Posted by Daniel Hall on June 23, 2008
This discussion about near-Earth objects reminded me of a point I heard Robert Pindyck make recently. It will probably sound obvious to all you Bayesians out there, but I don’t think I had heard it expressed so succinctly before.
The point was that in a world with multiple catastrophic risks, our concern about any one individual risk should not be measured by the independent probability of that catastrophe. Rather, we’ve got to discount each of the individual risks by the probability that one of the other catastrophes will wipe us out first.
Thus if we are very likely to be killed off by avian flu in the next century we should be much less worried about climate change.
I noted previously that critics of Easterbrook’s article like this guy want to dismiss asteroid impacts out of hand and focus on climate change. The irony in this discussion is that if we had a reliable method of diverting near-Earth objects — something that seems likely to be quite cheap — we would then be more worried about climate change, not less.