Primary calendar / poll closing times and delegate allocation megathread (Christmas is saved!)

[Erc]

Beginning to work through delegate allocation for both sides this year, and I noticed something a bit weird about the Democratic side:

The formula for the number of base number of delegates given to each state is given in the Call of the Convention (most recent version I could find is here)

You must be logged in to read this quote.

Where SDV represents the vote for the Democratic candidate in the given state in the given year, TDV represents the vote for the Democratic candidate nationwide in the given year, and SEV means the state's electoral vote.

Basically, this formula gives equal weight to total population (or its rough proxy in Electoral Votes) and number of Democrats (or its rough proxy in votes for Obama x 2 and Kerry).

Running the numbers and comparing them with the DNC's stated results (Appendix B of the call), the numbers seem a bit off, and not just due to rounding errors.  Some states are underrepresented by the DNC's calculation compared to the formula (TX by 11 delegates, FL by 5), while other states are overrepresented (Ohio by 5, NY/PA/MA by 4).

What seems to be going on here is that the DNC chose to use not the current EV figures, but the EV figures from 2004/2008!  This accounts for virtually all of the differences (apart from obvious rounding discrepancies and one oddity with NY which may be due to Fusion); note that in particular they are not averaging the EV counts over the last three elections, either.

Basically, the population weighting being done is on the basis of the outdated 2000 census, favoring the north and east over the south and west.

No one seems to have caught this so far; in particular, Texas (the state with the most to lose), goes along with the DNC's count in their delegate selection plan.

Blue states are overrepresented; red are underrepresented---this is essentially a map of which states lost or gained EVs in the 2010 reapportionment. 

It should be stressed this isn't a large effect: we're talking about 30-35 (base) pledged delegates here, or about 1% of the total.

[Erc]

Erc, you should alert the Texas Democratic Party of this mistake, so that they can get the DNC to fix it.  You'll be a state hero in Texas for getting them their fair share of delegates.

I got on their email list in order to access the Texas Democratic Primary/Caucus results in 2008, and I'm pretty sure one of my email accounts' inbox is now 90% messages from the Texas Democratic Party.  'Bout time I get something in return

[Erc]

-snip-

Maybe they were arithmetically challenged.

If it were based on one election, they would use:

AF = 1/2 * ( (SDV/TDV) + (SEV/TEV) )

If you wanted to use multiple elections you would then use:

AF = 1/3 * (AF2012 + AF2008 + AF 2004)

Someone may have got confused when rearranging terms, and instead of:

AF = 1/2 * (1/3 * (SDV2012/TDV2012 + SDV2008/TDV2008 + SDV2004/TDV2004) )

decided to use their form which eliminates the 1/3 used to average the vote share. It is also possible that they wanted to weight the higher turnout elections heavier.

Since TEV is a constant, it doesn't matter which form they use, and may have even led them to cover up their actual calculation, such that you had to reverse engineer it from the Appendix.

Yeah, I thought of something like that, but even if you weight the EV 1/3 2004, 1/3 2008, 1/3 2012, you're still off by a noticeable number of delegates; you need to be using 2000 Census EVs consistently to get their numbers.

And (with two very minor exceptions in ME & MN) the numbers are entirely consistent in all states that had no EV change between the two censuses, suggesting that the problem is entirely on the EV side of things.

I ended up calling the number suggested in Appendix B, and they were kind enough to get back to me right away.  They suggested that this issue might be fixed in a newer version of the Call to Convention (i.e. the ones that incorporate the Bonus Delegates from later primaries or Regional Clusters), which should come out in a couple weeks.  (Though where exactly I'd find an updated version once they do it, I'm not sure).

[Erc]

The Republican Party calendar, as it stands at present.  States have been resized so that their areas are proportional to the number of delegates they have (one small square = one delegate)

Dark Red: February
Red: March 1
Orange: March 2-14
Yellow: March 15-31
Green: April
Blue: May
Purple: June
Gray: Unclear

The unclear states at present:

Colorado may hold a caucus in February, but if so will not hold a straw poll at the same time.  It's as yet unclear if the RNC will allow this, and Colorado may still change its mind.  The likely backup option would be March 1.

North Carolina seems likely to choose a March 15 primary (yellow).

D.C. is likely to have its caucus/convention in late March (yellow).

I have no idea what's going on in North Dakota, Alaska, or the minor territories (Guam/Virgin Islands/Northern Marianas/American Samoa).

The existence of three delegates depend on this year's gubernatorial elections in Kentucky, Louisiana, and Mississippi; each state electing a Republican governor receives an extra delegate.

[Erc]

I would like all state to allocate their delegates proportionally. Winner-take-all seems to be unfair to me, since you can easily have 4 to 6 candidates who have different appeal across a state. If FL would say that the statewide delegates are allocated by WTA and the district delegates are allocated 2x1 (winner gets 2, runner-up gets 1), that would be fairer.

Not sure what the fairest method is but the Republicans need some consistency in their allocation rules, the rules are allover the map. The Democrats system has it's flaws (Super Delegates)  but the basic framework is at least consistent.  Proportional both statewide and by CD, 15% threshold, 1/4 of delegates allocated statewide and 3/4 by CD.  Also the more Democratic CD's get more delegates.

I think in the Republicans' view is that it's consistent with their governing philosophy.  They wouldn't want to dictate to the states what to do, and they certainly won't impose proportional representation on everyone.  Even this year's change, making sure that caucus states bind their delegates proportionally to the vote at the caucus, is a huge and controversial change -- but one that is going to give a real boost to transparency next year.

It should be noted that the Democratic system has a lot of flaws too, many of which we saw in 2008.  As you mentioned, there are way too many superdelegates, as you mentioned.  The method of allocating delegates proportionally by congressional district also means a hell of a lot comes down to whether your congressional district has an even or an odd number of delegates.  In a close, Clinton-vs-Obama style fight, those districts with an even number of delegates are likely to split the delegates equally regardless of who wins the district.

[Erc]

I would like all state to allocate their delegates proportionally. Winner-take-all seems to be unfair to me, since you can easily have 4 to 6 candidates who have different appeal across a state. If FL would say that the statewide delegates are allocated by WTA and the district delegates are allocated 2x1 (winner gets 2, runner-up gets 1), that would be fairer.

Not sure what the fairest method is but the Republicans need some consistency in their allocation rules, the rules are allover the map. The Democrats system has it's flaws (Super Delegates)  but the basic framework is at least consistent.  Proportional both statewide and by CD, 15% threshold, 1/4 of delegates allocated statewide and 3/4 by CD.  Also the more Democratic CD's get more delegates.

I think in the Republicans' view is that it's consistent with their governing philosophy.  They wouldn't want to dictate to the states what to do, and they certainly won't impose proportional representation on everyone.  Even this year's change, making sure that caucus states bind their delegates proportionally to the vote at the caucus, is a huge and controversial change -- but one that is going to give a real boost to transparency next year.

It should be noted that the Democratic system has a lot of flaws too, many of which we saw in 2008.  As you mentioned, there are way too many superdelegates, as you mentioned.  The method of allocating delegates proportionally by congressional district also means a hell of a lot comes down to whether your congressional district has an even or an odd number of delegates.  In a close, Clinton-vs-Obama style fight, those districts with an even number of delegates are likely to split the delegates equally regardless of who wins the district.

Agree on the flaws in the Democrats system but still makes more sense than the way Republicans pick their nominee.  Biggest Republican system flaw is in states like California that give the same number of delegates on a WTA basis to Nancy Pelosi's district as to Kevin McCarthy's.

As someone whose last two CDs have been MN-05 and MA-07, I've always viewed that as an upside to the system.

[Erc]

So, I've been working through the exact processes of the first several states in preparation for some future posts, and a more overarching question has come up.

What on earth happens to delegates that are bound to candidates who withdraw or are not placed in nomination due to Rule 40?  Early versions of the GOP Rules at least provided for the former possibility, but the current one does not seem to.

Currently, Rule 16 states that states have to allocate and bind their delegates in some winner-take-all or proportional fashion, based on the results of the primary/caucus vote in that state.  Each delegate's vote must be consistent with this, and they are recorded accordingly by the Secretary of the Convention.

Some states attempt to give provisions to deal with this eventuality, but I'm not sure these are consistent with the overall dictum to bind and allocate delegates based on the vote.

The most egregious of these is South Carolina, a Winner-Take-All state, which simply gives the delegates to the second-place (or failing that, the third) if the first-place candidate is not placed in nomination.  Nevada allows candidates to choose to release or reallocate their delegates if they withdraw before the State Convention.  Most other states I've looked at (with the exception of Iowa) provide some means for candidates to release their delegates--or in the case of Alabama, for the delegates to release themselves.

Are these (largely common sense, SC excepting) state rules consistent with the National Rules, and withdrawn candidates' delegates may be released, as one would expect?  Or are they bound to abstain?

Also in Rule 16 is the line: "Except as provided for by state law or state party rule, no presidential candidate shall have the power to remove a delegate."  What does "remove" mean in this context?  Does it cover this eventuality?

This is not to mention the headache of an (unlikely) second ballot; are delegates still bound by the primary vote in their state, regardless of what the state rules say?  As this could lead to an infinite number of ballots, I imagine not, but this is an awkward area in the GOP rules.

[Erc]

In most cases, when candidates withdraw from the race, they simply 'suspend' their campaigns, rather than actually ending them. This is done so 1) they can continue to raise money to help pay off their debt from the campaign, and 2) they can quickly re-enter if the frontrunner dies or something.

'Suspending' is different from 'withdrawing' - withdrawing is almost never done in presidential contests, and indicates a permanent end to one's campaign. While suspensions are generally permanent, they are not required to be that way, and can theoretically be nothing more than breaks from campaigning. The rule (in most states) is that if you merely 'suspend' your campaign, your delegates stay with you, unless you officially release them. This is why Santorum and Gingrich had to officially release their delegates before the 2012 convention - since they only suspended their campaigns rather than actually ending them, their delegates were still pledged to them, and would have been forced to vote for Santorum/Gingrich at the convention if they had been left unreleased.

Good point.  Many states do seem to cover the suspension possibility as well in their rules, though in many cases this is subject to interpretation, let alone the question of whether the state rules are even allowed to cover this possibility.

For example, New Hampshire keeps its delegates bound to a candidate “as long as he shall be a candidate before [the National] Convention.”  South Carolina rebinds its delegates to someone else if the candidate they were initially bound to is not placed in nomination.  Nevada actively asks its candidates if they want to keep their delegates; if they forget to respond (as may be likely for someone who has suspended their campaign), the delegates are released.

Also, of course, candidates who suspend their campaigns early (i.e. before mid-March) are incredibly unlikely to have their names placed in nomination due to Rule 40, as they won't have a majority of multiple states' delegations.  Even if they wanted to jump back in later, they wouldn't qualify for the national convention ballot, and it's unclear what becomes of their bound delegates then.  Of course, this assumes Rule 40 will be enjoining at the convention, which as Mr. Morden points out, is not clear at all.

[Erc]

Frontloading HQ raises a point, in the depths of a post about Texas, that it has been led to believe that the 3 RNC members from each state are supposed to be bound by the primary in each state as a result of RNC Rule 16(a)(1).  This seems to follow, as the RNC members are not elected by a statewide ballot.

Currently, however, many (if not most) states don't have a procedure for binding these delegates.  Will they all be given to the statewide winner? Allocated proportionally in early states? Should they just be considered to be extra At-Large delegates?  The RNC summary treats them separately, and sheds no light here.

If this interpretation is right, it gets rid of basically all the remaining Republican "superdelegates," excepting the few from states without any preference vote (CO/WY/ND, plus some of the territories).  This would, in principle, clear up some of the math, if we had any idea how they were going to be allocated.

[Erc]

Beginning to work through delegate allocation for both sides this year, and I noticed something a bit weird about the Democratic side:

The formula for the number of base number of delegates given to each state is given in the Call of the Convention (most recent version I could find is here)

You must be logged in to read this quote.

Where SDV represents the vote for the Democratic candidate in the given state in the given year, TDV represents the vote for the Democratic candidate nationwide in the given year, and SEV means the state's electoral vote.

Basically, this formula gives equal weight to total population (or its rough proxy in Electoral Votes) and number of Democrats (or its rough proxy in votes for Obama x 2 and Kerry).

Running the numbers and comparing them with the DNC's stated results (Appendix B of the call), the numbers seem a bit off, and not just due to rounding errors.  Some states are underrepresented by the DNC's calculation compared to the formula (TX by 11 delegates, FL by 5), while other states are overrepresented (Ohio by 5, NY/PA/MA by 4).

