The Delegate Fight: 2012

[Erc]

Something to look out for in a few months:

Since there is no threshold at all in Texas, there's a possibility (if the rounding goes well) that a candidate with as few as 0.34% of the overall vote in Texas can win a delegate.

Will Perry (or Cain, or Roemer for that matter) at long last win a delegate in this race?

EDIT:  This of course goes out the window entirely if the TX GOP ditches the primary, as looks increasingly likely.

[Erc]

Erc, I'm wondering if you've read this latest blog post by Nate Silver:

http://fivethirtyeight.blogs.nytimes.com/2012/02/25/the-g-o-p-s-fuzzy-delegate-math/

in which he talks about the large number of delegates to the RNC who are not exactly technically bound to any particular candidate.  In 2008, there was a media narrative that the Democrats faced a huge problem in that unelected Super Delegates might end up deciding the nomination, but that this problem couldn't have existed on the GOP side, since they have so few such "Super Delegates".

But reading this post, I'm wondering if this narrative should actually be turned on its head.  Does the GOP, in fact, have a significant share of delegates who are not exactly "bound" to any particular candidate, and thus there's a good chance that the primary season ends with Romney and Santorum furiously lobbying delegates to publicly back them, and give them a majority, so as to spare the party a convention fight?

Great post by Nate Silver, as this is exactly the sort of thing people should be paying attention to...but I think the risk of those bound/unbound technicalities is a bit less than that first chart might make it out to be.

A lot of these delegates are chosen at State Conventions, which mostly come very late in the cycle (I think the earliest is North Dakota's, at the very end of March).  At this point, it will be more clear how the race has shaped up, and if it really is a fight to the finish, supporters of each candidate will really be trying to elect loyal delegates who can be counted on to support their candidate in Tampa.  But maybe compromise candidates will be chosen if conventions have odd rules and end up getting extremely deadlocked (most likely in Iowa---though ND does this automatically in some sense, giving additional party leaders an almost-guaranteed ticket to Tampa)

A lot of the fighting for delegates will be done in odd venues---State Conventions, certain state executive committees, etc.---but I imagine the results of each will be rather clear after the event.  There will be a large number of delegates not chosen by strictly democratic means, but their voting intentions at Tampa will presumably not be in doubt.

Similarly, I imagine not-strictly-bound delegates chosen in primaries (OH/IL, for example) will remain loyal; Romney's organization was presumably good enough to pick reliable candidates, and anyone who signed on to be a delegate with Santorum pre-Iowa has got to be a True Believer.  There's always the possibility of a Paulista Fifth Column, but I'm skeptical.

I would very much be worried about Pennsylvania, which is going to be a complete mess.  It comes after another 3-week break in the season, though, which may give candidates enough time to get their act together.

As Nate Silver says, this probably matters most when one candidate is relatively close but not close enough to clinch on the regular bevy of endorsements alone---if one candidate is close enough that a few votes here and there might get him through on the first ballot---or enough switching beforehand makes sure that the convention is a coronation rather than a floor fight.

In the end, though, the media narrative matters.  The primary season effectively ends on June 3 (we all know who's winning Utah's delegates), and a floor fight seems possible, and the media start gunning for it over the following three months, whoever is in second place is probably going to fight tooth and nail for it---and all of these technicalities may get a lot of media attention.

[Erc]

Oklahoma

...

You are absolutely right.  In fact, I believe that's what I originally had, but I changed it at the same time I corrected the opposite mistake I'd made in TN's rules.

Yet another reminder that the Green Papers are not infallible.

[Erc]

March 13

Alabama

Overview
50 Delegates (2.19% of total)
Primary
26 At-Large (Proportional, 20%)
21 by CD (Proportional, 20%)
3 RNC Members

AL 2012 Presidential Preference Primary Resolution

21 delegates are assigned by CD, 3 in each of Alabama's 7 Congressional Districts.  If a candidate wins a majority of the vote (or is the only candidate above 20%), they win all 3 delegates.  Otherwise, the winner receives 2 delegates and the second-place finisher receives 1 delegates.

