Educated Guesswork

If one were to have foresight, one would have gone the route of developing an early warning system for tsunamis. That would not only have saved a lot of money in relief efforts, it would have gotten rid of all those annoying costs in terms of human life.

Other disaster prevention is similar. The costs of cataloguing all the bits of rock in our solar system which would cause a major disaster if they hit the pacific ocean and figuring out whether they're on a collision course with earth would be a paltry few million, clearly justified economically even with the extremely small chance of such a collision being imminent, but it isn't happening because, well, the powers that be are simply not interested in preparing for disaster.

One thing which stands out on that list is that pandemics and famine blow everything else away in terms of numbers of lives cost, and that's a trend which hasn't changed and isn't about to. AIDS deaths in africa are tremendous, and going to stay that way. Such deaths can't be insured against, because we already know they're going to happen. It would of course be not terribly expensive to reduce the problem considerably - mostly by distributing clean needles and educating the locals about germs and basic hygeine - but that isn't going to happen either, because there's very little interest in keeping people in Africa from dying, and a lot of interest in trying to tell them to not have sex.

Fair enough... but note that we're going to spend this disaster aid money anyway... I'm just trying to work out a better way to do it...

Actually, no, serving the interests of the victims isn't the point--at least, not the entire point. After all, there is easily enough "disaster relief" to be done in the world to soak up the entire GDPs of all the donor countries, should they simply want to "serve the interests of the victims" tout court. Rather, the donors want to (a) decide how much they'd like to donate, based on their own extremely complicated, subjective and largely arbitrary criteria, and then (b) allocate those donations according to a slightly less arbitrary set of criteria that includes the whimsical preferences of the donors, as well as some slightly-better-defined goals like helping the most needy first and doing the most good per dollar spent. It's not clear that the current system--a permanent relief infrastructure that receives both a steady stream of funding and occasional floods of directed donations, and allocates them both as per this complicated collection of criteria--is not optimal, given these constraints.

I've been thinking about this, and there is an important countervailing point: moral hazard.

One of the biggest problems with insurance is that it changes the incentives of the insuree to maximize his consumption. That is, such insurance reduces the likelihood that insured countries will act to reduce the probability/impact of these disasters. A lot.