3 ms·
this idea is bad because some actors can potentially earn more money from the manipulation towards a bad policy than they will lose as a result of that policy b
by bronzeage 8y ago
this idea is bad because some actors can potentially earn more money from the manipulation towards a bad policy than they will lose as a result of that policy being bad.
the whole reason corruption exists is that the punishment and loss potential from corruption is too small compared to the huge benefit that comes from affecting policy makers.
the worst policies will have the highest bets because they will always be a small price compared to the benefits from those policies.
all the companies will vote for obvious bad policies which will reduce their competition to zero, because they can afford to throw that money in the trash to keep their monopolies in place
- zack_amoveo 8y agoLets consider a game. A coin will be flipped, and we can bet as much money as we want that either heads or tails will win. You will win $100 if the price in this market stays >75% chance of heads for 7 days continuously. So you keep buying up shares of heads, moving the price from 50 -> 60 -> 70 -> 80. by the end you are betting $4 for the chance to win $1. I know that the result is 50-50, so if I bet $1 against you, and have the chance to win $4, then that means my expected profit of taking this bet is %200. Which makes this one of the best investments available on earth. Which means many people will want to participate. Every bet you make at the price 80-20 costs you $1, but is only worth $0.40. you are losing 60% of however much money you bet. So if you bet $200, then your expected loss would be $120. which is even bigger than the $100 you could potentially win by manipulating the market. It doesn't matter if I bribe a manipulator with $100, or with $1 billion. The same logic holds. Having a vested interest in the outcome of a decision does not give you any ability to manipulate us into making that decision. http://mason.gmu.edu/~rhanson/biashelp.pdf http://mason.gmu.edu/~rhanson/biashelp.pdf