Sunday, February 19, 2012

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

This morning, NBC News' Andrea Mitchell on Meet the Press brought up FHQ's delegate numbers from yesterday's Wall Street Journal. Al Hunt of Bloomberg News responded that the conclusion that Romney could wrap up the nomination -- surpassing the necessary 1144 delegate -- on either June 5 or June 26 (depending upon the date on which the Texas primary is ultimately scheduled) was misleading. Hunt is right. Those numbers -- based on an FHQ scenario analysis [see part one here] -- likely are misleading when taken out of context. However, the premise of the exercise was not to project when Romney or any other Republican candidate would or could amass the requisite 1144 delegates, but rather to assemble a reasonable baseline to which the actual and ever-changing delegate count can be compared. [...and, you know, actually utilizes the real Republican delegate allocation rules state-by-state.]

Is Mitt Romney likely to receive 49% of the vote in all the upcoming primaries? FHQ would contend that that is not all that probable. Yet, that scenario sets up a delegate accumulation that projects the current delegate count leader racking up wins, but wins at a level that will keep the growth of the delegate advantage at its slowest given the Republican delegate selection rules on the state level. In other words, Romney would win but would not win at levels (a majority of the vote in most states) that would trigger the winner-take-all allocation of either all of certain states' delegates or all of certain states' at-large delegates. Again, that is a level of victory that would give us a true sense of not only the contours of a stretched-out calendar but the nature of the delegate allocation rules changes in 2012 as compared to 2008.

Let's review the assumptions:
1) This includes the caucus states with defined binding or delegate selection rules (Alaska, Hawaii and Kansas) and primary states through the end of the primary calendar. [The Puerto Rico primary has undefined delegate selection rules at this time and it and the 20 non-automatic delegates are suppressed from the analysis.]
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):
Due to the inclusion of Alaska and Kansas in the count (a slight difference from the numbers reported in the Wall Street Journal), Romney, by gaining 49% of the vote in all the remaining states through the end of the calendar, would cross over the 1144 delegate barrier on June 5 -- with Texas on either May 29 or June 26. 

Given that this extends to the end of the calendar, the scenario analysis above is chock full of caveats. Let FHQ mention a few:
1) Again, this all follows the delegate allocation rules state by state. In some states, the automatic delegates are bound delegates. Where that is the case those delegates are included in the at-large total in the linked spreadsheet above. So, if you are studious enough to check my math against the rules, that is why there are a handful of states with at-large delegate totals for Romney that seem to have three too many delegates.
2) Illinois and Pennsylvania are loophole primaries in which delegates are directly elected on the ballot.  Even though both states send delegates to the convention unbound, FHQ has treated those delegates as if they have been allocated proportionally. There is a clearer transference of presidential preference in those two states -- under those rules -- than in non-binding caucus states. But, that is a point on which FHQ will admit that there is some room for debate.
3) As Al Hunt alluded to on Meet the Press -- well, in a sloppy sort of way1 -- this model does not account for momentum. It does not. A candidate could reel off a series of wins at some point on the calendar that would place upon the other candidates some undefined level of pressure to drop out of the race. This is as good a time as any to reiterate a point FHQ raised in part one: The math is not necessarily about getting to 1144 so much as it is about gathering enough delegates -- enough of a lead -- that makes it mathematically impossible for another candidate to overtake the leader (see Norrander, 2000). The decision-making calculus at that point will hinge not only on the pressure to drop out but the desire stay in and prevent the delegate leader from reaching 1144.
4) This model also does not account for the possibility that, unbound though they may be, delegates may [repeat: MAY] emerge in the intervening time from caucus state conventions who have expressed a preference for a particular candidate. As any of that information comes to light, it obviously impacts the calculations in the model above.
5) Similarly, another thing that is lacking above is any consideration of the unbound automatic delegates. If further endorsements are made by automatic delegates that also shifts the delegate count upward in a manner that may push the point at which the delegate leader surpasses 1144 to an earlier point.

Admittedly, that is a long list both of assumptions and caveats. Does that negatively affect the accuracy of the 49% model? Yeah, it probably does in some ways. But this was never going to be the way the delegate count was going to progress anyway. What this exercise does provide us with is something akin to a regression line through the delegate count across the remaining contests on the calendar (using the rules). Romney will overperform that level in some states -- though in only one thus far (Nevada) -- and underperform in others. The balance of those performances along with the addition of known unbound caucus state delegates, unbound automatic delegates and momentum affecting the dynamics of the race will determine when and if a candidate -- most likely Romney -- crosses the 1144 barrier earlier or later than in the above scenario.

1 Hunt constructed a scenario in which a candidate gains momentum by winning the upcoming Michigan primary and then sweeping the Super Tuesday contests. There aren't enough delegates there, but that could exert some pressure on the other candidates to drop out nonetheless. It ultimately comes back to the Southern Question FHQ proposed in the aftermath of the South Carolina primary.

Recent Posts:
Race to 1144: Maine Republican Caucuses (Updated Count)

Guam Republicans to Select Delegates at March 10 Convention

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

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

No comments: