For several weeks now, we’ve partnered with Good Judgment to conduct a forecasting tournament about the 2016 presidential election. One unique aspect of this tournament is that we’re asking forecasters not just the basic questions – such as who’ll win the nomination or the general election – but questions that don’t show up in conventional prediction markets.
For example, the tournament currently separates two key questions: Which Republican will have the most delegates after the Super Tuesday contests, and which will actually win the nomination? If there’s a difference in how people answer these questions, then it tells us the chance that Donald Trump might lose the nomination.
The folks at Good Judgment shared with me a subset of the data: the view of every forecaster (350 in all) who has answered both questions since the Super Tuesday question debuted Feb. 11.
Overall, these forecasters gave Trump a very high chance of winning the most delegates on Super Tuesday: The median answer is 85 percent. This fits with what the polls suggest and what most political observers expect.
But when asked about the Republican nomination, these same forecasters gave Trump only a 60 percent chance (again, the median). This difference suggests that there remains a real, albeit far from overwhelming, possibility that another candidate may win the nomination (these forecasters give Marco Rubio a 25 percent chance). This 60 percent figure is, interestingly, lower than what the prediction markets currently suggest (an 81 percent change for Trump as of this writing).
Part of the reason this difference exists is that these forecasters assign some probability – a median of 15 percent – to the possibility that no candidate wins enough delegates to win the first vote at the convention, which could lead to a nominee other than Trump. Usually these sorts of “contested” or “brokered” convention scenarios are hype or a fun parlor game. This year, the chance of such a scenario may be higher.
Here is another interesting question: How would a Trump nomination affect the Republican Party’s chances in November? The tournament gets at this directly by asking not only whether the Democrat or the Republican will win the general election, but who would win if Trump is the nominee.
Because not every forecaster thinks Trump will be the nominee, we might expect differences in how people answer each question – as long as they think Trump is different from a generic Republican.
Indeed they do. And not in a way that should make Republicans happy. If we look at all forecasts made for the general election, the Democrats have a 64 percent chance (similar to prediction markets). But when Trump is stipulated as the Republican nominee, the Democrats’ chances jump to 75 percent. If we look only at those forecasters who have made predictions for both questions – a true apples-to-apples comparison – the difference is smaller (67 percent vs. 75 percent) but still notable.
In short, the current message of the forecasters at Good Judgment is: Trump will hurt the party in November, but there is some chance that anti-Trump Republicans could derail his nomination.