Thursday, November 1, 2012

No Change in Presidential Race, Democratic Prospects Brighten in Senate But Not in House, The Voter Turnout Debate Rages On

First Polls Open In: 4 Days, 9 Hours
Projected Popular Vote Total: Obama +0.7% (up 0.1% from yesterday)
Projected Electoral Vote Total: Obama 290, Romney 248 (unchanged from yesterday)
Current Betting Odds: Obama 66%, Romney 34% (Romney up 1% from yesterday)
Current Popular Vote Betting Odds: Obama 57%, Romney 39%, within 0.5% - 4%

Not much has changed in the Presidential picture in the past 24 hours, other than Mitt Romney has 24 less hours to close what is, in my analysis, an electoral gap.  The news cycle continues to be dominated by hurricane coverage with only passing mentions of politics.  I expect the election coverage to pick up in earnest this weekend.

We are technically past the point of an October surprise - it is November, after all, but not too late for a late-breaking revelation, although one seems highly unlikely given that:
a. Obama has been in office for 4 years and has been vetted for smoking guns by the GOP as well as the Donald Trumps of the world and no smoking gun has been found.
b. Mitt Romney has essentially been running for President for at least 6 years and has been thoroughly investigated by two slates of GOP candidates without a smoking gun.

Given the relative stability of the Presidential race, I thought I'd devote some time to the races that I haven't been covering enough - the down ticket federal races for the U.S. Senate and House of Representatives.

In the Senate, Democratic prospects have improved significantly over the past several months.  Races that many, including myself, once believed would be easier wins for the GOP are now competitive, while Democrats have largely held safe the seats I believed would be safe.

Democrats have 30 seats that are not up for re-election and 13 races that I consider safe (>10% lead), giving them an effective starting point of 43 seats.

Additionally, there are two seats where an Independent will safely win who will likely caucus with the Democrats - Socialist Bernie Sanders will easily hold on for another term and former Governor Angus King is a lock to win in Maine and is believed to be headed to the Democratic Caucus.

On the GOP side, they have 36 seats not up and 5 races I consider safe.

Here is the polling averages associated with the remaining races and the build to my projection:

My current projection is for 52 Democrats, 46 Republicans and 2 Independents in the new Senate, or an effective working majority of 54-46 for the Democrats.

Of course, there are a number of exceptionally close races, most notably in Virginia, Montana, Wisconsin and Arizona.  And the polling in many Senate races is not nearly as broad as in the Presidential race, so the margin of error in these projections will be higher.

In the reasonable best case for the GOP, if they pick up all 4 of those ultra-close races, we would be looking at 49 Democrats, 49 Republicans and 2 Independents, which would still give Democrats a 51-49 working majority.  Basically, the tipping point races for control would be Ohio and Missouri.  The GOP would need both if Obama wins re-election and Joe Biden becomes the tie-breaker in a 50-50 Senate.  They would need only 1 of those 2 if Mitt Romney wins and Paul Ryan is the tie-breaker.

In the House, Democratic prospects are not nearly as bright.  Current aggregation of generic polling indicates that the GOP is at +0.5%, which implies a GOP majority of 243-192.  Because of the lack of polling data, I do not project individual House races, but the generic polling contrasted against the composition of House districts typically yields a pretty good proxy in aggregate.  The GOP have been huge beneficiaries of the last redistricting cycle.  Because of the shape and demographics of the new House maps, which were largely drawn by GOP legislatures because of the drubbing that Democrats took in 2010 in state houses, a national even split of the vote (that is if exactly 50% of people voted for each party, approximately in line with their demographics) would produce approximately 238 GOP House seats.  To get to a House majority, Democrats would need to be about +2% in the national vote, a margin they don't appear to be near as the election nears.

Turnout Models Reveal This Simple Truth - Turnout Determines All Close Elections
I can't remember a Presidential election where the question of voter turnout and the composition of the electorate has been the subject of so much debate in polling.

Gallup and the National Journal aren't getting broadly divergent polling results because they are talking to different types of voters - they are largely getting divergent results because they disagree on who will actually show up to vote.  I use these two as the two extreme in national polling - the last Gallup poll had Romney up 5%, the last National Journal poll had Obama up 5%.

Who is going to turn out in an election is the hardest thing that a pollster has to determine.  Asking people how likely they are to vote is often an unreliable barometer, typically many more people answering a survey SAY they will show up to vote than ACTUALLY do.

Leveraging history is tricky, and here is the rub between the two polls and competing schools of thought.

Gallup's turnout model looks largely like a 2004 model.  Gallup would argue (and the GOP talking heads would agree) that 2008 had a unique set of circumstances that drove up Democratic turnout.  Youth turnout surged to an all-time high.  Turnout from both blacks and Hispanics was higher than in any previous election - and not by a small margin.  Hope and change was in the air and Democrats were fired up.  It is certainly fair to argue that they are less fired up today.

The National Journal has a model that looks more like 2008.  They would argue that while there may be some dampened enthusiasm in some demographics - such as the youth vote, the black vote will show up for Obama in the end and the Hispanic population has grown significantly, meaning that even if Hispanic turnout is down, overall Hispanic share of the electorate will remain roughly flat.

Knowing who is right is very difficult, since, as I said, there are no highly reliable ways to know.

My belief is that the truth is somewhere in between.  The point of using models, as I do, that average and aggregate polls in multiple ways, is that meeting in the middle of the sets of assumptions from experts generally produces a far more accurate result than picking a model on either extreme because you like it.

We will know in a few days who is right.  But there are two things that this turnout debate makes crystal clear:
(1) Watching voter turnout on election day will give us a great indication as to who has won the election
(2) Get out and vote!  Whichever candidate you like, whether you and others like you turn out or not will determine whether he wins or not.

If you like this site, tell your friends.

No comments: