A number of journalists -- Reid Wilson at The Washington Post, Adam Nagourney at The New York Times -- have concluded that we could reduce government dysfunction if the U.S. and the states adopted California's system of non-partisan or "jungle" primaries, in which all candidates run against one another regardless of party, and the top two candidates advance to a runoff, even if they're in the same party. Political scientists -- John Sides, Seth Masket -- don't buy this. Masket has written:
Allow me to link here to the paper I did with Nolan McCarty, Eric McGhee, Steve Rogers, and Boris Shor ... showing that variations in the openness of primaries seem to have no relationship to the moderation or extremism of a state's elected officials.... The people that show up in primaries still tend to be very partisan voters, and parties are pretty skilled at advantaging the candidates they like even in difficult circumstances....I'm no expert, so I'll limit myself to a bit of speculation about recent Senate elections.
In theory, the openness of the top-two system encourages the entry of moderate candidates. But ideologically extreme ones can end up doing quite well, for the simple fact that voters have a very hard time distinguishing between ideologically extreme and moderate candidates....
All of this has produced a situation in which California's legislators are no more moderate, relative to their districts, under the top-two regime than they were before. Indeed, they may even be a bit more polarized.
Yes, I think it's possible that we might have been spared Ted Cruz in the 2012 Texas Senate race if Texas had had a jungle primary -- but it's not at all certain. Both Cruz and David Dewhurst received more votes (in both their primary and the subsequent GOP runoff) than Democratic primary winner Paul Sadler received in his party's primary -- so it seems possible that Dewhurst and Cruz would have advanced to the general election if there'd been a jungle primary. Cruz beat Dewhurst handily (57%-43%) in the GOP runoff. Would Dewhurst have picked up enough Democratic and independent votes to beat Cruz if they'd faced each other in a general election? Hard to say. And when I see that Dewhurst is now running for reelection as lieutenant governor by calling for President Obama's impeachment, I think there'd have been a high level of crazy no matter which of these guys won.
It's possible that Charlie Crist, who lost a three-way race to Marco Rubio in 2010 while running as an independent, could have beaten Rubio in a two-man race -- if there'd been a jungle primary, the two of them certainly would have advanced to the general election, beating Democrat Kendrick Meek. I'm amusing that Crist could have beaten Rubio one on one because if you add the vote totals of Meek and Crist, they exceed Rubio's by a tiny amount. But in a jungle primary system, Crist wouldn't have had to leave the GOP to run against Rubio; if he'd won, he'd have won as a member of the GOP, and he'd be serving as a Republican today.
Harry Reid won reelection in 2010 because he faced crazy Sharron Angle. Maybe, if she'd had to run in a primary that involved all the candidates, she'd have lost and the top two would have been Reid and less-crazy Republican Sue Lowden. Polls suggest that Lowden might have won that contest.
The same goes for Claire McCaskill in Missouri in 2012 -- if there'd been a jungle primary, the moderating influence of Republican and independent voters might have put her in a general election fight with John Brunner rather than the now-notorious Todd Akin. Polls suggest that he would have lost that contest.
And we can pretty much assume that a jungle primary in Delaware in 2010 would have picked moderate Republican Mike Castle for one of the final slots, rather than crazy Christine O'Donnell. If Democrat Chris Coons had been running against Castle, polls show that the Republican would have won handily.
So my guess is that jungle primaries would give us somewhat fewer crazy Republican senators -- but possibly more Republican senators.