The Revenge of Santorum Can't Get to 1144

There were a couple of fairly consistent complaints about the model FHQ put together for ABC News to project candidate delegate totals through the end of primary season:

1. What if Gingrich/Santorum drops out and helps Santorum/Gingrich consolidate the conservative vote against Romney?
2. Sure, Gingrich/Santorum can't get to 1144, but FHQ still hasn't really shown how Romney can get there. 

First of all, I think these are fair assessments/concerns of/about the model, and as is the case with anything like this, it always [always] helps to add extra eyes to process. But let me try to flesh out the explanation of the model a bit as a means of addressing the above issues and then have a look at what the terrain is like now that Super Tuesday is out of the way.

Issue #1: A Gingrich/Santorum Drop Out

FHQ will have to admit right off the bat that I thought the folks who responded to the numbers in the Santorum Can't Get to 1144 post by calling the model a fantasy were right on. It is a fantasy. It is a complete fantasy that one candidate will receive exactly one half of the vote across all the remaining states and by virtue of having that same level of support applied to each and every congressional district (where states allocate based on that subunit vote), win all the delegates from those districts. Again, that is not going to happen. But what I tried to do with the second model was to account for the number of candidates who got over the various thresholds in the states to get any delegates at all. That number was set to its lowest possible number: three in states with 15% thresholds and two in states with 20% thresholds. It seems reasonable that if one candidate -- as the model already assumes -- has 50% support that one additional candidate will break that 20% barrier and that two candidates would break the 15% mark in states with such a threshold. 

The model built in the reallocation of delegates not allocated to candidates because they did not meet the threshold. And because that was set to the lowest reasonable number, the model already -- in an indirect way -- assumes that Gingrich or Santorum is getting votes but not a proportionate share of at-large/all delegates. Those are reallocated to the candidates above the threshold. 

Let me illustrate this with an example. We'll call it Kansas. First of all, we'll assume that Rick Santorum receives 50% of the vote in Saturday's caucuses. Furthermore, we'll assume that Mitt Romney only just clears the 20% mark to get any delegates. Ron Paul and Newt Gingrich both split the remaining 30% with neither clearing the 20% mark (We'll assume an even 15%/15% split). The way the model treats this, Santorum would get a majority in each of the four Kansas congressional districts and wins all 12 delegates (3 delegates in each of the four congressional districts). By winning the statewide vote Santorum would also win the three automatic delegates. 

But where the fun comes in is with the 25 at-large delegates. Since only Santorum and Romney cleared 20%, the allocation will not be truly proportional. If it was, Santorum would receive 12 delegates, Romney 5, Gingrich 4 and Paul 4. That is not how it works though. The eight Gingrich/Paul delegates have to be reallocated to the candidate who cleared the 20% threshold. Reallocating those delegates pushes Santorum's total to 18 and Romney's to 7. The final allocation of the Kansas delegates, then is Santorum 33, Romney 7. 

What happens if Newt Gingrich drops out?¹  

I think we can all agree that no matter what, Ron Paul is not going anywhere. He will continue on and likely get anywhere from 10-15% of the vote. That means that Santorum and Romney are splitting roughly 85-90% of the vote (not accounting for uncommitted votes/voters). What happens if, as the model's initial detractors so adamantly claimed, Santorum receives all of the Gingrich support and Romney stays right at that 20% level? Again, we assume that Santorum is already starting out with a base of 15 delegates (12 congressional district delegates and 3 automatic delegates). What we are concerned with are the 25 at-large delegates. Ron Paul performs fairly well in caucuses so let's say he gets 15% of the vote, meaning that Santorum and Romney split the other 85% (Santorum 65%, Romney 20%). In that instance, Santorum claims 19 of the 25 at-large delegates and Romney, the other 6.² 

All that -- a 15% gain in the vote share -- and Santorum just gained one delegate. So the model already kinda sorta assumed a two person race, but even with Gingrich out of the way AND capturing all of his votes does not translate into all that many delegates. That won't make up the difference to get Santorum to 1144. 

In fact, if we plug the new post-Super Tuesday delegate count (97 delegates: 95 from the RNC and 2 additional automatic delegate endorsement) into the model it only gets Santorum to 959. And that's 959 with all the advantages described above. If we, in turn, grant Santorum all of the unbound delegates to this point in the race (198 delegates), he just barely cracks 1144.³

I think we can all concede that that is beyond generous and approaching the "act of God" that the Romney teams was discussing in the context of Santorum's delegate math.

Issue #2: How does Romney get there?

Speaking of Romney, what does his number look like in the same, again, fantasy model? Once we plug in his current RNC delegate count (339) plus his automatic delegate endorsements (19), Romney reaches 1236 delegates. And that's winning half of the vote and claiming winner-take-all delegates in states that allocate winner-take-all based on reaching that plateau either statewide and/or in the congressional districts. But not counting any additional automatic delegate endorsements or any endorsements from caucus states actually allocating unbound delegates somewhere down the line, Romney still has a cushion of about 100 delegates to work with to get to 1144. In other words, we could assume that Romney likely won't get 50% of the vote in Alabama this next week, but expect that if things continue on as is -- with the former Massachusetts governor as a "weak" frontrunner -- that he would/could make up for "losing" those delegates (relative to the model) by gaining automatic delegates and/or delegates from unbound caucus states. Of course, the time in between the former and the latter happening means that it would be later in the process before Romney actually got to 1144. 

Is that a shoo-in? No, but Romney has some cushion with which to work that his counterparts in the race do not.

--
¹ No, Gingrich will not drop out prior to Kansas but bear with me here for the purposes of this exercise.

² Bumping Paul down to the 10% level does not change the allocation. Santorum still receives 19 of the at-large delegates to Romney's 6.

³ With the number of caucus states waning, we quickly are approaching the point at which the unbound total is not going to grow anymore than the three automatic delegates per state (for those that don't bound them). That 198 is and will be a sizable majority of those unbound delegates.


Recent Posts:
2012 Republican Delegate Allocation: North Dakota

2012 Republican Delegate Allocation: Massachusetts

2012 Republican Delegate Allocation: Idaho

Are you following FHQ on Twitter, Google+ and Facebook? Click on the links to join in.