Saturday, February 18, 2012

A Very Rough Estimate of the Republican Delegate Math Ahead, Part One

One of the most frequent calls/requests FHQ receives from any and everybody is to provide a look forward and predict what will happen in any given upcoming contest or contests. And more often than not, I punt. Why? Much of what happens in presidential primary contests is seemingly predictable.

...in hindsight.

Well, often it is easy to say that the frontrunner coming into the election year ultimately won the nomination and move on. The reality, however, is that the "getting there" for the frontrunner can be much less easy to explain. The political science literature will lean on the amount of support various candidates have had entering the year of the election as quantified by FEC fundraising totals, poll position and more recently elite-level endorsements as a means of explaining the emergence of a frontrunner in a given nomination race. In recent cycles -- particularly 1996-20041 -- candidates were able to parlay success in fundraising, polls and endorsements during the invisible primary to early wins and put the nomination away fairly quickly.

Of course, one tie that binds those contests and not 2012 is the fact that the primary calendar is hugely different.2 In the era cited, frontrunning candidates had the ability to turn early success in both the invisible primary and the first handful of contests into momentum that would pay dividends on Super Tuesday. Early wins begat a small delegate lead begat many wins on one day begat a big delegate lead.  Afterward, challenging candidates, while still mathematically able to catch up, tend to be overwhelmed by a combination of the frontrunner's momentum in the contest and a delegate margin that is big enough that given the rules in the remaining states makes a continued challenge near futile. [Look, I cite it all the time, but since we are seemingly headed down a road toward a delegate count, go and acquaint yourselves with Norrander's End Game article from the Journal of Politics (2000). Contained therein are the calculations/equations you'll likely be hearing a great deal about over the next few months.]

But as I said, 2012 is different. There weren't 25 contests awaiting the remaining candidates a week after the Florida primary in 2012 as was the case in 2008. No, it was just Nevada, Colorado, Minnesota, Missouri and Maine with virtually no delegates directly at stake. Unless one candidate had come close to sweeping those contests and the ones preceding them, there really was not a viable knockout punch strategy for any of the candidates -- much less the race's nominal frontrunner, Mitt Romney.

So, when Neil King of the Wall Street Journal called this week to discuss how we might game this process out given the current dynamics, FHQ was typically hesitant. A presidential nomination race is a sequential process -- one step affects subsequent steps -- and attempting to parse out the possible scenarios can quickly become an exercise in futility. Yet, if we want to know if a knockout punch strategy is still possible, it might be helpful to actually examine a scenario or two to assess the odds of Romney extending his current delegate lead to a level that would be difficult for his opponents to overcome throughout the remainder of the calendar.

If you've read the Wall Street Journal today, you'll see some calculations attributed to FHQ. I thought that, in conjunction with Mr. King's piece, it might be helpful to show my work. Now, I cannot stress enough that what follows is not likely to happen. It is a thought exercise to help us all better understand the implications of not only wins for a candidate but how the candidates will potentially amass delegates moving forward in this race. I entered this process with a simple question/hypothesis: Can Mitt Romney deliver a series of wins on Super Tuesday that will end this contest on March 7? The answer there is pretty clear: No. Well, the answer is no based on the delegate count anyway. It is difficult to quantify the momentum coming off that sort of sweep. But even then, I think we can all agree that a Romney sweep -- John Avlon's North Korea-style election scenario -- of the Super Tuesday contests is unlikely.

The best test is to look at a couple of scenarios and compare them through Super Tuesday. King's WSJ piece alludes to these but let FHQ break them down.

Scenario #1: Super Tuesday South = South Carolina, Super Tuesday Northeast = New Hampshire

Again, these are hypotheticals. Obviously the dynamics of this race have changed since votes were cast in those early states. However, we do have information from those votes that may be useful if extrapolated onto other regionally similar contests. One but of information that is missing to this point in the race is that there has yet to be a midwestern primary. Michigan will be the first indication of what a vote in the midwest -- in a primary -- might look like. The Michigan numbers will -- or would for this analysis -- helpful for projecting Ohio. But alas...

Assumptions:
1) This only includes the primary states through Super Tuesday. Caucus states, due to their rules, were suppressed from the analysis.
2) Votes in the southern Super Tuesday states (GA, OK & TN) mimic the vote total in South Carolina, but with Santorum playing the Gingrich role (Santorum 40%, Romney 28%, Gingrich 17%). Delegates are then allocated according to the rules in each of those states. [South Carolina is a better reflection of the possible vote across the South than Florida because of the make up of the electorates in those two states relative to the rest of the South.]
3) Votes in the northeastern Super Tuesday states (MA & VT) mimic the vote total in New Hampshire, but with Paul and Romney splitting evenly the Huntsman votes from New Hampshire (Romney 48%, Paul 30%). Delegates are then allocated according to the rules in each of those states.
4) If there is a split in a state between the at-large (statewide) delegates and the congressional district delegates, the assumption is that the congressional district vote follows the statewide vote.

5) Romney wins Virginia and all 11 districts outright (+50%).
6) Michigan and Ohio are split roughly in half between a Romney and a non-Romney. In Michigan that means a split of the congressional districts and a split of the two at-large statewide delegates. In Ohio that means a split of the 16 congressional districts and a proportional allocation of the at-large delegates.
7) The threshold to win any at-large delegates in Ohio is 20%. The assumption is that only two candidates (Romney and non-Romney) clear that barrier.

