A commenter (yes, I get those!) asked who I was referring to when I talked about the "GOP establishment" -- those people whom I predict are going to crush Huckabee.
A lot of this comes out of the forthcoming book that I often mention, The Party Decides, by Cohen, Karol, Noel, and Zaller (you can read the manuscript in its current form here). The big finding in that book is that between 1980 and 2000, key party insiders - defined as governors (who endorse outside their state), senators, members of Congress, interest group leaders, major donors, opinion journalists, and a few others - have picked favorite candidates for the presidential nominations. They reward those candidates with endorsements and help channel funding their way. These endorsements (properly weighted) end up doing a much better job predicting who will win a party's nomination that polls, money, or momentum, whatever the hell that is. Indeed, the insider's candidate always got the nomination during this time period.
These authors just put out a new paper about the current cycle, which I mentioned previously. Figure 1 (at left - click to enlarge) demonstrates the above argument nicely. In the 1970s, the party insiders had not yet mastered the reforms that created the current system of primaries and caucuses. McGovern (1972) and Carter (1976) were able to win the nomination despite not being the consensus choice of party insiders. By 1980, however, the candidate with the most endorsements was always getting the nomination.
This year, the Republican party is kind of fractured. According to these authors, Romney has the plurality of endorsements, but only to the extent that Dukakis did back in 1988. That graph appears below:Additionally, only 8 of the 22 sitting Republican governors have even bothered to endorse this year. It seems clear, though, that the insiders have not coordinated on Huckabee. He could end up being this year's Howard Dean: the insiders aren't sure who they want, they just don't want him. Of course, they also don't want McCain, for reasons that this TPM commenter eloquently states.