Tuesday, November 06, 2012

The United States votes

For a political junkie, there is no more exciting event than that which falls on the first Tuesday after the first Monday of November in a leap year. Happy U.S. election day!  

Things to Watch

As any reader of Nate Silver's blog FiveThirtyEight knows, this election all comes down to Ohio. Remarkably, according to his analysis, Ohio has a better chance of deciding the winner than all other states combined.

This is somewhat unprecedented in modern U.S. electoral history. While Ohio was the deciding factor in 2004 as well, there was a clear alternative path (through Iowa, New Mexico and Nevada) for John Kerry to beat George W. Bush without Ohio. In 2000, while the focus is now on the incredibly close race in Florida, Al Gore would have been president while losing Florida had he carried New Hampshire (which he lost by just 7,000 votes) or his home state of Tennessee (which he and Clinton had won together twice and which he largely ignored and took for granted before losing it by only 80,000 votes out of 2,000,000 cast).

Were Obama to lose Ohio, Silver says he has just an 8% chance of winning the election. The most likely way for him to do this would be to win all of the Democratic-leaning swing states (Nevada, Iowa, Wisconsin and New Hampshire) and then win one of the Republican-leaning ones (Colorado, Virginia, North Carolina or Florida).

Were Romney to lose Ohio, Silver says he has just a 3% chance of winning the election and the path is more complex. He would have to largely run the table in the swing states, or win a state such as Pennsylvania, Minnesota or Michigan that his campaign has been making late plays for but which the polls and state dynamics say should be out of his reach unless he's winning in places like Ohio anyway.

(You can game out your own electoral scenarios at 270towin.com - other outfits have similar maps but this is the most simple to use.)

So, if Ohio is so critical, isn't that the only thing to watch? In a way, yes. But the polls in Ohio don't close until 8:30 p.m. Atlantic time and even if Obama or Romney exceed expectations that means the result won't be called until at least 10:30 - and if it is as close as expected, maybe not until the wee hours or several days down the road.

So I've looked at when various states close their polling stations to find early clues (canaries in coal mines if you will) as to what the result may be.

(Check out POLITICO's handy state poll closing map here.)

1. Indiana and Kentucky

At 7 p.m. Atlantic time the first polls close in eastern parts of Indiana and Kentucky. A good measure might be whether Romney wins Indiana in a by a big enough margin for the networks to call it immediately on the basis of exit polls alone (which would mean 15+ points, indicating a better than expected night for him). Unfortunately, they won't call the state until the western most polls close at 8 p.m., so we'll have to look for something more obscure.

Kentucky behaves in an abnormal fashion as part of Appalachia, and is not a good proxy for results outside of its region.

Certain counties in Indiana, however, may be able to tell us something. In particular, I would point you to Delaware and Madison counties whose polls close at 6 p.m. (EDIT: that's 6 local, 7 Atlantic) In 2008, Barack Obama became the first Democrat to carry these counties since Bill Clinton did in 1996. Typically, a Democrat does about 2 to 4 points worse than their national vote in these counties. (In 2008, Obama actually met and exceeded his national margin in these areas, but that was because he had invested resources in Indiana, unlike previous Democrats and unlike himself in this cycle). If Obama is winning these counties, it means he is probably winning the national popular vote by 3+ points and has the election in the bag. If he is losing them by a narrow margin, the national popular vote is probably a toss up. If he is losing them by more than 5 points, Mitt Romney is likely the next president of the United States.

2. Florida and Virginia

These two swing states close at 8 p.m. Atlantic, along with a number of other non-competitive states (Indiana, Kentucky, Georgia, South Carolina and Vermont).

Romney needs Florida to win. If he loses Florida, an unlikely complete sweep of the remaining swing states would give him a narrow 270-268 victory, however his campaign has already condeded Nevada, meaning that a sweep of the balance would mean a loss of 266-272.

So watch Florida, if Obama wins, it's all over but the crying. If Romney wins, he's still in the hunt.

Virginia falls into a different category. It is a critical piece of the electoral calculus for both candidates. Without Virginia, there is no realistic chance for a path to victory unless they win Ohio. It means absolutely no margin for error.

Obama's non-Ohio path runs through Virginia with wins in Nevada, Iowa and Wisconsin. Obama also has a narrow path without either Ohio or Virginia, involving winning Colorado which seems unlikely as it is more Republican-leaning than these states. Romney's non-Ohio path requires winning Virginia, the smaller swing states, and Wisconsin, which most see as a tall order.

3. North Carolina and Ohio

At 8:30 p.m. these states close. As we've discussed before, the winner of Ohio is the very likely winner of the election. As for North Carolina, Romney is favoured here. If Obama wins, then I suspect he has also won Florida and is headed to a large blow out similar to that which he scored over McCain in 2008.

4. A Romney landslide?

By 9 p.m. Atlantic, Ohio's polls will have been closed for a half hour but there is almost no way we will know the winner yet. At this hour, three states close that should favour Obama. If Romney wins any of them, he has won the election - most likely with over 300 electoral votes. These are Pennsylvania, Michigan and the second congressional district of Maine, which borders New Brunswick.


I made a prediction back on Oct. 25 as follows:

I stand by that prediction. However, were I to hedge, it would be in the direction of Obama winning by a larger margin. The polls seem to all be moving toward Obama to the tune of about one percentage point since Hurricane Sandy. If that movement continues, the last bit of it won't be captured by the polls and Obama may win by a large-ish margin. In fact, I would now wager that my above prediction of Obama is a low water mark for him. I see Obama's very high water mark at about 379 electoral votes. While I find it possible but unlikely that Obama would break 340 electoral votes, were this to occur, I suspect Hurricane Sandy would get the credit and go down in history as one of the great "October surprises".

The Senate and the House

The Democrats have gotten very, very lucky in the Senate races in Indiana and Missouri which they should have lost. My prediction is 53 Democrats + Angus King. If King aligns with the Democrats, the would mean a net gain of one seat for them. This is a bit more bullish than most prognosticators who have them at 49-52 seats + King.

I've not followed the House race as closely but it seems certain the Republicans will hold their majority with a gain or loss of up to 5 seats.