Read on Twitter
I was explaining the history of catastrophe (cat) reinsurance to a new hire. Then I looked at Austin and saw similarities. A 
...
In 1992 hurricane Andrew bombed Florida. At the time it was the most costly natural disaster in US history. https://en.wikipedia.org/wiki/Effects_of_Hurricane_Andrew_in_Florida
Change followed
...In 1992 hurricane Andrew bombed Florida. At the time it was the most costly natural disaster in US history. https://en.wikipedia.org/wiki/Effects_of_Hurricane_Andrew_in_Florida
Change followed
Insurance companies needed financial protection from tail events that were less than 1:100 probability... but (re)insurance models were largely historically based: fit a distribution to historical events. As such, nobody knew if the “bad tail” of the model was close to reality.
That’s a tough “out of sample” problem. In came modeling companies like RMS & AIR who built engineering based models and could run historic hurricanes through different paths & estimate damage. And they could estimate frequency of storms, and simulate. So that produced...
stochastic catalogs of hypothetical, but realistic, events. And risk bearing companies began to price (re)insurance based on probability of loss from these models. This was very different from actuarial pricing models of the time. So how’s this relate to Austin? Welp...
Current weather in TX is “out of sample”... and not just insurance pricing but also city infrastructure is based largely on historical events. There are many reasons to believe the underlying loss generating function (weather patterns) are not remaining stable (climate change)
So cities and businesses will look to models to understand the future possible outcomes. Some will not believe the future is different from the past. They will experience learning opportunities that result in human suffering and death. Some will mitigate risk through financial
protection, or physical infrastructure, or other means. And I expect the rest of my life will be spent watching organizational wrestling matches as we come to grips with “the future ain’t what it used to be”
Lots of opportunities for modeling... but also lots of frustration ahead
Lots of opportunities for modeling... but also lots of frustration ahead
addendum: I use “(re)insurance” above to mean “insurance and/or reinsurance”... reinsurance is financial protection that insurance companies buy to protect them from catastrophic losses. So it’s insurance for insurance companies. I work for a global reinsurance company, fwiw.
Here’s my employer’s take:
“With each additional record breaking hurricane, typhoon, flood, and fire the evidence continues to mount that we live in a world where climate change is influencing the frequency and severity of catastrophes” https://www.reinsurancene.ws/market-is-missing-the-point-on-climate-change-renre-ceo-odonnell/
“With each additional record breaking hurricane, typhoon, flood, and fire the evidence continues to mount that we live in a world where climate change is influencing the frequency and severity of catastrophes” https://www.reinsurancene.ws/market-is-missing-the-point-on-climate-change-renre-ceo-odonnell/
I’m obviously not representing my employer, because I don’t work where I shitpost. But I enjoy talking/thinking/bitching about analytical modeling on this platform. And I think what’s happening in TX is part of a bigger analytical discussion we should be thinking about.
