Part 1: Summary Statistics (last 7 days)
Obama vs Romney: 48-47% Romney (was 48-47% Obama)
President Obama job approval: 50-47% approve/disapprove (was 51-47% approve/disapprove)
Generic Congressional Vote: 45-43% Democratic (was 45-44% Democratic)
US Senate Projections: 46 Democrats, 45 Republicans, 1 Independent, 8 Tossups (was 47-44-1 Democrat)
Governor’s Race Projections: 30 Republicans, 17 Democrats, 1 Independent, 2 Tossups (was 30-17-1 Republican)
Part 2: Presidential Scorecard (270 electoral votes required to win): Obama 221, Romney 206, Undecided 111 (prior scorecard: 221-206 Obama)
(Note: blue = safe Democratic, light blue = leans Democratic, yellow = tossup, light red = leans Republicans, red =solid Republican)
There have been two debates (a “town hall” Presidential debate and a Vice Presidential debate) since the first Presidential debate, and while Barack Obama and Joe Biden were both animated and aggressive, it’s not apparent that undecided voters have been sufficiently impressed to return to the Democratic ticket. The electoral map therefore remained set in stone, with only one change necessary to our scorecard:
- Maine: “Solid Democratic” to “Leans Democratic”
One thing also worth noticing: for the first time, Mitt Romney has moved ahead in the national polling, despite the fact that Obama’s approval rating has remained at 50% for the past two weeks. Similarly, the US Senate races remain in flux, with 8 races (6 of which are held by Democrats) being “too close to call.”
There is one more Presidential debate this Monday. After that, the Presidential race becomes a turnout game both on Election Day and in the days preceding it. With a slew of critical states about to undergo early voting, the Presidential race is about to accelerate. Thus far, the bulk of the action has been with absentee and mail in ballots; even so, about 2.1 million votes have already been cast. That is over triple the number who had voted a week ago. Given that there is a time delay for statistics to be reported, it’s not a stretch to assume that 2.5 to 3 million votes have already been cast.
Part 3: What does it take for Romney to win?
We would like to independently analyze the Presidential election by using actual polling data from 2008 to provide the proper context for what we are seeing now. In other words, the last 7 days of polling at the statewide level don’t mean much in isolation unless you compare it against polling from the last week of the 2008 campaign. When that analysis is performed, here’s what we see:
If we take the polls in 2008 and 2012 at face value, it is apparent that President Obama has a 221-206 electoral vote lead. However, inclusion of 2008 polling data shows that he is running behind his 2008 poll showings by enough of a margin that we do not believe that this point that he could/would ultimately carry Colorado, Florida, Iowa, New Hampshire, and Virginia. In Electoral College terms, that means a 221-206 lead for Obama becomes a 267-221 Romney plurality – since last week, Virginia is now more likely to go to Romney.
That brings the epicenter of the election to Ohio. Obama leads there by a too close to call 48-46%. Depending on what you believe with regards to Democratic intensity (more on that later) and/or how the undecideds will break, we’re either talking about a 52% Romney victory or a 51% Obama victory. Assuming a similar turnout as 2008 (a questionable assumption, but we are making it in the absence of sufficient and timely early voting data), it would take, at a bare minimum, 77K more Romney votes in Ohio (or flipping of 39K Obama votes) for Romney to win with 285 electoral votes.
One caveat: it is true that the average of the last 7 days of polling in Wisconsin shows it “too close to call” with a 50-48% Obama lead. However, we are reluctant to predict that Romney would carry this state for two reasons: (1) Obama outperformed his 2008 polling numbers by 3 percentage points (his final poll average was 53-41%, while he carried the state with 56% of the vote), (2) because Wisconsin allows voters to register and vote on Election Day, we think that Democrats will use this law to their advantage in the Democratic strongholds of Milwaukee and Madison and carry the state, like they have in every election since the 1984 Reagan landslide.
Part 4: Early Voting Updates (as of October 18)
As this article is being written, absentee voting is underway in 47 states – tomorrow, Oregon and Washington commence their absentee balloting, while Pennsylvania on Tuesday becomes the last state to start accepting absentee ballots. In person early voting (the vehicle which helped President Obama carry several states) is now underway in 14 states. Critically, North Carolina begins its early voting today, and given the quality of early voting statistics available (the state maintains data on the early voter’s race and party), this will be the first state where we can adequately evaluate the intensity of Democratic “get out the vote” efforts relative to 2008. Here’s what we’ve seen so far:
North Carolina: 62K absentee ballots have been returned and accepted (since the beginning of October, it looks like about 4K returned ballots are counted per day). From those accepted ballots, the racial breakdown is 89-7% white/black and 54-27% Republican/Democratic. This is identical to the 2008 racial breakdown of 89-7% white/black and 54-28% Republican/Democrat. However, the real story will be told with in person early voting. In 2008, those voting before Election Day went for Obama 56-44%, while McCain carried the Election Day vote 57-41%.
So what are comparing against? The 2008 in person early vote was 67-29% white/black and 54/28% Democratic/Republican – these were the percentages it took for Obama to carry the state, and will therefore be our yardstick.
Iowa: Iowa has traditionally had a strong Democratic early vote. In fact, McCain carried the state 50-48% if you were to look only at Election Day voters; early voters supported Obama 60-38%. This time, Democrats got off to a strong start (nearly 70% of the first day’s voters were Democrats). This advantage has continually eroded over time, and a 57-25% Democratic lead from a week ago is now 50-30% Democratic, with 265K early votes cast so far. While this is still above the 46-29% Democratic edge from 2008, we are now about at the point where the Democratic advantage is equal to what it was in 2008.
Ohio: While meaningful results are tough to come by (absentee/early voting data is being compiled at the county level, and Ohio has 88 counties), we do know that 492K early votes have been cast (a quadrupling in a week’s time). In counties that voted Democratic in 2008 and 2010, 10% of the 2008 vote has already been cast. That figure is 8% in Republican counties and 7% in “swing” counties. It’s also worth noting, however, that data in predominately Democratic Cleveland and Columbus is updated on a daily basis, while in many Republican counties, we are lucky to get weekly updates. In fact, the 53 counties where data is a week out of date voted 56% for McCain – about 8% more Republican than the statewide average.
Florida: Currently, only absentee ballots are being accepted (in person early voting doesn’t start until October 27). We know so far that 464K absentee voters have been cast, and that vote is 45-40% Republican by voter registration (statewide voter registration is 41-37% Democratic). The media has made much of the fact that Republicans had a larger voter registration edge with absentee voters in 2008, although it’s also true that the Obama campaign is encouraging more of its supporters to vote by mail this time, instead of voting “in person.” We’ll have a better assessment when in person early voting starts.
Looking ahead, the next seven days will essentially decide who the next President will be. 14 states (including Louisiana this upcoming Tuesday) will start in person early voting, and four of those 14 states (Colorado, Nevada, North Carolina, and Wisconsin) are swing states.