Fat tails thinking and non-thinking

Disasters are bad no matter how you look at them. And then there are disasters that boggle the minds of even those who are hardened in the art of disaster management.

They aren't just a bit worse than anything that came before, they are twice as bad.

They don't burn a few more acres than the last wildfire, they burn twice as many.

They don't just take the lives of a few more people than the last terrorist attack, they take the lives of twice as many.

 
Second Largest
Largest


Wildfires, 1988-99
Nevada (1999)
Yellowstone (1988)


Acres burned
288,200
1,585,000


Earthquakes
Pakistan (2005)
Sumatra (2005)


People killed
80,361
283,106


U.S. Terrorist Attacks
Oklahoma City (1995)
9/11 (2001)


People killed
168
2,974

 

That relationship seems to hold across the board, and the table isn't cherry-picking. The worst earthquake in history, in Shaanxi on January 23, 1556, claimed over 800,000 lives—much worse than Haiti, Pakistan, and Sumatra combined.

Disaster damages don't follow nice little bell-shaped distributions we all study in high school. The tails of the distributions are much wider than others, "fat" in statistical jargon.

What to do?

Well, politicians have the tendency to hope that something bad doesn't happen on their watch, and chances are that it won't. The problem for the rest of us is that often it's just a matter of time. The goal isn't to survive a four-year term. The goal is to survive. It isn't very comforting that an ever warming atmosphere will load the dice to make precisely these extreme events more likely.