26 delegates are assigned At-Large.  If a candidate wins a majority of the vote (or is the only candidate above 20%), they win all 26 delegates.  Otherwise, the delegates are distributed among all candidates above 20%, in proportion to their share of the vote among all threshold-meeting candidates.  Fractions are rounded to the nearest whole number.  If too many or too few delegates are assigned, the last-place finisher loses delegates or the first-place finisher gains delegates, respectively.

RNC Members

Bill Armistead
Paul Reynolds
Bettye Fine Collins - Santorum

Results (as of 4/6)

Santorum - 22
Gingrich - 13
Romney - 12

Results have been certified.

American Samoa

Overview
9 Delegates (0.39% of total)
Convention
6 At-Large
3 RNC Members

A Territorial Convention meets and chooses 6 At-Large delegates.  "The delegates from American Samoa are chosen in such a way so that they best reflect the presidential preference of the Caucus/Convention participants."

RNC Members

Victor Tofaeono - Romney
Te'o J. Fuavai - Romney
Amata C. Radewagen - Romney

Results (as of 3/14)

Romney won all delegates.

Hawaii

Overview
20 Delegates (0.87% of total)
Caucus
11 At-Large (Proportional)
6 by CD (Proportional)
3 RNC Members

Based on the caucus vote, 11 At-Large delegates and 6 CD delegates (3 for each CD) are assigned.  In each jurisdiction, starting with the first-place finisher, delegates are assigned in proportion to the candidate's share of the total vote in that jurisdiction, rounding all fractions up.  Repeat for the next-place finishers until all delegates are assigned.

RNC Members

David Chang
Ted Liu
Miriam Hellreich

Results (as of 3/14)

Romney - 9
Santorum - 5
Paul - 3

Mississippi

Overview
40 Delegates (1.75% of total)
Primary
25 At-Large (Proportional, 15%)
12 by CD (Proportional, 15%)
3 RNC Members

12 delegates are assigned by CD, 3 in each of Mississippi's 4 CDs.  25 delegates are assigned based on the statewide vote.  In each jurisdiction, if a candidate receives a majority of the vote, they receive all delegates.  Otherwise, delegates are assigned amongst all candidates meeting a 15% threshold, in proportion to each candidate's share of votes among threshold-meeting candidates.  Rounding is done to the nearest whole number.  If too many or too few delegates are assigned, the last-place finisher loses delegates or the first-place finisher gains delegates, respectively.

RNC Members

Joe Nosef
Henry Barbour - Romney
Jeanne Luckey

Note that Henry Barbour was in the Perry camp until he dropped out.

Results (as of 3/14)

Santorum - 13
Gingrich - 12
Romney - 12

[Erc]

With Arizona called for Romney, he wins the state's 29 delegates.

[Erc]

Going to wait until the morning before seriously tackling Michigan.  A first glance would seem to suggest Romney has at least 13 delegates in the bag, and Santorum has at least 7.  I expect a majority of the remaining districts will break towards Santorum.

[Erc]

Michigan by-CD results.  Again, not all precincts are reporting so there may be resulting errors.

CD 1: Very Likely Santorum.  Santorum has a 687-vote lead in the whole counties that comprise CD 1.  Mason County is split between CD 1 and CD 2.  Mason County is not reporting more detailed results, but Santorum won the county as a whole, so it is very unlikely Romney will make up the difference there.

CD 2: Santorum.  Santorum wins by at least 8500 votes.

CD 3: Santorum.  Santorum wins by at least 1100 votes.

CD 4: Santorum.  Santorum wins by at least 2000 votes.

CD 5: Too Close To Call.  Romney has around a 900 vote lead outside of Tuscola County (with some uncertainty, mainly due to incomplete reporting of absentee votes by precinct in Saginaw Twp). Tuscola County is not reporting by precinct, and Santorum leads there, but it seems unlikely that Santorum will make up the difference there.  However, the AP shows Santorum with a 200-vote lead (thanks cinyc!), so I don't know what to make of it.

CD 6: Santorum.  Santorum wins by at least 4000 votes.

CD 7: Santorum.  Santorum wins by around 700 votes.

CD 8: Romney.

CD 9: Romney.

CD 10: Likely Romney.  Division of Sterling Heights City and Tuscola County is unclear, but it seems Romney won the district by at least 2000 votes.

