What is liable to happen if, as expected, Charlie Crist defects from the Republican primary to run as an independent?
I'm going to rely a bit more than usual on my intuition. The numbers, frankly, aren't much help anyway, since they show anything from a 17-point Rubio lead (Rasmussen) to a 2-point Crist advantage -- and all of this polling has heretofore been based on a hypothetical. But here's my best guess.
I think that Crist is going to have a great deal of trouble holding onto his remaining Republican support. Right now, Crist is preferred to Rubio by about one-third of Republian registered voters, according to Quinnipiac. But, even though there are some moderate Republicans in Florida (38 percent of the Republican presidential primary electorate in 2008 described themselves as moderate or liberal), it's asking a lot of a Republican identifier to continue to support a candidate who has just ditched the party -- particularly when it seems to be motivated by expediency rather than ideological resolve (Crist's lack of grivatas, even when compared with someone like Arlen Specter or Joe Lieberman, will hurt him here). Plus, Rubio, already very well liked by conservatives, will get to do a bit of a victory lap as the sole viable candidate on the primary ballot, and can immediately turn his attention to broadening his appeal among party moderates. This might not even require all that much effort; Rubio's favorability rating among Republicans is 67-7 (!) according to Quinnipiac, and while quite conservative, he seems from a distance to avoid coming across as any kind of dangerous extremist.
If you play around with the "what-if calculator" that Pollster.com's Mark Blumenthal created, and give Rubio, say, 75-80 percent of the Republican vote along with a bare minimum of 25 percent of the independent vote and a few percentage points of Democratic support, he seems to have a floor of about 36 percent of the vote overall, whatever reasonable assumptions you might make about turnout. Indeed, given the mood of the electorate, I would be very surprised if Rubio wound up with less than 36 or 37 percent of the vote, unless there were some major scandal (always possible) or he ran a terrible campaign (also possible).
In a three-way race, what this means is that Rubio is guaranteed at least second place. He could still lose if the race were Crist 37, Rubio 36, Meek 27, or Meek 37, Rubio 36, Crist 27, but you can't get more than one-third of the vote in a three-way race and do worse than second.
This is advantageous, since it means that nobody much will defect from you for strategic reasons. Say, for instance, that your order of candidate preference were Rubio, Crist, Meek. If the polling at some point in late October were Meek 40, Crist 35, Rubio 25, a voter might consider defecting from Rubio to Crist to prevent Meek from being elected, thinking that his vote would otherwise be wasted. But this doesn't come into play if it's Meek 40, Rubio 35, Crist 25 -- then switching to Crist only helps to ensure Meek's election. Nor does it come into play if the polling is Crist 40, Rubio 35, Meek 25 -- then Meek isn't really in the running and you'll simply vote your first prefernce (Rubio) over your second (Crist).
By the same logic, third place is a dangerous place to be in a three-way race: then you can get into a vicious cycle where voters ditch you because they're afraid of wasting their vote, which in turn makes your problems worse and worse. This is precisely why support for third-party candidates often does tend to collapse at the end of an election if the candidate looks as though he's not going to be able to get over the hump.
In the near-to-medium term, if Crist wants to avoid falling into third place, he probably needs to start appealing to Democrats and Democratic-leaning independents almost immediately. There's a little bit more breathing room over on that side of the aisle because Kendrick Meek is relatively unknown by the electorate: just 26 percent of the electorate have an opinion of him, according to Quinnipiac, versus 57 percent for Rubio (and 84 percent for Crist). But Meek (whom -- full disclosure -- I met for breakfast a couple of months ago) is a humble and likable guy who is liable to grow on people. There are plenty of people who won't like his politics, and -- yes -- a very few people at the margins who won't like him because of his race. But Crist will need a fairly broad amount of support from Democrats and Democratic leaners in order to have a shot in a three-way race in which Rubio will almost certainly finish in at least the mid-high 30s, and his best way to achieve that is to prevent Meek from getting traction in the first place. Having denied Meek viability, he could then tact back toward the center or the center-right come the fall.
But, this isn't easy: the more Crist thrusts to the left after having being reborn as an independent, the more flip-floppy he'll look. It's quite a needle to thread and it's possible that he's simply waited too long to make this move.
We probably haven't seen the last of the fireworks here. At some point between now and November, it's possible that Rubio will go through a period of negative media attention which could give Crist an opportunity to re-gain momentum. With highly competitive races in the state for both senate and governor, featuring female, black and Latino candidates, turnout is liable to be quite high. Lastly, there are wild cards that nobody wants to bring up and which could impact the race in various ways.
But Marco Rubio certainly looks like the favorite here, and I'm not sure that Crist is particularly more likely to pull off the upset than Kendrick Meek. Going indie is likely the right move for Crist, who had no shot to win as a Republican and probably had little future in the party. The upside if he's successful, moreover, is quite high. But the more I look at this race, the more treacherous his path seems to be.