Monday, October 20, 2008

Why McCain is Farther Behind Than You Might Think

I'm going to take a quick break from education to write one -- and only one -- post about the presidential election.

The consensus on the election seems to be that Obama is in the lead, but McCain's not out of striking distance (as he says: "we have them right where we want them"). The national polls are a quick snapshot, but they don't reflect the way we elect our next President. The next President will have to win at least 270 electoral votes; not a majority of Americans (though let's hope, whoever wins, that they win both). I hear a lot of talk about the polls, but very few people seem to really break it down. So, here is a breakdown of the states that will decide who wins:

After each state I have placed 5 numbers in parentheses (yes, it's a lot of information -- but it shows just how iffy looking at polls or predicting the race is). Here's an explanation of each number:

1.) A simple average of McCain's average lead/deficit in all recent polls, calculated by RealClearPolitics.com
2.) McCain's average lead/deficit, as calculated by a regression trend, in the polls as calculated by Pollster.com
3.) A snapshot of McCain's current lead/deficit, as calculated by averaging polls and factoring in trends, demographic information, and what's happening in neighboring states as calculated by FiveThirtyEight.com
4.) A prediction of McCain's margin of victory/defeat on election day, given all the information above in addition to historical trends, calculated by FiveThirtyEight.com
5.) The odds of a McCain victory as calculated, again, by FiveThirtyEight.com

In short, the 5 numbers are: 3 estimates of how far ahead/behind McCain is right now, an estimate of what that gap will be on election night, and the odds that he'll win that state. I've rounded all decimals to nearest whole number to make it (slightly) easier to read. All margins are presented from the McCain camp's point of view. In other words, -5 means that he's 5 percentage points behind according to that calculation.

Ok, so McCain needs to win 270 electoral votes in order to win the election. Depending on which website/calculation you believe, there are about 150-160 electoral votes that are solidly polling in McCain's favor or strongly leaning that way. Which leaves about 110-120 electoral votes that he has to win from states that are either toss-ups or in which he's currently behind. For simplicity, I'll take the calculation of 155 leaning/solid McCain electoral votes that RealClearPolitics has and start from there. Assuming he wins the solid states plus Georgia and Montana, he would have 155 electoral votes. Then he needs to start winning some of the closer states. Below are all of the states currently declared "toss-ups."

McCain is slightly ahead in:
West Virginia, 5 electoral votes (+2|+5|+3|+6|91%)
North Dakota, 3e.v. (N/A|+4|+3|+4|84%)
Indiana, 11 e.v. (+4|+3|+2|+2|65%)

Obama is slightly ahead in:
North Carolina, 15 e.v. (-2|-3|-1|-1|38%)
Missouri, 11 e.v. (-3|-1|-2|-1|35%)
Ohio, 20 e.v. (-3|-3|-3|-2|28%)
Florida, 27 e.v. (-2|-4|-3|-3|25%)
Nevada, 5 e.v. (-4|-3|-4|-3|25%)

All those races are currently deemed too close to call by most pundits. If McCain makes a big surge he could, conceivably, win all of them. Let's just say for now that he wins all of the states in which he's currently clearly ahead in addition to every single toss-up state -- which would be quite an accomplishment. That would give McCain 252 electoral votes. Which means that even if McCain pulls that off he would still need to find 18 more electoral votes from somewhere. Here are the states that are currently "leaning" toward Obama:

Colorado, 9 e.v. (-6|-7|-6|-6|10%)
Virginia, 13 e.v. (-8|-9|-6|-6|8%)
New Hampshire, 4 e.v. (-9|-6|-8|-6|8%)
New Mexico, 5 e.v. (-8|-8|-9|-7|7%)

In other words, in order for McCain to win he has to do the following:
1.) Win all of the states that are currently leaning in his direction
2.) Win all of the states that are currently too close to call
3.) Win 18 electoral votes from states in which he's currently far behind (which almost certainly includes VA)

McCain may only be 6 points down in the national polls, but it's not that simple. A national resurgance is not enough. He basically has to win all of the closest states, and then he also has to win at least two other states where he currently has a 1 in 10 chance or less of winning based on a thorough analysis of historical trends, polling data, and demographics.

Another sign that Obama is further ahead than you might think is that all of the closest states right now are ones that Bush won in the last election -- some by large amounts (for example, he won NC by 12 points). The closest Kerry state is NH, which 538 gives Obama a 92% chance of winning and in which most of the recent polls have had him up double digits. Indeed, VA is the must-win state for McCain -- he has virtually no chance of winning if he loses VA -- and it's one that Bush won by 8 points in the last election.

Lest you think that there are other options beyond VA, NM, NH, and CO, here are the next closest states:

Minnesota, 10 e.v. (-10|-7|-9|-8|4%)
Pennsylvania 21 e.v. (-12|-16|-11|-10|2%)
Michigan 17 e.v. (-12|-20|-11|-10|2%)
Iowa 7 e.v. (-12|-11|-13|-13|1%)
Wisconsin 10 e.v. (-11|-8|-11|-10|1%)

The odds of him winning any of these states are slim to none. Which means that there are only a few ways he can win. It basically boils down to two possibilities:

1.) Win all of the close states + 18 electoral votes from VA, NH, CO, and NM
2.) Pull out a miracle

I'm not sure which one is more likely. Any number of things could happen in the next two weeks, but a McCain victory is probably not one of them.

And that's why McCain is even farther behind than you might think.

Update: One more thing I forgot to include that makes like difficult for McCain: early voting. The latest poll shows a dead heat in NC, but Obama is up 23 points among those who have already voted. 538 pointed out the same trend in a number of other battleground states as well last week. So if McCain didn't face a difficult enough task as it is, it looks like Obama supporters are more enthusiastic and better organized in crucial states, which means that:

1.) The polls may be understating Obama's support by underestimating the turnout of his supporters
2.) McCain will already be far behind in the vote tally when Election Day dawns
3.) It will be even harder for McCain to make up ground than you might think

3 comments:

Anonymous said...

Interesting breakdown. Never knew my home state (VA) would be so critical in the presidential elections. I think the last time VA went 'blue' was for LBJ!

Corey Bunje Bower said...

A number of the traditional battleground states have already been decided, so the focus is shifting. If, on election day, Obama wins VA then McCain's night is over. If McCain wins VA, however, Obama has a ton more states he could still win that would put him over the top.

j* said...

Nice analysis. I'm crossing my fingers that you're right and shopping for shoes I can wear to dance in the street. Most of the polls I've visited have Virginia leaning blue.