CD 11: Romney.

CD 12: Romney.

CD 13: Likely Santorum. (apparently, according to the AP)

CD 14: Romney.

I didn't bother digging through Wayne County, but I find it very unlikely Santorum wins a district there.

That's a total of 7 CDs for Romney, and 7 for Santorum.  This results in a 15-15 split of delegates in favor of Romney.

[Erc]

Macomb precinct results are here.

MI-10 went to Romney.   MI-13 went to Santorum.  MI-05 is close, but Santorum leads by 269 votes with 7 precincts out per the AP.

If Santorum's MI-05 lead holds,  it should be 8-6 Santorum, which means he won the most delegates 16-14.  If not, it's a tie and Romney wins the delegate tally 16-14.

Yeah, I found the Macomb county results, and Romney does indeed win CD 10.

I do not feel like digging through Wayne County, so I will trust the AP there.  Number of votes in CD 13 is so small compared to the other CDs, it's pretty hilarious.

I am skeptical of the AP's CD 5 results, personally.  CD 5 breakdown:

Arenac, Bay, Genesse, and Iosco counties are entirely within CD 5.  In these counties:

Romney - 18632, Santorum - 18126 (506 vote margin for Romney).

Saginaw County is split with CD 4.  CD 5 contains the entirety of Blumfield Township, Bridgeport Township, Buena Vista Township, Carrollton Township, City of Saginaw, Spaulding Township, City of Zilwaukee, and Zilwaukee Township.  In these municipalities:

Romney - 1274, Santorum - 1622 (348 vote margin for Santorum).

Saginaw Township is split with CD 4.  Due to splitting of precincts, it's impossible to say the exact vote split.  However, the election-day vote in the CD 5 portion appears to be approximately:

Romney - 1561, Santorum - 1132 (429 vote margin for Romney, approximate).

The Absentee Vote is not reported by precinct (and many municipalities in Saginaw County aren't reporting AV at all, making me suspect that some may be from outside Saginaw Twp).  Assuming the Saginaw Twp AV vote breaks down by the same proportion by CDs as the election-day vote, and the %Romney isn't different in each CD (in actuality, Santorum did better in the CD-4 portion):

Romney - 403, Santorum - 119 (284 vote margin for Romney, approximately).

This is an approximate 871-vote margin for Romney outside of Tuscola County.  The most uncertainty is in the absentee vote totals in Saginaw Twp, but I'd imagine this margin is still accurate to within 150 votes either way.

In Tuscola County as a whole, the split was:

Romney - 2081, Santorum - 3089 (1008 vote margin for Santorum)

Slightly more of Tuscola County is in CD 10 than CD 5.  It thus seems quite unlikely, barring weird voting patterns, that Santorum would make up the difference in Tuscola County.

For main-page purposes, I'm going to move CD 5 back to Uncommitted, for now (Romney and Santorum trade CD 10 and CD 13).  The remaining delegates are split 15-13 for Santorum.

[Erc]

I found an actual MI GOP source for the CD breakdowns.  Since the MI GOP is what matters here, we should take this as the most authoritative source:

MI GOP Primary Results by CD

This has some major differences from the AP results.  Note that Romney is ahead in CD 5 by a few hundred votes, as expected.  CD 14 is the really odd one, though, with more than double the number of votes reported here than in the AP results.


Both AP and MI GOP counts are missing thousands of votes (compared to the statewide total), so they obviously haven't finished their tallying.

[Erc]

For main-page purposes, I'm going to move CD 5 back to Uncommitted, for now (Romney and Santorum trade CD 10 and CD 13).  The remaining delegates are split 15-13 for Santorum.

Why do you get 15-13?  Are the statewide two delegates split or do they both go to the winner?

The reports are rather conflicting on that point.  The original delegate plan (pre-sanctions) had the At-Large delegates being assigned proportionally, with a 15% threshold.

Post-sanctions, it hasn't been as clear how the remaining 2 At-Large delegates will be assigned.  Some sources have said WTA, but the campaigns and sources within the MI GOP seem to indicate they'd be allocated proportionally (i.e. 1 for each), and I'm going with the latter interpretation.

