Wednesday, February 20, 2013

Cheaper than asteroids!

This week, while everybody was watching videos of a meteor in Russia, a larger asteroid passed by the earth. It might hit us next time it comes around.

Space agencies track the risk from things like asteroids, but it's unclear how much to spend on this sort of thing. The Planetary Society writes:
Near-Earth object surveys have found (we think) 98% of the largest objects that present the most risk, reducing the actuarial risk due to asteroid impacts from 250 fatalities per year to 64 per year. Based on past discovery rates and projecting forward through proposed future projects, over the next 16 years, we should achieve 90% completion of discovery of asteroids larger than 140 meters in diameter. The effect of this 16 years of work -- at a cost of roughly a billion dollars -- will be to reduce the actuarial risk to 33 fatalities per year. If you see asteroid surveys as a form of insurance, then you're spending about two million dollars per fatality avoided.
Is $2 million per life a good price?  It's repellent to even think about putting dollar values on lives, but we do it all the time.  If you buy the cheaper car instead of the safer, more expensive one, for example, you're trading off money and safety.

US government agencies define the “value of a statistical life” somewhere around $7 million. If they're deciding whether to institute new safety regulations on seat belts or air pollution, for example, they want to know whether spending billions of dollars is worth it. “Worth it”, for American lives, is around $7 million.

That's a pretty arbitrary number, and it isn't the same for all lives. What's the dollar value of a life in Haiti or Cambodia? I don't know, but I know the US government sure wouldn't spend $7 million to save one.

GiveWell currently estimates that its top-rated charity, the Against Malaria Foundation, saves lives for about $2,300. That's a pretty great deal, considering. In my whole life, I don't expect to earn $7 million. But I do expect to save hundreds of lives by donating to cost-effective charities.

The nice thing is that economies are not zero-sum. It's not just a question of shuffling money around; sometimes there are win-win solutions.  Some changes (like immigration) create more well-being for everyone, and we should aim for those.

But in the meantime, it's nice to know that you can save people's lives for a lot less than it costs NASA.


  1. In fairness, the number of potential lives lost to asteroids is dominated by the tail of civilization-destroying events (if one has no time preferences whatsoever, and thinks a stable long-lived civilization is plausible, and isn't convinced by the simulation argument).

    Of course, on that analysis, the impact of AMF is also dominated by similar effects. But in that case the sign is less clear, and I have not seen a compelling case that its a cost-effective alternative (back of the envelope calculations suggest it isn't competitive).

    So the real issue is not about the price of a life, it is about your expectations and preference regarding the future.

    (I don't know to what extent people who support asteroid surveys have this motivation in mind. I think most effective altruists who care about asteroids do have this motivation. Personally I don't think it's a cost-effective way to reduce catastrophic risk.)

    But I'm just picking nits. I agree with the spirit of the post, despite disagreeing with all of the particular recommendations :)

    1. I figured someone would call me on my failure to do justice to the asteroid issue. :)

      It seems to me AMF is better in the short term, and preventing things that might destroy civilization is better if you think we're likely to have civilization for a while. What's the argument for AMF having a negative effect under the conditions you named?

    2. Not a negative effect in the long-run, just an unclear effect. I would say that AMF is better in the short run and roughly neutral in the long run, and more explicitly forward-looking things are probably better in the long run and roughly neutral in the short run.