There is very good reason to think that the Republican candidate for president will get the party’s nomination on the first ballot. That reason is that this has happened at every Republican convention in every election cycle for the last 68 years. The last brokered Republican convention was in 1948. So the _outside view_ has to be that the probability that the Republican candidate for president will not get the party’s nomination on the first ballot is close to zero.
Well, not so fast. Here’s a point that was made by Nate Silver some time ago. Sure, there hasn’t been a brokered Republican convention in almost 7 decades, but conventions don’t happen very often, so the sample size of data is pretty small. Furthermore, there have been a number of near misses. Republicans had a close call in 1976 when the then sitting president Gerald Ford lacked enough delegates to secure the nomination before the convention but somehow managed to gain enough support at the last moment to edge out challenger Ronald Reagan. So if we start the clock at 1948, we’d have a 1/17 probability (round up to 6%) of a brokered convention as a baseline.
And there are some positive reasons for thinking it might happen this year. At the moment, there simply isn’t a clear frontrunner for the Republican nomination. ‘“It’s hard for me to see why a round of brokering in Cleveland isn’t the most likely outcome,” says Daniel Heninger, deputy editor of the Wall Street Journal’s editorial page. “None of these candidates looks likely to pull away and capture the majority of primary delegates before the party’s nominating convention in Cleveland next July.”' Fair enough, but there’s this kind of talk during almost every election cycle on one side or the other. A slightly more nuanced take is this: Trump appears to hover around 25%-35% in the primary polls. There’s some evidence that his ceiling is 1/3 of the Republican primary voters. If (and it’s a big “if”) that’s true, and if (here’s another big “if”) candidates like Cruz, Rubio, Bush, and Kasich manage to split the anti-Trump vote, then there’s a clear path to a failure to get a Republican nominee after a first round of voting at the convention.
All that said, media consumers like me need to be careful. Many journalists would love to be able to cover a brokered convention (and I would too if I were a journalist), so they’re likely to suffer from the positive valence effect. I need to be careful so that I don’t internalize this sort of bias. With all of this in mind I’d say that the current likelihood that the Republican candidate for president not win the party’s nomination on the first ballot is about twice the baseline, rounding down 1 point in the hopes of avoiding the positive valence effect: 11%. Of course, this number will need to be updated often as the convention gets nearer.