Swing States Favor Obama

Six weeks before election day, a host of recent, state level opinion polls favor Obama over Romney in November 2012 Presidential race, despite a less than stellar economy (see also here). Those states, and the number of percentage points by which they favor Obama, are as follows:

Colorado (3 and 6),
Florida (-1 and 1 and 5),
Iowa (7),
Michigan (9 and 12),
Minnesota (8 and 9),
Nevada (7),
North Carolina (2 and 4),
Ohio (4 and 5),
Pennsylvania (2 and 8 and 12),
Wisconsin (12) (where GOP Vice Presidential nominee Paul Ryan was elected to Congress), and
Virginia (3).

In Nebraska, which allocated some of its electoral votes by Congressional district, rather than statewide, there is one Congressional district (the second) where Obama and Romney are tied.

Among the states that are safely for Obama are California, Connecticut, Illinois, Maryland, Massachusetts (where Romney served as Governor), New Jersey, New Mexico, New York and Washington State. (Washington D.C. is a safe win for Obama as well.)

Among the states where Romney leads are Arizona (3 and 10), Georgia (6), Indiana (6), Missouri (6), South Dakota (15), Tennessee (7), and Texas (11).

Many of the Obama leads in these swing states are not so great that they couldn't be reversed in a month and a half, and Presidential polls in one state are not independent of Presidential polls in another. But, Romney needs almost all of them to win 270 votes in the electoral college and hence to win the Presidency. One of the last natural turning points in the campaign for either candidate is the October 3, 2012 Presidential debate which will be held at the University of Denver.

If the Presidential election were held today, Obama would win. With a seven percentage point shift in Obama's favor (e.g. winning Virginia by a 10 percentage point margin instead of a 3 percentage point margin as current polls indicate), Obama would win by a landslide. With a seven percentage point shift in Romney's favor, he would win, but not by a landslide. Romney couldn't win the Presidential election even with a three percentage point shift in his favor from current polling in swing states. Romney needs more than Florida, North Carolina and Virginia to win.

Even a really notable "October surprise" would be unlikely to swing current polls much more than seven percentage points either way in the nation's swing states, and usually late breaking news has something less than a seven percentage point impact on polling. The closer we get to election day, the less likely it is that new information will matter. By late October, many people will have already cast their ballots via mail in voting and early voting.

National polling shows a closer race (five recent national polls favor Obama by two to six percentage points, while a sixth national poll is tied), but the race isn't decided by the national popular vote. It is decided by electoral votes in swing states. A surplus of support in states that aren't close doesn't matter.