Election Predictions of Political Pros: Clinton vs Trump
I predicted back in 2013 Clinton would run. And if she ran she’d be the nominee and subsequently she’d be the next president.
With 6 months to go, I wanted to offer a look at the odds of that prediction being accurate.
The media paints a picture of a close race. BS. They are motivated to do that because it sells advertising and it’s what people want to believe/feel.
When you look at people who’ve built their reputation on predicting elections (not selling ads), you get a different picture. Back in 2012, I used this discrepancy between media/public feeling/perception and pro perception to make 54% on a 10k bet in 3 days.
Prediction Pros. These aren’t talking heads on TV. They’re political data junkies. They don’t just look at some polls, they look at ALL polls. They have formulas to adjust polls which polls are skewed to favor R’s or D’s. They look at the state polls, county polls, economic data, likely turn out models, key demographics inside key counties inside key swing states. They don’t compete to call who will win an election – boring, too easy, and they’d nearly all being agreeing with each other – they compete on the exact margin on election day campaigns will win by.
Here are the people I’ve found to be the most credible/reliable calling who will win and by how much:
538: 353 Clinton vs 184 Trump. Inside the political prediction world, Nate Silvers has achieved celebrity-like status in 2008. He just released his prediction on 7/1/16.
Sabato: 347 Clinton vs 191 Trump. A professor of PoliSci at University of Virginia, he puts out a free newsletter (I recommend) and calls everything from Presidential races to Governorships to individual Senate and (much more difficult) House races. His prediction as of 6/23/16 (link is live updated).
ElectionProjection.com: 349 Clinton vs 189 Trump. This is a normal guy who blogs but is obsessed with guessing political outcomes accurately. I began following him back in 2004 when he started and he’s a damn good guesser. His prediction as of 7/1/16 (link is live updated):
Princeton Consortium: 330 Clinton vs 208 Trump. This is a highly mathematically based model that the closer you get to election day, the more certainty their model has. PC’s prediction as of 7/1/16:
Outcome based analysis. Every pro is calling a Clinton electorate vote blow out and they differ (not by much) in calling that margin of victory.
Probabilities. It’s good to think of a prediction as the most likely outcome. That said, a more nuanced perspective looks at probabilities to get a sense of likelihood of an outcome. When it comes to betting, upsets happen. The predicts outcomes above are black and white. Here is the grey:
Probabilities of Clinton Winning (taken on 7/1/16):
73% Vegas market makers.
81% FiveThirtyEight’s probability model.
Summary. 6 months before the election, all pros are seeing a likely Clinton blow out (measured by Electorate Votes). While Vegas-like betting markets are putting odds, around 70-ish%, most political pros are 80-ish%.
Betting. When I made my last bet, there was a significant market inefficiency. Currently, the market inefficiency is (=83/73) ~14%. Meh, margin but not enough to make the risk/reward worth it. Last betting cycle the inefficiency was (=53/14) nearly 4x greater.
Jared Fielding 