Mar 12, 2016

Another prediction market oddity

Over the past month I've blossomed from toe-dipping prediction market novice to hot-blooded prediction market day trader.

Today during my daily PredictIt trawl, I encountered another strangely priced market dyad (strangely priced dyads have previously proven quite profitable).

"Will Hillary Clinton win the Democratic nomination?" was trading at $0.80. Meanwhile, "Will a federal criminal charge be filed against Hillary Clinton in 2016?" was trading at $0.30.

So Hillary has an 80% chance of winning the Democratic nomination while simultaneously having a 30% of being indicted? Granted, these outcomes aren't mutually exclusive – there exist scenarios where Hillary secures the nomination only to have charges filed against her later in the year. Naïvely, that would be surprising – bringing charges against a party frontrunner would be a big kerfuffle, but not as big as bringing charges against a party nominee. Maybe FBI Director James Comey (who is "very close" to the Clinton investigation) wants to blow up the 2016 presidential election by indicting Clinton in the fall, but I'm going to go ahead and put scenarios like that in the "highly unlikely" bucket.

My intuition is that there's a misvaluation here. Either p(Clinton wins Democratic nomination) = 80% is too high, or p(Clinton is indicted) = 30% is too high. Which is it?

Let's take p(indictment) first. If we're considering the conjunction of winning the nomination and being indicted as highly unlikely (say, p(conjunction) < 0.05), then we're mainly thinking about the likelihood of Clinton being indicted before she secures the nomination. Is there a 25% chance of that?

Given that David Petraeus received a light punishment for more serious but similar violation, setting a lenient precedent, Clinton's apparent unconcern, and the fact that using personal email for government business is fairly common, I would be surprised if Clinton was indicted. I think it's unlikely. Less than 25% likely? Probably.

Considering the alternative, p(wins nomination). 80% seems pretty on point, if not low. Clinton could fail to secure the nomination either by losing to Sanders or dropping out. Despite her defeat in Michigan, she still is on track to defeat Sanders, and I think there is only a small chance (say < 5%) of her dropping out.

Conclusion: "Will Clinton be indicted in 2016?" is currently being overvalued by the market. I picked up some "No" shares.

[rereads: 1, edits: phrasing tweaks]