[Erc]

To make clearer my skepticism of the AP results by CD:

It's clear the AP is missing a bunch of results.  If you sum up their results by CD, Romney is 23,405 votes short of his statewide total (according to the Michigan SoS), and Santorum is short 16,582.

Most likely they just omit precincts that are split or that they are unsure about which CD they're in...it's 99% of precincts they know about reporting.

The situation for the MI GOP count is a bit better, but a bit weirder.  Romney is only short 893 votes, but Santorum has 2,532 more votes than in the SoS count.

[Erc]

Update on Alaska:

A Presdential Preference Poll is held statewide on March 6th, Super Tuesday, but only 17 out of 40 District Conventions are held on the 6th, and it is the preference of attendees at the latter that matter.

As a result, we should really not try to extrapolate any delegate information from the results from Alaska on Tuesday (not all people who vote in the Presidential Preference Poll will attend the District Conventions, even if held on the same day and in the same place; there are registration fees for the latter, for example).

We will not know the complete delegate distribution from Alaska until the final District Convention is held on March 24 in Nome.

[Erc]

Alaska: March 6

Overview
27 Delegates (1.18% of total)
Conventions
24 At-Large, Proportional
3 RNC members

AK GOP Rules

The 24 At-Large delegates are allocated proportionally based on the vote held on March 6.  The proportionality is not based on the raw vote; instead, each district convention has its say weighted by the number of delegates to which it is entitled at the State Convention.  Number of delegates per district can be found here.

All rounding of fractional delegates is done up for the number of highest finishers so that 24 delegates are allocated; all other rounding is down.

If any candidate entitled to delegates drops out between March 6 and the State Convention (April 26-28), the State Convention chooses officially 'Uncommitted' delegates in their stead.

RNC Members

Randy Ruedrich
Ralph Seekins
Debbie Joslin

Preliminary Results (as of 3/22)

Romney - 8
Santorum - 8
Paul - 6
Gingrich - 2

A final canvass of the votes has been completed and the resulting delegate count calculated.  Santorum picked up a delegate that I had originally assigned to Gingrich.

[Erc]

AP has updated their by-CD results and now Romney has a 500-odd vote lead in CD 5.

I still don't like the look of some of their Wayne County results, but I now think everyone is in agreement as to who won which CDs.

This makes a final delegate total of Romney 15, Santorum 15 (with an outside shot at a 16-14 split if Romney is awarded both At-Large seats).

[Erc]

I was mistaken as to the role of the Wyoming Precinct Caucuses (the Green Papers are not gospel, the Green Papers are not gospel).

My apologies to the folks I (wrongly) "corrected" over the last week or so.

Very tentative delegate allocation is:

Romney - 11
Santorum - 9
Paul - 6

(An alternative delegate allocation featuring Romney with 10 and Gingrich with 1 is also feasible [and CNN uses it], but I imagine Gingrich will get winnowed out at the County Conventions).

As usual, this doesn't account for a whole host of factors, all of which are exacerbated by the low turnout.  More details in the main Wyoming post.

[Erc]

Late March
or, The Treasure Fleet stops in San Juan in March, late in the month.

Puerto Rico: March 18

Overview
23 Delegates (1.01% of total)
Primary
20 At-Large, Proportional
3 RNC Members

If a candidate wins a majority of the vote, they receive all 20 delegates.  Otherwise, the delegates are allocated proportionally, apparently with a 15% threshold (rounding details unknown).

RNC Members

Carlos Méndez - Gingrich
Luis G. Fortuno - Romney
Zoraida "Zori" Fonalledas - Romney

Results (as of 3/18)

Romney won a majority and all 20 delegates.

Illinois: March 20

Overview
69 Delegates (3.02% of total)
Primary/Convention
54 by CD, "Loophole"
12 At-Large, Convention

IL GOP Bylaws
Sample Ballot

Each CD has between 2-4 delegates assigned to it, depending on its share of the vote for John McCain in 2008.  The exact allocation per CD can be found at the Green Papers.  Voters vote for delegates directly, casting their vote for 2-4 (depending on their district) delegates, and the top 2-4 (as appropriate) vote winners have their ticket punched for Tampa.  Delegate candidates' Presidential Preference is listed by their name on the ballot, though the delegates are not officially bound.  

A presidential preference poll is also held; this has no bearing on delegate allocation.

12 delegates are chosen by the State Convention on June 8-9.  It is unclear how the delegates attending the State Convention are chosen.

Ballot Access

Santorum has no delegates on the ballot in 4th, 5th, 7th, and 13th CDs, and thus can win at most 44 out of a possible 54 points delegates.

RNC Members

Patrick Brady - Romney
Richard Williamson - Romney
Demetra DeMonte

Preliminary Results (as of 3/20)

Romney - 42
Santorum - 12

The AP has yet to call 11 delegates, so some of these may change as the final results come in.  Remember that a final 12 delegates will not be selected until June 9.

Louisiana: March 24, April 28

Overview
46 Delegates (2.01% of total)
Primary/Caucus/Convention
20 At-Large (Proportional, 25%)
18 by CD (Caucus/Convention)
5 At-Large (Smoke-Filled Room)
3 RNC Members

Primary: March 24

20 delegates are assigned proportionally, based on the percentage of the statewide total vote received, amongst all candidates meeting a 25% threshold.  Fractions are rounded to the nearest whole number.  Any remaining delegates (left over due to rounding errors or the threshold) are officially Uncommitted.

Caucus: April 28

District Caucuses choose delegates to the State Convention.

State Convention: June 2

Attendees meet by CD and choose 3 delegates per CD, for a total of 18. 

The convention as a whole also chooses the 20 At-Large delegates; these are mostly bound by the results of the primary, except for any designated as Uncommitted (there are 5 such delegates).

5 additional At-Large delegates are nominated by the Executive Committee and elected by the convention as a whole.

RNC Members

Roger F Villere, Jr.
Ross Little, Jr.
Ruth Ulrich

Note that Ruth Ulrich is also the vice-chairwoman of the RNC.

Preliminary Results (as of 3/25)
Santorum - 10
Romney - 5
Uncommitted - 5

Remember, an additional 23 delegates (as well as those 5 Uncommitted At-Large delegates) are chosen later in the process.

Missouri: March 15-24, April 21, June 2

Overview
52 Delegates (2.27% of total)
Caucus/Convention
24 by CD
25 At-Large
3 RNC Members

Missouri Caucus and Convention information

A Presidential Preference Primary was held on February 7.  This had no impact whatsoever on delegate allocation.

County Caucuses: March 15-24

County caucuses elect delegates to Congressional District and State Conventions.  Most counties caucus by March 17; however, Jackson County and the city of St. Louis caucus on the 24th.  The number of delegates elected per county is proportional to the county's vote for John McCain in 2008.  Detailed allocation can be found here.

It is unlikely that reliable projections can be made from these caucuses, as no straw poll is held and there is no centralized reporting of any kind of results.  As such, we will be in the dark until April 21.

CD Conventions: April 21

Each CD elects 3 delegates, for a total of 24 statewide.  Attendees vote for individual delegates, who will have stated their presidential preference.

State Convention: June 2

The convention as a whole chooses 25 At-Large delegates.  Attendees vote on full slates of delegates.  If no slate wins a majority in the first ballot, the top two slates are voted on in the second ballot; the winning slate goes to Tampa.

RNC Members

David Cole
Lance Beshore
Ann Dickinson

[Erc]

so you have Romney at under 50% of delegates awarded so far?

Just barely over.  A bit more over if you include supers in states that haven't voted yet.

I ran through a somewhat optimistic (but not crazily so) scenario for Santorum recently...Romney just sort of hovers around that 50% mark for the rest of the campaign, and would need maybe only a third of the superdelegates to clinch on the first ballot.

I'll post a more detailed set of projections after Super Tuesday, but it is going to take some doing for Romney not to effectively clinch this by June 3.

[Erc]

I'll post a more detailed set of projections after Super Tuesday, but it is going to take some doing for Romney not to effectively clinch this by June 3.

Romney's going to clinch this two days *before* California votes? 

Yeah, when Obama wins South Dakota and Montana on June 3, 2008.

[Erc]

As noted elsewhere, the Michigan GOP has clarified that the 2 At-Large delegates are to be assigned WTA, so the final total is, as cinyc suspected, 16-14 in favor of Romney.

