13 Jun 2012

The mathematics of disaster

Someone at a conference made this wise observation:
The probability of disaster is often VERY small (~0) while its damage can be VERY big (~infinity).

These numbers, multiplied together, result in "undefined" expected damages from a disaster, which leads to a political deadlock when discussing the right policy for trying to prevent the disaster.

Apply this scenario to climate change, tsunamis in Japan and Indonesia, big failures in the water business, etc. to see why we were "surprised."
The best policy in these circumstances (besides risk-averse over-investments in physical defenses) is to put fluctuating prices on as many dimensions of disaster as possible, via insurance, water prices, carbon taxes, etc. Fluctuating prices make it easier for people to learn how to manage "everyday" fluctuations and thus prepare for BIG ones (cf. stockmarkets).

Bottom Line: Continuous information and price incentives makes it easier for people to prepare and manage essential services than occasional bureaucratic interventions (regulations) or rescues (bail outs) that are often too little, too late.