Part of the problem is that there are so many polls. I remember the 1980 race between Reagan and Carter very well. It had a very similar dynamic. An incumbent President who had failed over his four years. A challenger who was painted as out of touch and radical. The polls showed a virtual dead heat going in to election day. Reagan won in a landslide. The difference was the single debate between Carter and Reagan a week before the election. The Gallup poll had Reagan down by 8 points a week before the debate. He won by 10 points.
There simply were not the same number of polls in that era. A voters tsunami was occurring after the debate but it was harder to pick up because of the fewer number of polls. I remember reading back in August of this year before the political conventions that more polls had been done on this year's Presidential race by that time than were done in total in the Reagan-Carter race.
The increased polling has given us more information but it has been shown time and again that more decision points do not necessarily mean better decision making. Human beings have cognitive limits on how they can process and use information. Simpler is usually better.
With that in mind I attempted to take a step back from the noise and reconsider what I said in my blog post in mid-September "50 Days and Counting" on how I thought this election would be determined.
50 days from now this election is going to be determined by three big factors.
1. The Presidential debate. Can Romney convince voters that he is change they can believe in?
2. Will Republican voters be energized to vote like they were in 2010?
3. Can Romney win the Independent vote?Governor Romney answered the first question in the debate of October 3.
The second question really involves turnout. Bear in mind that when Obama won in 2008 that voters who identified with Democrats were 39% of those who went to the polls. Republicans only made up 32% of the total. That 7% margin in party identification made a huge difference. Of course, many of the polls we are seeing are assuming similar turnout models this year. Whether that will occur is still a big question.
The question of whether Romney wins the Independent vote also looms large. Independents made up 29% of the voters in 2008 and they favored Obama 52%-44% last time. Hard core Democrats will not turn from Obama. Therefore, for Romney to win he must turnover other Obama voters from 2008 to his side. The bulk of these will be Independent voters or very soft Democrats.
Therefore, the election on Tuesday is going to be determined by two simple numbers.
Turnover-the % of Obama voters from 2008 who will switch to Romney in 2012
Turnout-the % increase in the Republican/Democrat turnout margin at the polls
Simply stated, Turnover + Turnout should predict the final popular voting margin in this race. The chart below that I created roughly indicates the popular vote margin for Romney at improving levels of turnover and turnout compared to 2008. Looking at the chart below, if no Obama voters (0% turnover) switch to Romney and Democratic/Republican turnout percentages remain the same (0% turnout) this election should result in a -7.4% voting margin for Romney all other things being equal. That was the result in 2008.
The horizontal yellow axis reflects the potential "net" turnover from Obama to Romney. The recent Washington Post/ ABC News poll showed that 13% of Obama voters plan to vote for Romney. A recent Reuters poll showed that 5% of McCain supporters from 2008 (I know, this is hard to believe) planned to vote for Obama in 2012. That would be a net turnover of 8% from Obama to Romney.
If that level of turnover should occur Romney will win the popular vote unless Democrat turnout exceeds Republican turnout by over seven percentage points (the advantage the Democrats had in 2008). The vertical green axis reflects the change in the turnover margin. Historically, the Republicans have lagged Democrats in turnout by two or three percentage points, not the seven that we saw in 2008. On the chart above that would be turnout improvement at the 4% or 5% level. That means that Romney needs to turnover Obama voters at the 5%-6% level to win.
A number of polls have been criticized as they are assuming a Democrat margin of eight or nine points. That would be -1% or -2% on the green axis. As you can see above, if that is reality, Obama will likely win as almost no amount of turnover of Independent voters will save Romney. Romney need s
The blue color in the chart above shows areas where Obama is likely to win the popular vote. The red is Romney country. If either candidate wins the popular vote by at least 1% it is unlikely that they would not also win the electoral college based on most simulations I have seen. The purple color is where the popular vote margin is within 1%. This is an area where we could see a split between the popular vote and electoral college. Remember the 2000 election where Al Gore beat George W. Bush by .49% in the popular vote?
I don't pretent that the grid above has any predictive power. It merely demonstrates the importance of both turnover and turnout in Tuesday's election. Romney needs both to win.
Ohio
I live in Ohio and there is clearly a lot of interest in my state right now. I have been looking at recent polls in Ohio as well as studying past election results. I wrote about Ohio previously in a prior post,
The Columbus Dispatch released a poll this morning that has Obama leading Romney 50%-48%. I was particularly interested to see if they had breakdowns on Obama and McCain voters from 2008 and how they say they are going to vote in 2012. In that poll, 11% of Obama voters in 2008 say they are going to vote for Romney. 5% of McCain voters in 2008 say they are going to vote for Obama. That is a net turnover of 6% favoring Romney. Let's apply that to the 2008 results.
| (in millions) | Obama | McCain/Romney |
| 2008 Ohio Results Baseline Vote | 2,940 | 2,678 |
| 11% of Obama voters switch to Romney | -323 | 323 |
| 5% of McCain voters switch to Obama | 134 | -134 |
| 2012 Projected Votes Based on Turnover | 2,751 | 2,867 |
On the turnover factor alone it appears that Romney can defeat Obama. However, the Columbus Dispatch poll shows Obama leading Romney by 2%. The poll results seem to suggest that Democrat turnout will be close to what it was in 2008 with new voters and Independents also tilting toward Obama. We will see. The poll has a margin of error of 2.5%.
However, I could not help but notice the numbers when I compared the projection above to the actual vote totals in 2004. Unbelievable! I could not make this up if I tried.
The projection above is within 10,000 votes of Kerry's total in 2004 and 8,000 of the Bush total that year.
It is interesting that the total vote in Ohio for the two major party candidates in 2004 (5,600,000 total votes) and 2008 (5,618,000) was almost identical even though voting nationally was up approximately 7%. Cuyahoga County, Ohio's largest county which also is heavily Democratic, actually had 12,000 fewer votes cast in 2008 than in 2004 despite Obamamania. Turnout there will be critical in determining Ohio's results in 2012. Romney has to lower the Obama margin there to win. That is something that McCain could not do.
(in millions) Kerry Bush
2004 Ohio Election Results 2,741 2,859
It looks to be a real barn burner. If anyone tells you they know what is going to happen they are kidding themselves and kidding you. This is a year that some people are going to be either really right or really wrong in their prediction. I will make my prediction tomorrow. I kid you not.
No comments:
Post a Comment