Prediction markets for politics are silly, for SCOTUS nominees even sillier. They are dominated by the uninformed, insiders, and hedgers.
I'm referring to markets like this, which is allowing you to bet on who Trump will nominate to replace Anthony Kennedy. Right now, punters give Kavanaugh an 80% chance, and Merrick Garland (Obama's old pick) a 1% chance.
predictit.org/Market/3232/Wh…
predictit.org/Market/3232/Wh…
In theory, such markets are supposed to "work", giving us the best measure of likely outcomes. Instead of making predictions, experts can put their money where their mouth is. Insiders who wouldn't leak to the press may nonetheless buy bets on such markets.
In practice, this doesn't happen. To begin with, the majority of those making bets are the uninformed. This noise swamps any signal. Even if you are an insider knowing Trump has all but decided on X (95%), you still can't afford to bet this amount to move the market to 95%.
In practice, insiders would not simply bet on the markets, but also manipulate them. If Trump has already decided on Kavanaugh, you are only gong to make 20 cents on the dollar for this bet. Therefore, you leak to the press it's some other choice.
If, as an insider, you can convince Politico to run a story tonight that it's Hardiman, you can easily double your money. You can even buy Hardiman bets first, leak Hardiman to the press, then sell that position and buy Kavanaugh, tripling your money.
Finally, there is also the hedgers. Certain politicians, if chosen, will likely have costs/benefits for your business. Therefore, you can hedge that result in prediction markets.
For example, Trump was certain to lower the corporate tax rate, so therefore you'd buy Hillary shares. If she won, you'd pay higher taxes, but win money on her victory. You'd lose that bet money if Trump won, but win on taxes.
The point is that such betting markets have problems. Rather than capture the "wisdom of the markets", they are screwy things that can't be trusted.
