a larger surface area would not be sufficient to make me care more. However it seems it would be more probable / there would be more resources to have something worthwhile with it, provided that it is not "wasted". I don't have a clear handle on what the "good" produced is but just having several acres of neurotissue around is not the finished stage.
It seems to me that there's a simpler argument for the same conclusion with the same premises. Consider any small population-changing intervention. If it's small then it's approximately reversible. Either it or its reverse will have positive rather than negative impact. It such interventions are cheap enough, the impact will still be positive after accounting for cost. If you got staggeringly unlucky and picked an intervention of near-zero net impact, pick a different one instead.
Of course all the premises are somewhat questionable, but that's true whichever formulation of the argument one uses.
It's not obvious to me why small interventions should be reversible - can you explain? The fact that lives of type X (e.g. people born in a particular place) are cheap to create and prevent doesn't mean it should be done independent of your population ethics: if you think X-type lives are neutral, it's not worth changing the number of them. It needs to be cheap to create/prevent at least two different types of lives which are clearly different in expected utility. That way, someone who thinks that one type is neutral will find the other type highly non-neutral and be in favor of changing its population size.
For example, even if it's cheap to change the number of people living on $500 per year, someone who thinks those lives are barely worth living wouldn't do it. But if it's also cheap to change the number living on $50,000 per year, then the same person would be in favor of increasing that number. The idea is that nobody should view both types of lives as neutral, since they are very different and most people think it's very good to improve an existing person's income by 100-fold.
Because it seems like it would be an awful coincidence if the current situation were right at the end of the range of the available possibilities. That would mean, e.g., that there's a small gap between where we are now and one more child being born in Bhutan, but a really big gap between where we are now and one fewer child being born in Bhutan.
That's by no means a watertight argument. It could be, e.g., that for some reason it's really easy to get people to have more children and really hard to get them to have fewer, or vice versa. But it seems really unlikely.
For the avoidance of doubt: I didn't think it does, nor did I think you think it does.
That's pretty much exactly what I meant by "If you got staggeringly unlucky ... pick a different one instead". My apologies if that was too cryptic.
Seems really likely to me. For instance, having more children is associated with poverty. It's a lot easier to make many people poor than to make many people rich.
For a plus charity, you could donate sperm. I've heard that the price for sperm is far below the market clearing price. I'm not sure what would cause that. According to Wikipedia, if you freeze your sperm yourself and sell it on the internet, you can make a lot more, which suggests that that's true. There's also places where you're not allowed to make money or the pay is strictly limited.
If the price really is below the market clearing price, then the market is limited by the number of donors, so donating enough sperm to have a child results in one more child being born. You're not just causing one fewer person to donate sperm, or that much more sperm to be wasted, or something like that.
I don't think this follows. If you think that Minus is better than Plus, it does not follow that Minus is better than Plus+Minus. Likewise, if you think that Plus is better than Minus, it does not follow that Plus is better than Plus+Minus. The key is that with certain population ethics, the change caused by Plus or Minus is context-dependent.
For instance, suppose you have average utilitarianism and utility is proportional to income per year. Also assume an existing population of two people each earning $500 in a poor country.
Minus causes one less person to be born in the poor country. This doesn't change the average, so with Minus you have spent $1000 for no change in utility at all.
Plus costs $6000 and causes there to now be 3 people. Overall utility is (5000 + 500 + 500)/3 = 2000
Minus + plus costs $7000 and causes there to be 2 people with overall utility of (5000 + 500)/2 = 2750
Plus has a cost per added utilon of 6000 / (2000 - 500) = 4. Minus+plus has a cost per added utilon of 7000 / (2750 - 500) = 3 + change
Therefore plus+minus is better than plus which is better than minus.
I assumed that the effect of the intervention is "small enough relative to the world" for your population ethics to be smooth. For average utilitarianism in particular, this corresponds to the percentage change in population being small. In your scenario, this isn't true, since there are only 2 people, but in the real world it holds up very well. Just 1% of the world population is 70 million people, and virtually no intervention (except for things like existential risk reduction) could cause such a large population change.
Yes, they did. In real world, "Plus" option means "one more person born in a middle-income country, in a poor and uneducated family". And even that is expensive.
That might be the most cost-effective Plus option, actually - if you crudely model the cost of one extra birth as proportional to the child's future income, then diminishing marginal utility of income means that it's better to promote births in poorer countries (up to a point). The optimal income level at which to do a Plus intervention (in terms of maximizing the cost-effectiveness of Plus+Minus) depends on the cost of preventing a birth in a poor country. If the cost is high, you'd want Plus to be in a richer country due to the "overhead" of the Minus intervention, but if Minus costs almost nothing, you'd want Plus to be in a country only slightly richer than the Minus country.