Thinking about polls

I think Mike Huckabee has a low probability of winning the general election in November. But that does not significantly differentiate him from the rest of the Republican field in my opinion. I think the entire field has a fairly low probability of winning the general election next year due to a combination of an internally divided along multiple axises Republican Party, President Bush as an anchor, rhetorical, policy and political as well low levels of enthusiasism and comparative fundraising success compared to Democratic candidates. Finally the stands a Democrat needs to be popular among the Democratic primary electorate universe are either fairly popular in general, or rather benign. However the stands Republicans need to make to win their primary nomination battle are significantly divergent from the stands they need to take if they are to win.

So when the CNN poll came out yesterday there was widespread analysis that this sinks his chance at the GOP nomination as he is beaten by all Democratic challengers. I disagree that this is the death knell of the Huckabee campaign on electability grounds for a couple of reasons.

The first is that electability so far has not played a determining role in either campaign this cycle. The following chart shows the CNN spreads in twelve match-ups.



McCain is the most 'electable' candidate, and John Edwards is the most 'electable' Democratic candidate according to these head to heads. Yet McCain is charitably in 4th place nationwide and in his firewall state of New Hampshire, is in 3rd place with a small chance of getting a net positive post-Iowa bounce. Edwards is in 3rd place despite being the most electable and liberal of the top three Democrats. His path towards winning is far clearer and has fewer dependencies than McCain but it is not a straight forward path to the nomination that Clinton has, or the double dependency path that Obama has.

Huckabee really is not performing much worse at this point in time in the head to heads than Romney, and only a couple points behind Guiliani. Both Guiliani and Romney have significant name recognition advantages at this time over Huckabee, but if Huckabee was to win the nomination, the first quarter billion dollars of advertising will bring his name recognition up to 95%. Once name recognition equalizes, soft Republicans who identify as independents or non-affiliated but vote 90% GOP will vote for their party's candidate. Nothing surprising there.

But getting beyond the horserace, one has to wonder what the objective of the Republican Party and its primary stakeholders is for this election cycle. It is analytically absurd to assign singular motives and objectives to an combined, collective decision making body such as a political party, but it is convienent shorthand.

Right now it looks like 2008 will be a very tough year for Republican candidates. Assuming that we don't get two straight 'freak' elections, I think it is safe to assume that some GOP challengers will beat Democratic incumbents somewhere (Georgia is the best bet), but overall, the betting line is for the Democrats to win expanded majorities in the House and Senate as well as the White House.

If that is the case, then an objective of winning the intra-party faction fight which has been simmering for a while is a high value and obtainable objective. The nominee is the de facto party leader and the coalition behind the nominee is the de facto flagship coalition within the party. Huckabee represents a significant faction within the GOP, and it is one that has threatened to bolt in order to demonstrate its own importance.

If a significant proportion of Republican decision and opinion leaders, and more specifically theo-/social conservative opinion leaders believe that this is a losing year for the general election nominee irregardless of who the nominee is, then the incentive to win the intra-party fight and support Huckabee no matter what becomes far stronger which will ensure that he stays in for a while.