8) Statewide votes translate to the congressional district as well. 

Delegate count (given those assumptions -- Click link to see full breakdown):
1) Romney would pick up an additional 192 contest/bound automatic delegates (distinct from automatic delegates who are not bound by primary results) in the primary states. That total would be 221 with Arizona
2) That would bring his binding delegate total to 265 or 294 with Arizona.
3) Santorum (or candidate in second place X) would pick up an additional 143 delegates or with Arizona 172 delegates.
4) That would bring Santorum's total to 146 or 175 with Arizona. 
5) Paul would pick up 19 delegates in the northeast (because of NH) bringing his total to 27.
6) Gingrich would pick up 8 delegates in Oklahoma for clearing the 15% threshold (because of SC), bringing his total to 37.
7) Post-Super Tuesday, then, it would be: 

  • Romney 265 or 294
  • Santorum 146 or 175
  • Gingrich 37
  • Paul 27
Notes:
1) The delegate margin would increase by about 90-150 delegates for Romney depending on Arizona.

2) That assumes, again, that one non-Romney consolidates that non-Romney vote. 
3) In this case, the assumption is that Santorum is that candidate.
4) Virginia helps Romney neutralize other losses across the South on Super Tuesday.
5) Arizona along with modest gains in the northeast and breaking even across Michigan and Ohio is what drives Romney's delegate margin up.

Scenario #2: Romney's Rosy Outlook

The intent of this scenario is to provide a kind of best-case scenario for Romney, albeit a limited one. FHQ will throw the other candidates a bone here and assume that Romney gets up to 49% of the vote across the board. That allows Romney to take a great many delegates, but not trigger the winner-take-all allocation of at-large delegates in the states where the allocation is conditioned on one candidate receiving a majority of the statewide vote. This is another way of saying that FHQ will allow for the proportional allocation of those at-large delegates.

Assumptions:
1) This only includes the primary states through Super Tuesday.
2) Romney wins 49% statewide and in the congressional districts. This is more likely in some states than in others, but recall that this is a baseline sort of scenario for comparison's sake only. 
3) Related to #2, it is probably out of reach for anyone to get to the 66% threshold in Tennessee, so I'll treat it like the rest: Romney gets 49% statewide and on the congressional district level.
4) This may be a shortcut and kind of undermine the "best case scenario" argument, but I'll assume that the remaining vote and delegate allocation centers around one candidate (Santorum) instead of it being split among Santorum, Gingrich and Paul.
5) Romney wins Virginia and all 11 districts outright (+50%).

Delegate count (given those assumptions -- Click link to see full breakdown):
1) Romney would pick up an additional 281 contest/bound automatic delegates (distinct from automatic delegates who are not bound by primary results) in the primary states.
2) That would bring his binding delegate total to 354.
3) Santorum (or candidate in second place X) would pick up an additional 110 delegates.
4) That would bring Santorum's total to 113 or Gingrich's to 139.

Notes:
1) The delegate margin would increase to about 215-230 delegates for Romney.
2) That assumes, again, that one non-Romney consolidates that non-Romney vote. 
3) Due to the nature of the rules, Romney would win all the delegates in Michigan, Arizona and Virginia.
4) Again, that likely places some pressure on the non-Romney candidates and makes their quest -- compared to Scenario #1 -- more a matter of keeping Romney from getting to 1144 rather than achieving that level of delegates themselves. This is a point that is receiving very little discussion. How that narrative exerts pressure on or has the RNC exert pressure on the other candidates will be an important factor to watch. We may see that manifest itself in the form of automatic delegate endorsements (or additional elected official endorsements).

In other words, the odds of a momentum contest are quickly dissipating as we look ahead to what the state of the race is likely to be post-Super Tuesday and giving way to a nomination race focused on the gradual accumulation of delegates. The above comparison gives the impression that under certain circumstances -- and there are a lot of assumptions there -- the delegate count will be close after Super Tuesday. However, it is important to note that those sort of analyses fail to capture the dynamics of the race at any given point (momentum, polling snapshots, etc.). It is easy, then, to count delegates without factoring in how primary/caucus results and those very same delegate counts impact the race.

Now, a number of folks have already looked at the delegate math ahead and have come to the conclusion that this race will go on for a while. FHQ agrees but urges caution in counting delegates too far in advance. There are any number of permutations that could occur and thus momentum to potentially develop from that. It could also be that it becomes easier to project as we gather more information from future contests.

FHQ will take a crack at one such permutation in Part 2 tomorrow; an extension of the baseline model above through the end of the calendar. As a baseline, it will give us some understanding of just how far the delegate race may extend.

--
1 One could add to this McCain's clinching of the 2008 Republican nomination as well. There are some elements of this pattern in the Arizona senator's run to the nomination then.

2 The other major difference in 2012 relative to the past is the volatility of support for the candidates in the polls. That isn't to be discounted.


Recent Posts:
A Follow Up on the Maine Republican Caucuses


2012 Republican Delegate Allocation: New York

April Primary Given the Heave Ho in Texas


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

No comments: