Saturday, September 20, 2008

The Six States To Election Victory

Recent polling that has come out in the last two days has capped off quite an impressive shift in both the national and state polls for the presidential election. In essence all the gains that McCain made after the republican convention have been wiped out and even more impressive, Obama is leading by larger margins in most swing states than he was before either convention. 

While this may cause many democrats to celebrate and republicans to be concerned, this interpretation of the polls is not quite accurate. If one inspects the polls carefully, it is easy to understand why we are seeing such movement in polls trying to project electoral vote counts. 

The first aspect to remember is that the president is not elected by a national popular vote. Thus, national polls that do not control for likely voter location are not incredibly useful in determining likely electoral vote outcomes beyond overarching trends. Since most states are winner take all systems, as soon as a candidate gets 50%+1, all of the extra votes a candidate gets in that state are wasted. So for example, in New York, even though Obama will likely win the state with about 70% of the vote, 19% of this vote is wasted. Thus, when national polling is done, it is hard to weed out this wasted vote effect. In my opinion, national polling is pretty worthless unless you have a candidate up by more than 15 percentage points. 

So, since the president is elected via the electoral collage, when trying to project the outcome of the race, the key statistic to look at is how candidates are doing in individual states. Luckily for us, most polling outfits agree with me and this election marks an unprecedented amount of state polling. 

Before we go to the analysis of the seesawing projections of who is going to win the election, it is important to remember there are 538 electoral votes and that to win the presidency, a candidate needs 270.

So with our electoral math hat on, let's take a look at state polling numbers.  The most striking result is that of the 538 electoral votes, about 459 (or 85%) electoral votes are immovable. These are electoral votes from states that are strong supporters of the respective candidates, and have showed no signs of moving since June, with very high poll numbers in favor of a respective candidate by at least 12 to 15 percentage points. States in this category include NY, CA, TX, AL, GA, etc. When we break down these 459 electoral votes by candidate, we find that Obama has 259 electoral votes, while McCain only has 200. 

This is where things get interesting. The remaining 79 electoral votes are split between only six states: CO-9, FL-27, NM-5, NV-5, OH-20, and VA-13. Every time one of these states moves back and forth between the candidates, it drastically affects who is projected to win the presidency. Hence, the seesawing that we are getting in the polls and projections is not because the nation as a whole is really changing its level of support for the different candidates, but rather, is the result of small movements in these six states where split between support for both candidates is very close. 

The net result of this is to interpret polls that keep showing drastic shifts in the outcome not as drastic shifts in levels of supports for the candidates, but rather as an artifact of how we elect a president in this country.  Very small amounts of support shifting in very few states can have a drastic impact on the outcome, which is what the polls are picking up. Thus, if Obama falls behind by a large margin in a outcome projection, it is not because he has lost vast amounts of support, but rather that a very small number of people in one of the six states has sifted their support. 

This volatility is likely to continue until actual election day because the numbers needed to switch one of these six states to either candidate is around 1-3%. 

Now, an astute mathematician who has been following the math may have noted that based on the safe electoral vote count, Obama is in a much stronger position since he has to win far fewer of the six swing states than McCain does (Obama only needs 11 more electoral votes to get to 270, compared to McCain's 70). I think that this line of reasoning is spot on and is why you have see Obama and Biden shifting away somewhat from their 50 state strategy to focusing on these six key states in this critical period of the election. For those that want to know the current break down in these six states, Obama leads in NM and CO, while McCain leads in VA, OH, FL, and NV. Under that scenario, Obama would win with 271 electoral votes. 

No comments: