Friday, October 31, 2008

Can Obama Lose?

Do Obama fans need to worry?

The answer is yes; but just a bit.

Though it is still possible Obama could lose, you’re much better off being an Obama fan today than being a McCain. Right now, four days before the U.S. election, the polls show Obama up by 6% nationally – based on the RCP average of polls. Currently Obama has 353 electoral votes – again based on the RCP averages of individual state polls. If we use only the most recent polls (which may or may not indicate the beginning of a trend) Obama’s electoral votes rise to 392. (270 electoral votes are needed to win.) The only way McCain can win is by the “perfect” – from his point of view - confluence of a number of factors.

How could that happen?

Well one way is, that with four days to go, he continues to eat away at Obama’s lead. If you look at the chart above you can see that from October 26 to October 30, Obama’s lead in the projected national vote dropped from 8% to 5.9%. That’s a 2.1% drop in just 4 days! If McCain can do the same in the next 4 days, Obama’s lead would be only 3.9% on election day: close but no cigar.

But the margin of error in most national polls is 2.5 – 3.5%. So if McCain can continue his effective momentum of the most of past week; and if the polls are at their maximum error; and if all that error was in Obama’s favour, then McCain can achieve a tie – or close to it – in the national vote. (Technically he would still lose by 0.5% - but that is truly too close.) That’s a lot of IFs. What’s more, from yesterday – Oct 30 – to today – Oct 31 – Obama went up 0.1% in the polls. So it would appear that the momentum McCain generated earlier in the week, with his focus on taxes, has stalled: either just slowed a bit or he has milked that issue for all it was worth.

Another way Obama could lose is if all the "undecideds" break for McCain – rather than splitting more or less evenly, or just staying home 'cause they are too undecided to bother to vote.

In this scenario Obama loses all the races where he has less then 50% today. On the national level, as of today, the RCP average of polls shows Obama with 49.9% and McCain with 43.9% of decided voters. So, if ALL the undecided voters vote; and if ALL of them decide for McCain, then McCain wins the election by 50.1% to 49.9%! But that’s only in the overall popular vote. And as we learned in 2000, American elections are not won by the overall popular vote, but by the electoral college and in the state by state contests.

So how does this “undecideds to McCain” scenario play out on state by state level?

As we see in the chart above, as of Oct 31, Obama is leading in states with 353 electoral votes, based on RCP averages of polls. But if he has less than 50% of the vote in enough states with enough electoral votes, and if all the undecided vote, and they all vote for McCain, Obama would lose 84 electoral votes, and thus lose the election.

Is that the case? Here is a list of states where Obama is leading but where, as of today, he has less then 50% of the decided vote.

State EV Obama McCain
--------------- -- ----- -------
Florida 27 48.5 45
North Carolina 15 48.9 46.3
Ohio 20 49.2 43.4

So if all the undecides in those states break for McCain, he picks up 62 electoral votes; but Obama still wins the election: 291 electoral votes to 247.

But again, the state polls have a margin of error. So maybe 50% for Obama is not REALLY 50% fro Obama. To be completely comfortable, and totally immune to the “all undecideds vote McCain” scenario, Obama needs 53.5% in any given state. If we apply this threshold the following states now come into play.

State EV Obama McCain
--------------- -- ----- -------
Nevada 5 50.3 43.3
New Mexico 5 50.3 43
Colorado 9 50.8 44.3
Virginia 13 51 44.5
Pennsylvania 21 52.3 43
Iowa 7 52.5 41.5
New Hampshire 4 52.8 40.5
Wisconsin 10 53 41.3
Minnesota 10 53 40
Oregon 7 53 37.8

If McCain can pick off the closest 4 of these races, or if he can pick off Pennsylvania and one other, he will in fact win the election.

But that would require all the stars aligning just right. ALL (and we are talking 95% or more here) of the undecides would have to decide to vote and to vote for McCain, and the polls would have to wrong pretty much at the maximum end of their margin of error, and all of that error would have had to be in Obama’s favour.

Is there any other way Obama can lose? Yes: low turnout. All the polls are based on a model of likely voters. And all agree that first time voters are more heavily tilted to Obama. All have assumed some of these first time voters will vote (and in fact one of the major differences in the poll results is based on how many of these first time voters they assume will actually vote.) If first time voters significantly under perform the pollster’s assumptions, and if some of the other scenarios described above come just partly true, Obama could still lose. But most news stories indicate that new voters will out perform the pollster assumptions – and by a lot.

So any way you look at it, a McCain victory, while still possible, is a long shot.

If you are still worried (or God forbid – hopeful) than here are a few early signs you can look for on Nov 4th.
1. Voter turn-out. High turn-out works to Obama’s favour. Low turn-out should make Obama supporters start to worry a bit

2. Florida. Obama is currently leading in Florida by 48.5 to 45. That’s a 2.5% lead. If McCain wins this state, Obama fans should up their worry level to yellow: caution is called for.

3. Virginia. Obama is currently leading in Virginia 51 to 44.5. That’s a 5.5% lead. If Obama takes Virginia by anything less than 2.5% Obama fans should increase their worry level to yellow. If McCain wins Virginia, then Obama fans should start chewing their nails: we are into threat level orange!

4. Pennsylvania. Obama is currently leading by 8.5%. If McCain makes this close, go to code orange. If McCain pulls off an upset and wins Pennsylvania, go to threat level red! It will be along and nerve wracking night.

Myself? I don’t anticipate Obama losing. But that won’t stop me from following the results obsessively on election night, and worrying at the first sign of “trouble.” Obama might not be perfect (far from it in my opinion) but the election campaign has convinced me that a McCain victory would be pretty damn bad. If nothing else for what it says about Americans' judgment.


Anonymous Anonymous said...

I hope I'm not wrong but if there were a 24 hour news cycle in 1984 - the last few day would be filled with experts and pundits pointing out how Mondale could still win.

2:17 pm  

Post a Comment

Links to this post:

Create a Link

<< Home