What seems to be going on here is that the DNC chose to use not the current EV figures, but the EV figures from 2004/2008!  This accounts for virtually all of the differences (apart from obvious rounding discrepancies and one oddity with NY which may be due to Fusion); note that in particular they are not averaging the EV counts over the last three elections, either.

Basically, the population weighting being done is on the basis of the outdated 2000 census, favoring the north and east over the south and west.

No one seems to have caught this so far; in particular, Texas (the state with the most to lose), goes along with the DNC's count in their delegate selection plan.

Blue states are overrepresented; red are underrepresented---this is essentially a map of which states lost or gained EVs in the 2010 reapportionment. 

It should be stressed this isn't a large effect: we're talking about 30-35 (base) pledged delegates here, or about 1% of the total.

The latest versions of the Democratic allocations appear to have fixed these problems!  Apart from possible one-delegate rounding errors in Colorado, Indiana, and New York, the math seems to check out.  Never fear, Texas Democratic Party!

[Erc]

I'm planning to keep my thread running throughout the process, you could always sticky that in lieu of or alongside this?

[Erc]

Updated the OP to include the primary calendar.  It looks like there's still some contradictory information though about some of these states holding caucuses or conventions that may or may not hold preference votes.  Let me know if you spot any mistakes.

Only a couple notes:

1) I believe the ND GOP doesn't have caucuses per se, but instead 47 Legislative District Conventions, which are spread out, lasting from January 19 - February 28.  

2) The Democrats Abroad Primary starts March 1, but lasts until March 8.

3) EDIT: I have no idea what's going on in the Virgin Islands.  One GOP source claims March 19, while another claims March 10.  The latter claims VI GOP never filed a delegate selection plan, so they reverted automatically to their 2012 scheme, including putting it on the exact same date (even though March 10 is now a Thursday, not a Saturday).

This is a matter of taste, but you may also want to list the latter stages of delegate selection for the states without GOP preference votes (CO/WY/ND).  

• ND has its state convention (25 delegates) April 1-3.
• WY has County Conventions (12 delegates) March 12 and its State Convention (14 delegates) April 14-16
• Colorado has county assemblies at some point between the caucuses and March 26, its CD Conventions (21 delegates) April 8, and its State Convention (13 delegates) April 9.

[Erc]

Minnesota GOP caucuses start at 8PM EST.  Democrats say registration starts at 7:30 PM EST, presumably they also start at 8PM.  Don't know when/if either side shuts the doors.

[Erc]

Do we know if Colorado and Wyoming will still be holding non-binding caucuses just for the sake of it? If so, what will be considered the "results"?

Colorado will have no preference poll and leave their delegates unbound until the state convention. I've read conflicting information about Wyoming, but I don't believe they're having a poll either (same with ND).

None of the three states are even allowed to have a poll if they're going to be allocating their delegates in the manner they're doing so.  If they were to have a poll, they'd have to allocate their delegates proportionally based on the results of the poll.  Those states wanted their usual convention processes to matter, so they chose not to have polls.

Note that Wyoming and North Dakota have actually already begun their processes; their local caucuses/conventions are spread out over the course of weeks to months.

In North Dakota, 85% of the caucuses have in fact already happened; they started January 19th and continue until February 28th.

In Wyoming, Niobrara County held its caucus on February 16; some take place over the next week, but most are on March 1 itself.

This is how the allocation of delegates for the Vermont Democrat primary works:

• 11 district delegates are to be pledged proportionally to presidential contenders based on the statewide primary results as Vermont has only one Congressional District.
• In addition, 5 delegates are to be pledged to presidential contenders based on the primary vote statewide.

            3 at-large National Convention delegates
            2 Pledged PLEOs

Why don't they simply allocate all 16 delegates statewide? Why so complicated?

Democrats have some weird rules on the national level that they are very strict about, and they don't make exceptions even in one-CD states when applying the rules strictly makes the result look bizarre.

[Erc]

Resolving one of the final confusions about the calendar:

The Virgin Islands GOP has confirmed that they are holding their caucus on March 10.  This was not their preferred date, but their failure to submit a Delegate Selection Plan on time meant they had to go with the same date as in 2012, even though that means they have to hold their caucus on a Thursday.

[Erc]

Interesting that many states close before MN's caucuses start; wonder if that will have an effect here.