Third in a series of diaries on Mitt Romney's delegate math. (#3)
I currently project that Mitt Romney will win no more than 914 delegates which is 230 short of what he needs to win the nomination on the first ballot. This nomination is in the hands of "unbound" delegates. Many more delegates are released after 1 ballot, some after 2.
For information on the projection and purpose of this diary, as well as me "showing my work," see below.
Two major revisions in this version are a downward revision of Mitt's delegate number to 914 from 952 based on a poll from Texas showing him below the threshold require to win congressional district delegates there. Recent other polling, especially Michigan, saw Romney pick up a handful over the last projection, but he still suffers a net loss of -38.
The purpose of this series of diaries has been to analyze, based on the current state of things whether Mitt Romney has enough delegates to win a first-ballot nomination, given a scenario favorable to him. Depending on results as they happen, I may change the purpose of this diary.
The purpose of this series of diaries is not to accurately project the final delegate tally. Bias is deliberately built in at each step to make results favor Romney so that the projection is an "upper limit" to test whether Romney can get a first-ballot nomination.
The purpose of this series of diaries is not to assess what the likelihood of Rick Santorum, Ron Paul, or Newt Gingrich being the nominee is. These are separate questions. It seems unlikely to me that even with only a plurality of delegates, Romney wouldn't be the nominee. He simply has to accept the best offer to get whatever amount of delegates he needs to win. Vice President Santorum? Supreme Court Justice Ron Paul? Who knows? This accommodation could occur before the convention.
There is also added uncertainty due to the "penalty" assessed on the states that held "winner-take-all" primaries early. It is entirely conceivable that the "Automatic" delegates (that's Republican for Superdelegate) or others, even if they are not Romney supporters, could, for example, vote to reinstate the delegates from those states and put Romney over the top. Back room deal? Absolutely!
The truth is that no one knows what the real delegate total is even at this moment, nor can they, because many are unbound and may change their mind. In fact, even if a deal is struck, the "unbound" delegates could revolt and reject it. This makes projections even flimsier. What's more, there are so many delegates that are assigned at conventions, that shenanigans are possible--apparent even at this stage (see, e.g., Maine). The only ones we can be sure about are pledged delegates from winner-take-all states. But this diary isn't about an accurate tally overall. It's about whether Mitt can win under reasonably constructed circumstances based on present conditions.
Finally, a certain number of delegates haven't really been decided yet. These state conventions could start getting really interesting the more critical they become.
Some Terms Defined
"Bound delegates" are those delegates that are legally required to vote for the candidate to whom they are assigned. Some bound delegates are what we might think of as "Superdelegates," some are selected at state conventions or district caucuses. Depending on the state, they may be bound for 1, 2, or 3 ballots, or, "until released [by the candidate]." Candidates that are not legally bound are in some states called "morally bound." I do not count those as "bound delegates."
"Automatic Delegates" are delegates who are the 3 RNC members for each state or territory. Most times—but not always—they are unbound.
"Unbound Delegates" are delegates who may vote for whomever they please. Sometimes they are the RNC members from the state or territory, sometimes they are all of a state's delegates, or sometimes a portion. Some are selected from the grassroots, some are not.
It is important to note that delegates that are almost always selected at state conventions. Quite often, there is no reason other than the delegates selected to the state convention require it, that the number of delegates sent to the RNC will be apportioned how the caucus results occurred, especially in states where the delegates are unbound.
If the result at the end is contingent on a handful of unbound delegates, I can see all kinds of chaos.
Based on the work done in my first two diaries, I have refined my methodology. The rules for each state are based on this document from the RNC Counsel's office.
Most states aren't even polled at all. Those that are are almost never polled by congressional district. Yet in many states, delegates are selected based on who wins the most votes in a congressional district.
1. In winner-take-all states where there is no state poll, I assign all of the state's delegates to Mitt. This distorts the result in favor of Romney intentionally.
2. In winner-take-all states where there is a state poll and Romney is polling within the margin of error of a plurality, I assign all of the state's delegates to Mitt. This, again, distorts the results in favor of Romney intentionally.
3. In proportional states where there is a poll, I simply assign the number of delegates based on the state poll numbers, rounding up. For example, if the state has 15 delegates and Romney is polling at 33% in the state, I assign him 5 delegates.
4. In proportional states where there is no poll, I simply assign the number of delegates based on Romney's national poll numbers, rounding up. For example, if the state has 15 delegates and Romney is polling at 33% nationally, I assign him 5 delegates. The rounding up slightly biases the poll in favor of Romney.
5. In states where delegates are awarded partially on a winner-take-all basis, I award the winner-take-all delegates according to Rules 1 and 2.
6. In states where there are delegates awarded on a Congressional-district-level winner take all basis, I nevertheless assign the delegates according to Rules 3 and 4.
7. I assign all unbound Automatic delegates to Romney. This intentionally distorts the results in favor of Romney, but I provide a total both without them and with them proportionally assigned. Only 117 GOP "Automatic" delegates are unbound. (That's Republican for "Superdelegate".)
8. I use the delegate estimates from states that have already reported their results that are most favorable to Romney.
9. I assign Hunstman's delegates to Romney as his endorsee.
N.B. Some states have percentage thresholds below which no delegates are earned. Sometimes this is statewide. I only account for this in states with polling, e.g. Texas.
I am currently using a weighted average of 30% as the national polling number, again, to the high side of what is real at this very instant, based on the TPM Polling Average.
Here are some current delegate counts:
There are 2,286 total delegates. 1,144 are needed to win. Romney current has 134, including CNN's projection of 11 for Maine. There are 2,038 remaining, meaning Romney must win 43.4% of those left.
29 delegates, all bound but not legally for 1 ballot. (Penalized 29 delegates).
Projection: 29 bound.
30 delegates, penalized 29. Delegates bound for 1 ballot. Statewide delegates are proportional for those over 15% threshold. Winner-take-all on a CD basis. Gingrich is currently below the threshold.
Projection: 11 bound.
43 total delegates. 3 RNC, 30 CD and 10 statewide. 40 are bound for 1 ballot. Chosen at state convention on May 30-June 2.
Projection: 13 potential bound, 3 unbound = 16.
27 total delegates. 3 RNC, 3 CD and 21 statewide. The statewide delegates are apparently really proportional. Delegates are bound for 2 ballots.
Projection: 8 bound, 3 unbound = 11.
32 total delegates. 3 RNC, 6 by convention and 23 CD and statewide. Only 23 are bound for 1 ballot, 9 unbound, including 3 RNC. 23 are proportional from caucus with a 15% threshold. 6 are selected at state convention on June 21-23, 2012. Automatic delegates are bound. You would think Mitt would outperform his national polling here, but I don't have any data for that.
Projection: 8 bound, 9 unbound = 18.
41 total. 3 RNC, 27 CD, and 11 statewide. Proportional by CD with 15% threshold (what a mess). Statewide proportional with a 15% threshold. As with Idaho, I expect Mitt will outperform his national polling here, but I also have no data.
Projection: 13 bound, 3 unbound = 16.
28 delegates. 3 RNC, 3 CD and 22 statewide. All delegates unbound, elected at state convention.
Projection: 11 unbound.
66 delegates that are "morally" bound. 3 RNC, 48 CD, and 15 state-wide. Winner take all for statewide delegates if candidate gets 50% otherwise proportional with a 20% threshold. CD are winner take all.
Projected: 24 unbound.
43 delegates. 3 RNC, 40 bound until released with 25 statewide and 15 CD. 15% threshold in statewide and CD but 50% is winner take all in CD and at-large.
Latest poll: Gingrich 34, Romney 31, Santorum 15.
Projected: 15 bound, 3 unbound = 18.
58 delegates, 55 bound for 2 ballots. 3 RNC, 27 CD, 28 statewide. At-large has a 20% threshold, but winner-take-all if >66%. In CDs, >66% = winner-take-all. If top candidate has 20-66% of the vote, winner gets 2, second place gets 1, unless second place has less than 20%, then winner-take all. If winner is less than 20%, top 3 candidates each get 1 delegate.
Projection: 24 bound, 3 unbound = 27.
17 delegates. 3 RNC, 3 CD, 11 statewide. AL are proportional for all candidates over 20%, winner take all if over 50%. CD winner take all.
Projection: 4 bound, 3 unbound = 7.
49 delegates. 3 RNC, 33 CD, 13 AL. Proportional for candidates over 15%; winner-take-all for 50.001% or more. CD winner take all. Santorum and Gingrich are not on the ballot.
Projection: 46 bound, 3 unbound = 49 (Winner take all if over 50%).
Beyond Super Tuesday:
155 total delegates. 3 RNC, 108 CD, and 44 AL. 153 bound for 3 ballots. Proportional. CD proportional with 20% threshold. Latest poll: Romney 16%.
Not above threshold for CD delegates.
Projection: 7 bound, 3 unbound = 10 delegates.
Total: 914. Includes all types of delegates.
(1) Romney must begin winning at least 44% of delegates from each state now or else he will not achieve a first-ballot nomination.
(2) The presidential nomination process is in need of major rationalization and democratization in both parties. The rules are beyond Byzantine. They are Talmudic if I might coin a term. The voters, even in closed primaries and caucuses, will not decide this nomination; party insiders will.
Tue Feb 21, 2012 at 8:23 AM PT: I might or might not do it, but I would be very interested to see people reporting from the state conventions. All hell could break loose at those conventions if people are organized.