Kentucky Delegate Predictions

Yesterday I gave Oregon a try, today I'll try Kentucky, which votes on the same day.

Kentucky has 51 pledged delegates to award on its May 20 contest.  They break down as follows (from the Green Papers:


  • 17 delegates by statewide total


    • 11 At-Large delegates

    • 6 PLEO delegates


  • 34 district delegates


    • 5 delegates to KY-01

    • 5 delegates to KY-02

    • 8 delegates to KY-03

    • 5 delegates to KY-04

    • 5 delegates to KY-05

    • 6 delegates to KY-06


Normally I'd use the latest Survey USA poll, due to their accuracy so far this year (with notable exceptions in Southern states and Missouri), but their latest April 15 poll was Clinton 62%, Obama 26%, a gap of 36 points.  I have trouble believing that poll is anything but a massive outlier.  The poll has Clinton getting 23% of the African-American vote, as well as winning 18-34 year olds  by 60%-31%.

Survey USA is often held up as a beacon of accuracy, but they have a tendency, noticed by others as well as by me, to drop big outliers weeks before a contest, only to suddenly fall in line for their last poll.  They did this in Ohio (putting her up by 20% until the last week) and in Pennsylvania (putting her up by 18% a week out).  I would be surprised if Survey USA's last Kentucky poll shows anything greater than a 20% split.

Survey USA's previous Kentucky poll on March 31 found Clinton ahead of Obama 58% to 29%, a gap of 29 points.  This one has more believable breakdowns, and I'm going to go with these assumptions for my breakdown, although Clinton may very well get close to 70% as their last poll indicates.

For the purposes of this prediction, I will guess Clinton wins 65%-35%, in keeping with SUSA's first poll if they split undecideds.

...

The PLEO delegates will split 4-2 for Clinton as long as she clears 58.3%, but she won't get any more unless she crosses 75%, which ain't happening.

The At-Large delegates will split 7-4 as long as Clinton passes 59%, but she can't get any more unless she breaks 68%.

Subtotal: 11-6 statewide.

Then it's on to the district delegates.

KY-01: 5 delegates.
Represented by Republican Ed Whitfield, this district encompasses the western side of the state, bordering rural counties in Tennessee, Missouri, Illinois, and Indiana.  Obama will get stomped here, but will Clinton get above 70%?  It's possible, and unfortunately I think likely given the Survey USA polls, though I would love to be proven wrong.  4-1 Clinton.

KY-02: 5 delegates.
Represented by retiring Republican Ron Lewis, a rural district just south of Indiana.  Same dynamics here, can Obama avoid falling below 30%?  Until polling shows otherwise, this one is 4-1 Clinton.

KY-03: 8 delegates.
Represented by Democrat John Yarmuth, who has endorsed Obama in February, this is an urban district surrounding Louisville.  Will Obama be able to win this district?  I guess the question we should be asking (if we are accepting that Obama faces a 65-35% loss statewide) is can Obama avoid a 3-5 loss?  Whoever gets above 56.25% gets the extra delegate.  If Obama improves his standing in this state that shouldn't be a problem given the nature of this district.  However, if he's getting rolled over statewide he probably is lucky to just hold Clinton under that number.  4-4 for now.

KY-04: 5 delegates.
Represented by Republican Geoff Davis, this district is located in Northern Kentucky and borders the rural Ohio districts that gave Clinton over 70% of the vote.  Unless SUSA is way off the mark, he doesn't avoid that here.  4-1 Clinton.

KY-05: 5 delegates.
Represented by Republican Hal Rogers, this district takes in Eastern Kentucky and borders rural districts in Tennessee and Virginia that gave Clinton almost 90% of the vote.  Obama just needs to maintain viability (15%).  4-1 Clinton.

KY-06: 6 delegates.
Represented by Democrat Ben Chandler, this central district includes the cities of Lexington and state capitol Frankfurt.  Clinton will need 75% of the vote to get five delegates here, which I don't think she will manage given the cities here.  4-2 Clinton.

...

All together that's 35-16, a nasty loss for Obama that certainly won't be canceled out by Oregon, and one that would give her more net delegates than any state since Super Tuesday (only Arkansas, New York, and California have given her more).

However, I am confident that once Obama campaigns here his numbers will improve, and he will at least hold Clinton to 61 or 62%.  Just by doing that he will decrease the delegate loss substantially, reducing districts KY-01, KY-02, and KY-04 to 2-3 losses, and possibly winning KY-03 by 5-3.  if he can do that then the delegate count changes to 30-21, a much more palatable loss.

Tags: 2008, delegates, Kentucky, Primary (all tags)

Comments

18 Comments

Re: Kentucky Delegate Predictions

I thought delegate counts don't matter anymore?  Isn't that the new set of rules we're playing by?

by Chili Dogg 2008-04-25 12:11PM | 0 recs
Re: Kentucky Delegate Predictions

I hope not.  Because it seems like every single person has their own version of the popular vote count.  But there is just one delegate count, so I hope we stick with that.

by Skaje 2008-04-25 12:22PM | 0 recs
hehehe...we need

like an American Idol of Popular vote counts. So we can have the popular Popular vote count.

by kindthoughts 2008-04-25 01:10PM | 0 recs
Re: Kentucky Delegate Predictions

This delegate allocation system is the most idiotic thing ever.

by alvic63 2008-04-25 12:12PM | 0 recs
Re: Kentucky Delegate Predictions

It is confusing, but I am glad we do not use winner-take-all primaries the way Republicans do.  Proportional allocation encourages participation in all 50 states.  Maybe there's a better way to do it, but at least this way every state is a battleground, unlike the general election.

by Skaje 2008-04-25 12:23PM | 0 recs
Re: Kentucky Delegate Predictions

I've been thinking winner take all by CD might be a better system. I have problems with a national popular vote model, there is some very good arguments for keeping the electoral college system but if you did it winner take all by CD nationally, each CD gets 1 EV it might work nicely.

by Obama Independent 2008-04-25 12:44PM | 0 recs
Re: Kentucky Delegate Predictions

Perhaps just making all delegates at-large, split them directly according to the percentage that one wins statewide.  That could work, and would remove concerns that the district breakdowns favor one candidate or another.

by Skaje 2008-04-25 01:23PM | 0 recs
Re: Kentucky Delegate Predictions

Thanks for the analysis.  Very interesting.

by mefck 2008-04-25 12:13PM | 0 recs
Re: Kentucky Delegate Predictions
Kentucky will be the best state for Clinton after Arkansas
70-30 I would predict
by darlene25 2008-04-25 12:23PM | 0 recs
really???

a 40% split?

by kindthoughts 2008-04-25 01:11PM | 0 recs
Re: Kentucky Delegate Predictions

Great diary!  Please continue to get your nerd on here on MyDD.  We need more of that and less candidate piefights!!

by SpideyDem 2008-04-25 12:41PM | 0 recs
Re: Kentucky Delegate Predictions

So, are the Clintons now counting delegates again??

by Spanky 2008-04-25 12:49PM | 0 recs
Re: Kentucky Delegate Predictions

They should be.  They can argue about popular vote all they want, but if Obama goes into Denver up 150 pledged delegates he ain't losing.  She needs to narrow the gap to even have a chance.

by Skaje 2008-04-25 01:20PM | 0 recs
Re: Kentucky Delegate Predictions

Good wrap, looking forward to the updates.

by Jerome Armstrong 2008-04-25 01:18PM | 0 recs
With Kentucky, Can walk across the USA

When Clinton wins Kentucky, you'll be able to walk from the Pacific Ocean to the Atlantic Ocean without touching a state that Obama won. You'll also be able to walk from Mexico to Canada without touching an Obama state. In contrast, it will be impossible to walk from sea to sea, or border to border without touching a Clinton state.

Ok, doesn't really matter that much, except that it might indicate that Clinton could have an electoral vote advantage over Obama in the general election against McCain, but I thought it interesting.

by Zzyzzy 2008-04-25 01:33PM | 0 recs
Re: Kentucky Delegate Predictions

Thank you!

These are interesting and take a lot of work on your part. How about a diary doing the same analysis for WV? (If you have done that and I missed it, I'm sorry)

by arkansasdemocrat 2008-04-25 01:42PM | 0 recs
Re: Kentucky Delegate Predictions

West Virginia should be a pretty simple one because it only has 3 districts.  The question is whether Clinton passes 58% in any of them, if she doesn't the state is a 15-13 split, if she passes that number in all three of them it's 18-10.  I'm thinking the latter is more likely due to the surrounding counties.

by Skaje 2008-04-25 01:49PM | 0 recs
Re: Kentucky Delegate Predictions

I was in Kentucky a couple months ago.  The cab driver started telling me how Barack Obama doesn't say the Pledge of Allegiance and believes in the Koran instead of the Bible.  Nice...

Perhaps the dumbest thing about the delegate allocation system is the way so much hinges on whether a district happens to have an even or odd number of delegates.  If it's even, there's a high threshold to get any sort of advantage, but if it's odd, all you need is 50.01% to claim an extra delegate.  Leads to really weird results.

by Steve M 2008-04-25 02:22PM | 0 recs

Diaries

Advertise Blogads