Ramaswamy and the physics of multi-candidate primaries

He's probably going to rise in the polls

I have a confession: I haven’t yet watched last night’s Republican debate. But I think I can make this prediction regardless. In fact, maybe it doesn’t matter that I haven’t watched, because I’m not sure my subjective impressions would help all that much — it’s not as though what I’m looking for in a presidential candidate maps very well to the median Iowa caucusgoer.

The prediction, stated with say 85% confidence, is that Vivek Ramaswamy will rise in the polls over the next month or so.

This prediction comes from some fairly basic mechanics that you’ll often see in multi-candidate primaries, predictable enough mechanics that they almost seem like gravitational forces:

1. Ramaswamy is not all that well-known and so has room to grow. In the recent Quinnipiac national poll, 51 percent of registered Republican voters hadn’t heard enough about him to form an opinion. The same was true of 41 percent of Iowa caucusgoers in the recent NBC/DMR/Selzer Iowa caucus poll.

2. Ramaswamy is rapidly becoming better known and last night’s debate will help with that. Here is Google search traffic over the past 24 hours among the leading non-Trump candidates. (Historically I’ve found Google search traffic to be a pretty reliable indicator of who will or won’t get a surge from a debate.) Ramaswamy got by far the biggest bounce in search traffic last night and it’s sustaining into this morning:

   

   Here’s the same data over the past 30 days, which doesn’t include the debate, but shows he was also getting more attention in the pre-debate period — surpassing Ron DeSantis as of 7-10 days ago. This won’t have been fully priced into his polling yet either.

   

3. It’s very hard not to gain ground in polls when you get a rapid increase in name recognition. Right now, Ramaswamy is at about 10 percent in national polls with 50 percent name recognition. So he has a 20 percent yield — of the voters who know him, 20 percent have him as their first choice. That’s actually not too bad — historically, early-stage primary polls are more predictive if adjusted for name recognition. Even if his yield were to fall to 15 percent (there’s no particular reason to think it will) but his name recognition rose to 80 percent, he’d move up to from 10 percent to 12 percent a first-choice candidate.

4. His polling rise will almost certainly beget more media attention, making this a self-perpetuating dynamic. Remember this from the 2012 Republican primary?

   

   Look at all those surge-and-collapses: Rick Perry, Newt Gingrich (twice!), Herman Cain, Rick Santorum, and to some degree Michele Bachmann. On average, they took 6-8 weeks to play out: three or four weeks on the upswing, then a roughly equally rapid descent — with the exception of Santorum, who actually sustained his momentum into the voting period and won a few states. This race isn’t exactly like 2012 — Trump is a much clearer frontrunner than Romney was — but it follows some of the same dynamics. The media is bored with Trump’s ostensible main challenger (DeSantis) underperforming, and the Democratic primary being noncompetitive. It will have a high propensity to chase shiny objects.

5. Most primary voters like multiple candidates, and that makes multi-candidate primaries intrinsically volatile. If you’re a liberal Democrat reading this, you probably wonder how voters like any of the Republican candidates. But put yourself in the mindset of a conservative Republican in Iowa. They’re like kids in the candy store: most of them like Trump and they also like several of the other candidates. In the aforementioned NBC/DMR/Selzer poll, for instance, the average Iowa caucus voter had a favorable impression of 4.2 GOP candidates. Voters can shift rapidly between the several choices they like; it doesn’t take that much to go from Snickers to M&Ms. So “momentum” and “vibes” can be self-fulfilling over the short run.

In nerdier terms, the object of multi-candidate primaries is to be the Schelling focal point and draw attention to yourself from voters who have several choices they like. Trump understood that in 2016 and Ramaswamy may understand it as well, given the number of media appearances he makes. You’d much rather get too much media attention than too little, even if some of it is negative.

Will this hypothesized Ramaswamy surge be sustained? I have no prediction about that, which would require me to do more “scouting” (i.e. rely more on my subjective opinion) than I want to do right now. But most surges aren’t — “expectations” rise, which means candidates are judged against a higher standard, and voters and the media go chasing the next shiny object.