[Erc]

Politico reporting that Santorum may not qualify for upwards of 18 Ohio delegates because he didn't qualify for any in nine different congressional distircts.


I mean, we can talk about "momemtum" day and night but if THIS is how Santorum's campaign is being run, combined with failing to qualify for the VA ballot -- then wow. I knew Romney's campaign was organized, but I think it's the failure of the other campaigns to have any semblance of organization that is helping him more.

He is at least on the ballot in 6 of those congressional districts (he isn't in the other 3, for the purposes of district delegates), so he has a shot of 'winning' those delegates.  I assume the Ohio GOP will coordinate with his campaign when choosing whatever additional delegates need to be allocated.

[Erc]

Making a projection in Washington is difficult, due to the enforced runoff balloting at the LD/County conventions (once sub-20% slates are eliminated, it's French-style between the two top placers).  This means that tactical voting plays an extremely important role in determining what slates are chosen to go to the State Convention.

If one is completely agnostic about this, and just assumes supporters of candidates who don't make the last runoff distribute themselves evenly (or just don't vote), Romney easily wins the entire Tampa delegation.

If one assumes Gingrich supporters predominantly go to Santorum (while remaining agnostic about Paulistas), Santorum should pick up the delegates from CDs 5 (Spokane) and 6 (Olympic Peninsula), but the convention as a whole is dominated by Romney supporters.  Paul is a non-factor.

Of course, in addition to these tactical voting concerns, there are the more mundane concerns...LD results aren't known yet, so it's possible someone other than Romney will be able to dominate certain King County (and environs) LDs and change this analysis.  And of course, Stealth Paul, etc.

I'm going to tentatively go with the Gingrich -> Santorum tactical voting prescription.  To be honest, I think it's a decent lower bound for Romney, and as he's the man to beat we may as well be conservative with his delegate projections.

Romney - 34
Santorum - 6

 

[Erc]

I need to display Maine and Nebraska's CDs in order to get the proper delegate total to show up in the state itself; pay no attention to the '1' value in each CD (though I have colored the Maine CDs appropriately for reference).

The numbers off the coast of Florida correspond to Puerto Rico and the insular territories (Virgin Islands, American Samoa, Guam, and the Northern Marianas).  All these territories have already voted, all overwhelmingly for Romney.

Atlas conventions for the colors:  Green = Romney, Blue = Gingrich, Orange = Santorum, Yellow = Paul (ME CD 2).

Of course, most states are not winner-takes-all, so don't take a map like this too seriously.

[Erc]

Chronological delegate assignment:

For states that have multiple processes or stages (Louisiana, for example), the conclusion of the first media-reported process/stage that has any bearing on delegate allocation is what is reflected here.

Gray: Already occurred
Orange: April 21
Red: April 24
Blue: May 8
Green: May 15-29
Yellow: June

All insular territories have already voted.

[Erc]

Did you colour Alaska wrong or does the caucus just not pick the delegates tomorrow? Anyway, thanks very much. I love the way Georgia is the 4th most important.

The poll taken tomorrow does not actually determine the delegate allocations; a separate poll is taken upon entrance to the district conventions, which take place over the next few weeks.  A sizable number of the district conventions do take place tomorrow (a bit under half of the total), but they do not finish until Nome has its on the 24th.  Even for the district conventions that do take place tomorrow, there could be a substantial difference between the two votes; there are registration fees for the district conventions, for example.

[Erc]

The easy calls:

Romney wins all of Idaho's 32 delegates as he got a majority in the state.  

Romney wins all of Massachusetts' 38 delegates as no other candidate got above 15% of the vote.

In North Dakota, the delegation (which automatically includes around 7 superdelegates) may assign itself proportionally based on the statewide vote.  If this happens, the vote would presumably be 11 Santorum, 8 Paul, 7 Romney and 2 Gingrich.  I'll check later to see how reasonable this expectation is, but I'll stick with this allocation for the time being.

In Virginia, Romney won a majority statewide and won all At-Large delegates.  Apparently, Paul won VA-03 and does pick up 3 delegates in the state to Romney's 43.