Come to think of it, that “Efficient markets in doom“ title could also apply to another thing I’ve been wondering about.
There’s this idea that Robin Hanson among others has pushed about how we should decide absolutely everything based on prediction markets.(Futarchy is perhaps a poor choice of names) There’s a few problems I have this this, but there’s one in particular that could be killer.
There’s this apocryphal story I will try to recount that shows this point. (Maybe from Barry Ritholtz?) Basically, there was trader who was on a stock exchange floor when somehow the floor started to believe that a US-nuclear exchange was coming. One of the traders there started selling everything, figuring that none of these assets would be worth anything. His boss came by to him and said “What the heck are you doing?” and the the trader said ”Haven’t you heard what’s going to happen?“
The trader’s boss then tells him “Look, if the world ends, neither the trades nor the money will actually matter. Always trade as if the world is going to continue.“
So, the problem with prediction markets is right there. For businesses, organizations, and nations, there’s always the risk of terminal failure. If they end, nobody’s going to pay your winnings. That means that catastrophic risks will always be poorly represented.
(I haven’t read all the prediction market literature, so I’m wondering if Hanson has an answer to this.)
I think Hanson’s aware of this problem – otherwise he and Eliezer would’ve bet on their AI FOOM debate. The literal end of the world is not represented in the outcome space, but that doesn’t mean what is represented isn’t useful.
I mean this is a problem with even something like a Trump victory, if you assume that he’d manage to destroy the value of the dollar.
I believe the world will end in a month. You don’t. Therefore I propose the following trade: You will give me a loan of a thousand dollars (which I will spend on hedonistic pleasures over the coming month) and I will pay you back at whatever interest we agree on. This is a very attractive proposition to you, given your beliefs, because I will be willing to agree to a very large interest on a loan I believe we’ll both be too dead to care about. This is a very attractive proposition to me, given my beliefs, because I’ll get $1000 free money I never have to pay back.