Electoral College: Reversion to the Mean

Two states change categories today. One moves toward Obama, the other moves toward Romney. Lets take a look:

Last Friday the five poll average moved to a 5.2% Obama lead in Nevada. With today’s update, it drops back down to 4.6%. With that I put Nevada back in the list of “Lean Obama” close states. In the last year, Romney has NEVER held a lead in Nevada in the five poll average. Winning Nevada would be a stretch for him. But it is once again somewhat close. There ARE situations you could imagine, even with less than two weeks to go, that would erase Obama’s advantage in Nevada. As we get closer though, a 4.6% lead looks a lot bigger than it might have a month ago.

After a brief dip into Romney territory in the middle of October, Virginia returns to where it has been for most of the last year… a small Obama lead. As of now, the five poll average has Obama with a 1.0% lead in Virginia. This moves Virginia back to Obama’s column. The lead is narrow, this could change, but the movement lately has been toward Obama.

If you look specifically at the state charts such as the two above, this is actually the picture in most of the states that are in contention. You see an Obama decline starting about a week before the first debate, which accelerates into more of an implosion after the first debate. That decline hits bottom sometime between the VP debate and the second presidential debate, and then Obama has been trending upward ever since, but is still quite a long way from his highs at the end of September.

Now where does this put our summary?

That is right. Our old friend is back. Obama winning by 332 to 206. This is where the “Current” line has been more often than any other position. If you look at the chart showing the history of the race for the entire year you can see that the most time has been spent here, and while there have been occasional movements in both directions away, time after time things have just reverted to 332 to 206. This just seems to be where this race “wants” to be. This seems to be the “mean” that the race always reverts back to after one event or another moves it away.

This looks like a pretty strong position for Obama, right? Romney is only ahead in ONE of the close states at the moment. He has to move several back over to his side to win. This does not look great for Romney.

Well, no, this isn’t great for Romney. His situation is considerably worse than it was right before the second debate. He clearly peaked and is now declining again… at least in terms of the electoral college. There have been some people saying that they are seeing Romney momentum right now. Uh, no.

But… it is still close. Time to look again at the individual close states:

• North Carolina (15 ev) – 2.0% Romney lead – 0/5 polls after last debate
• Florida (29 ev) – 0.6% Obama lead – 0/5 polls after last debate
• Virginia (13 ev) – 1.0% Obama lead – 0/5 polls after last debate
• New Hampshire (4 ev) – 1.0% Obama lead – 1/5 polls after last debate
• Colorado (9 ev) – 1.2% Obama lead – 0/5 polls after last debate
• Ohio (18 ev) – 1.6% Obama lead – 1/5 polls after last debate
• Iowa (6 ev) – 2.2% Obama lead – 0/5 polls after last debate
• Wisconsin (10 ev) – 3.5% Obama lead – 0/5 polls after last debate
• Nevada (6 ev) – 4.6% Obama lead – 0/5 polls after last debate

We still have Romney with 191 electoral votes without any close states. Add in the one close state he is ahead in, North Carolina, and you get to 206 electoral votes. That is 63 electoral votes short. In yesterday’s update if you moved all states toward Romney by more than 1.2%, you would have a Romney win. The tipping point was Ohio. The tipping point is still Ohio.

The easiest way for Romney to win is to hold North Carolina, then pull Florida, Virginia, New Hampshire, Colorado… and Ohio… to his side of the fence. This is looking like a much harder job than it was before. More states are slipping away from Romney… but they are still close… they are not out of reach.

Not as close as they were though. Obama’s lead in Ohio is increasing. It is now 1.6%. But 1.6% is still tiny. You can easily imagine events that would cause a 1.7% or 2.0% move in the polls across all the close states. If Obama pulls ahead by a bit more in Ohio, then maybe we can say he starts to look like a clear winner. But at the moment, it is still close.

Obama is ahead. If the election was held today, he would be the favorite by a decent margin. As of today, FiveThirtyEight has Obama at a 68.1% chance of winning. I’ve seen higher estimates elsewhere, up to 93.4%. My gut says it is somewhere in between. Maybe 80%.

But there is another interesting possibility percolating that I’ll mention before I close. For the last two weeks or so, most of the places doing poll averaging of national polls have shown Romney ahead. I don’t believe ANY of the sites that I track that are doing analysis of the electoral college have shown Romney ahead at any point. Obama has always been ahead.

If this situation persists through election day, we may have a repeat of 2000, but with the parties reversed and the key state being Ohio instead of Florida. Romney wins the popular vote, but Obama wins the electoral college and therefore the presidency.

This kind of situation happens when you have something like Romney building up increasingly large margins in states he was going to win anyway, like Texas. These additional gains make no difference to the electoral college, but they do matter to the electoral vote.

I would find this extremely entertaining for no reason more than being able to watch all kinds of people in both parties suddenly have a completely opposite opinion on the value of the electoral college than they espoused 12 years previously.

This is still not necessarily an expected outcome… FiveThirtyEight puts the odds of the scenario above at 6.5%. (For completeness, they give 1.8% odds to the reverse scenario where Romney wins the electoral college and Obama wins the popular vote.) 6.5% isn’t exactly super likely, but it is nothing to sneeze at. If there was a 6.5% chance of a hurricane hitting your city, you would be starting to make preparations.

Meanwhile though, other analysts put these odds much higher… specifically, Sam Wang puts the chances of this at 25%.

I normally say you can basically ignore the popular vote except as an early indicator to moves in the states… but this is a situation where watching the popular vote is worth it. If we continue to see a popular vote / electoral vote split it will make for a very interesting election night.

Note: Chart and map from the Abulsme.com 2012 Electoral College Prediction page. Both assume Obama vs Romney with no strong third party candidate and show polling as it currently exists. Things will change before election day. On the map red is Romney, blue is Obama, gold states are too close to call. Lines on the chart represent how many more electoral votes a candidate would have than is needed to tie under several different scenarios. Up is good for Obama, Down is good for Romney.

Edit 2012 Oct 24 18:53 UTC to fix a few wording issues.