Nervous nellies on the Right have their panties in a bunch over a couple stories I won’t link to about alleged Romney people lamenting how strong Obama is and how there is no way Romney can win. This story has the same ring of truth like the Politico story with 36 anonymous GOP insiders lamenting how selecting Paul Ryan lost the election. Each of these stories has no basis in reality but are definitely planted stories by major advocates of one political campaign — guess which one? So instead of letting your friends grasp their blankey a little tighter, let’s take a reality check on repeatedly debunked stories on Team Obama’s irrational over-confidence, changing polls, and the electoral college.
I’ve blogged a few times about the seemingly absurd level of confidence emanating from senior Obama campaign officials. I have no problem with them believing in their candidate but they genuinely do not believe they can lose (which is a fatal flaw for anyone in a dogged election battle). Polling in the race remains air tight both nationally and among the Battleground States. The Real Clear Politics average of polling is invariably the most accurate gauge of any state race or the national contest. So I decided to assess the electoral map coming out of the Conventions.
As we have pointed out over-and-over again, the vast majority of this polling has been of both registered voters and for a voter turnout quite similar to the 2008 Democrat wave election when party ID shifted from dead even in 2004 to a seven point advantage for Democrats. Both of those factors skew heavily Democrat which when accounted for tell a very different story than the “sky is falling” crowd.
Nate Silver of the New York Times 538 blog analyzed registered voter polls versus likely voter polls in 2010 and found registered voter polls produced results 4-points higher in favor of Democrats than the far more accurate polls of likely voters. 2010 was a strong Republican year and without Obama on the ballot it is understandable why the 2010 figures don’t perfectly apply to 2012 polls. For these reasons a 4-point adjustment seems a bit much with a Presidential ballot likely to increase Democrat enthusiasm to vote versus off-year elections. As such, for my own purposes, I adjust registered voter polls 2-points in favor of the Republican to reflect the quantified differences between registered voter polls and the more accurate likely voter surveys.
The shift from registered voter surveys to likely voter surveys had the expected impact and, as of today, President Obama leads five of the Battlegrounds by ~1% easily within the margin of error for every survey: Iowa +0.2, Virginia +0.6, Ohio +0.7 (ex-PPP poll because how many times do you need to see their bias before you ignore them too), Wisconsin +1.4, Florida +1.7. And if I backed-out PPP from every state, Romney would be even closer if not winning some of these states but let’s move on.
This is where it gets really interesting. A two-point swing would move each of the above mentioned states into Romney’s column and hand him the Presidency without winning Colorado, Nevada, Michigan or New Hampshire (let alone even considering Pennsylvania).
Party Identification in Polls
This has been a sticking point throughout this election season. The typical polling service constructs and adjusts its polls to reflect the electorate as it was on election day in November 2008. They do this for consistency sake but not for accuracy sake. In an exhaustive post worth everyone’s read anytime this topic comes up, Jay Cost at the Weekly Standard completely debunked anyone believing the current polling reflects the electorate who will show up on election day in November. His bottom line is that the election in 2008 had a legitimate partisan advantage of more self-identified Democrats showing up at the polls than Republicans by 7 percentage points — a great showing for Democrats. Today, his highly educated opinion is that Democrats advantage will shrink to only 3 percentage points (read his post for the well argued explanation). I personally think the advantage will be no greater than 2 percentage points but no matter. The key is the big drop in democrat enthusiasm and self-identification that is true in every single poll no matter how biased the source.
With that, if we see only a 2% shift towards Republicans and not even the 4-5% shift Cost and myself believes will occur Romney wins Iowa, Virginia, Ohio, Wisconsin and Florida. That’s based on the electorate right now. Before Romney’s ~$200 million war chest gets unleashed. Before Obama has to debate and defend economic policies with sub-2% GDP, 23 million unemployed, and a $16 trillion deficit. Before the real battle for undecideds begins (which is right now). Therefore, with just the slightest adjustment for a voter identification turnout in November that will certainly be more favorable to Romney than current polling reflects, the electoral map looks like the following: