Post-S.C. Republican analysis

The race on the Republican side went about as I predicted in both Nevada and South Carolina on Saturday. The McCain-Huckabee margin was a bit closer for first and the Romney-Thompson margin was a bit closer for third, but the ranking was right and the margins were about right. In Nevada, similarly I was correct in predicting a wide margin win for Romney but I under-estimated how wide and rather than placing a clear second, McCain was lumped together with the others at 10% (+/- 3). Giuliani again amazed me in showing how soft his previous support had been (though his 4% of the vote was enough to win him his FIRST delegate to the convention).

The Nevada numbers are largely meaningless due to the blow out by Romney there. In South Carolina though, the exit polls tell an interesting story and paint us an interesting picture as we move on to Florida in the GOP race.

I predicted a margin of 37% for McCain (he won 33%), because I thought we would dominate among veterans and that there number would be, at least, as high as in the 2000 primary. However, veterans made up only 25% of the electorate (compared to 27% in 2000) and McCain got only 36% of the vote from them (compared with 48% in 2000, though it was only a two-way race then). His margin among veterans over Huckabee was only 36-29, a +4 improvement over his general score, a far cry from the +18 he had over Romney among veterans in Michigan.

Moreover, as in New Hampshire and Michigan, McCain's strength came from indpendents. Among registered Republicans, Huckabee won 32% to 31%. Among Indpendents, McCain won 42% to 25%. In Florida, the primary is closed and only Republicans will be allowed to vote.

My prelimary prediction therefore is that Huckabee will win the Florida primary. Why? Because there, unlike in other states, no only with McCain not be able to count on the support of independents, he will also have to compete with Giuliani for moderate Republicans.

Looking back to the Florida primary exit poll from 2000, we find that 58% of voters considered themselves either "somewhat conservative" or "very conservative". Among that group in South Carolina, McCain lost 35% to 26%. Moreover, Thompson is likely to drop out and (even though he may endorse McCain) his supporters are conservatives which will most likely go to Huckabee. Indeed, among the "very conservative" metric, Thompson placed second, ahead of both McCain and Romney.

In Florida, McCain will not only be battling Giuliani for moderates. Despite positioning himself as a "conservative" in contrast to his moderate-to-liberal record, Mitt Romney has done best among moderates in the previous contests.

If we assume that voting patterns by ideology hold up and that 60% of Thompson supporters go to Huckabee, 25% to McCain, 10% to Romney and 5% to Giuliani and assume that the ideological make up of the Florida primary will be the same as 2000, Huckabee would get 27% of the total vote in Florida will out gettting one vote from a moderate.

In 2000, Flordia Republicans broke down this way:
Very Liberal (VL) 2
Somewhat Liberal (SL) 6
Moderate (M) 34
Somewhat Conservative (SC) 38
Very Conservative (VC) 20

So Huckabee has essentially 58% of the voters to himself (more or less) while McCain, Giuliani and Romney will have to fight for the remaining 42% of voters.

This does not look promsing for anyone but Huckabee, especially if Thompson withdraws from the race.

Assuming Thompson is out, and omitting Paul, here is how I would paint Florida (subject to change):

Huckabee 33 (0VL + 0SL + 3M + 18SC + 12VC)
McCain 33 (0VL + 1SL + 15M + 12SC + 5VC)
Giuliani 18 (2VL + 5SL + 10M + 1SC + 0VC)
Romney 16 (0VL + 0SL + 6M + 7SC + 3VC)