Walter Shapiro thinks Republicans might want to change their rules to prevent the possibility of an ugly result: Sarah Palin winning lots of delegates in a multicandidate field with winner-take-all delegate rules, even in the event that most Republicans don't think she should be president.
I discussed in part 1 why she has no chance if she is only the choice of about 35% of Republicans and unpopular with the rest, but should the GOP try to change its rules just in case? We're talking about a Jesse Jackson type of candidacy: able to win a sizable percentage of the primary vote, but very unpopular with the rest of the party (for the nomination: it doesn't matter whether she would they would actively dislike her, or just act on their belief that she'd be a disaster for the party in November).
A 35% Palin won't win the nomination, and generally won't disrupt much of anything, given GOP rules. Yes, if we further assume that the 35% is a national average, and she would range from 20% to 45% depending on the state, it is true that she would capture a few winner-take-all states, and with them a handful of delegates. But that's all. However, under the Democratic Party rules, which mandate strict proportional representation, such a candidate would be far more disruptive. A 35% Palin would win 35% of the delegates. That's bad enough; by controlling a large chunk of the delegates at the convention, Palin would be in a better position to make demands on the nominee in return for good behavior than if she arrived with only a handful of delegates (see, at a lower level of support, Ron Paul in 2008).
Even worse, however, is the possibility of a true disaster: a deadlocked convention. If it was really true that Palin was seen as likely to win 35% of the delegates (but no more) in a system governed by proportional representation, the normal incentives in the system start to break down. The logic of winnowing has always been that there's no particularly strong incentive for a candidate to remain in the race after he or she can no longer win, because showing up at the convention with 10% or 20% of the delegates doesn't give you very much (that is, for candidates who are trying to win, as opposed to ideological candidates who just want to be heard). However, to the extent that the convention is likely to be deadlocked, the opposite logic applies: everyone will want to be a player at the convention, and the admission price will be delegates. As long as the frontrunner can be held under 50% of the delegates, then any candidate who controls a group of delegates should be able to get something valuable.
Now, granted, the party has a lot of ways to pressure losing candidates to drop out; odds are that in the event of Palin 35% and proportional representation rules, only one other candidate would survive South Carolina, and that candidate would wind up with well over 50% of the total delegates. But it's risky; if party insiders cannot choose between, say, Romney and Pawlenty, it is possible that both could be viable into Super Tuesday, and if they both remain, then others might decide to stay in as well, perhaps (like Huckabee '08) concentrating on a handful of states.
All in all, changing the rules would simply shift from one set of speculative headaches to another set of less speculative problems. And I should emphasize again that I don't really think the basic assumption, about Palin winning around 35% of the primary vote but being unacceptable to the rest of the party, is by any means inevitable, or even likely. Still, the Democrats showed the dangers of a PR delegate selection system with the Jackson candidacies in 1984 and 1988, and the Republicans would be foolish to ask for that result.