Friday, October 7, 2011
What is a good rate? What good is a rate?
Windstorm insurance rates in Florida (well, and other rates in FL and elsewhere) are created through a scientific, economic, and political decision process. This is somewhat exemplified in the prior Goldilocks posting. The end result of this process is a negotiated agreement on the characterization of the hurricane risk for the next year. The hurricane risk, socially constructed, is understood and communicated through a number which, in effect, is a prediction or prediction or forecast or what have you about the upcoming year's windstorm maximum possible loss.
So, how do we know if this numerical prediction is any good?
Well, in weather and climate, predictions goodness is judged in part by its skill and value. Something similar may be done here. I think. Maybe skill can be judged by how well premiums collected one year compares to that year's losses. A problem with this methodology however is that a) premiums include way more then what is paid for the risk itself and b) tracking annual rates is seemingly near impossible because that data is not readily available. Instead, maybe skill of risk characterization can be judged by the ability to maintain insurability of the risk.
As for value (to society), this can be judged by the affordability of the rate. Therefore judging a rate goodness may be assessed by considering a 2x2 matrix (which I have adapted from others' work on climate forecasts). I've filled it in partially by some initial case study guesses. Alas, I am still on the hunt for a rate that is both affordable and insurable. Ha!
Posted by Jessica Weinkle