Can Obama Win?

August 11th, 2012 by Lee Eldridge

Today we’re going to take a look at the upcoming presidential election from a different angle. We’re not going to talk about Paul Ryan, political agendas, fundraising or the campaigns. We’re not going to talk about personalities or character. We’re not even going to talk about the two candidates. Today, we’re just going to look at recent elections, supporting data and how understanding the past helps us to predict the future.

Reagan and Carter

Reagan vs Carter
We’re going to start with something a little more anecdotal. In the summer of 1980, the polls had President Carter beating challenger Ronald Reagan. These same polls showed ex-President Ford beating Carter. The republican establishment was concerned. They wanted Reagan to pick Ford as his vice president. Basically a co-president. Reagan declined and picked the first George Bush. Reagan ran on a platform of conservative values and American greatness. He went on to beat Carter. In hindsight, did Carter have a chance? The economy was bad. Unemployment was rising. Interest rates were high. People were concerned about the future. This is not the type of atmosphere where voters reward a president with four more years.

Unemployment
A friend asked me about a year ago if Obama had a chance at winning. My initial response was that it all depends on the unemployment numbers. An oversimplification, but still a valid point. Presidents don’t get reelected with high unemployment. I commented that if unemployment was above 8.5%, that Obama would lose. And if unemployment fell below 8%, he could make the case that his policies were working. Unemployment sits at about 8.25%. But that’s just the official number. Real unemployment is much higher — closer to 10.6%. Millions have been fleeing the workforce. The economy is flat. It’s been the movement from people leaving the workforce that has moved the official unemployment rate down from 10.2%. I wrote about this here. And you can read more about this from James Pethokoukis here.

Here’s a portion of James’ article:

Only in a world of lowered, New Normal expectations was the July jobs report anything less than another disaster for U.S. workers.

– This continues to be the longest stretch of 8% or higher unemployment since the Great Depression, 42 straight months.

– If the labor force participation rate was the same as when Obama took office in January 2009, the unemployment rate would be 11.0%.

– Even if you take into account that the LFP should be declining as America ages, the unemployment rate would be 10.6%.

– If the labor force participation rate hadn’t declined since just last month, the unemployment rate would have risen to 8.4%.

– The broader U-6 unemployment rate, which includes “all persons marginally attached to the labor force, plus total employed part time for economic reasons,” ticked up to 15.0%.

Incumbents Lose the Undecided Vote
In presidential elections, the undecideds rarely vote for the incumbent. Why? Because if you haven’t made up your mind about the president after four years, you’re probably not going to vote for him. In elections where the incumbents have won (such as Reagan in 1984 or Clinton 1996) or when they’ve lost (such as Carter in 1976 or Bush in 1992), approximately 80% of the undecideds vote for the challenger. So if you see a poll that shows Romney and Obama each at 45%, what you really have is a Romney win of 53-47. Polls typically show 3-5% of people supporting “some other candidate”, but most will end up voting for one of the two primary candidates anyway. In the last two presidential elections, only about 1% of the vote has been cast for a third party candidate.

Likely Voters
One of the keys to look for in the polls are if they’ve polled likely voters. Some poll adults. Some poll registered votes. When polling adults or registered voters, the results tend to favor the democrats. But polling likely voters tends to be the most accurate predictor of elections. Rasmussen Reports polls likely voters, and as of today, they have Romney leading 47-43. And if you follow the 80% rule from above, then you’ve got a Romney win of 55-45.

Barone Table 2

Mid-Term House Elections Predict Presidential Elections
This is an interesting chart comparing mid-term elections in the House of Representatives to the following presidential elections and House elections. Notice how the House votes in 1998, 2002 and 2006 almost perfectly predict the results in 2000, 2004 and 2008. If this trend holds true, then Romney will beat Obama by something close to the seven-point margin that republicans beat democrats in the House in 2010.

Approval Rating
President’s don’t typically win reelection when their approval rating is below 50%. One exception was George W. Bush who had an approval rating of 48% with registered voters in 2004, but was above 50% with likely voters. As of today, Rasmussen has Obama’s approval/disapproval ratings at 45-53, and Gallup has them at 43-51.

Over Sampling of Democrats
This has been an interesting phenomenon this year. In 2008, Obama beat McCain by seven points, and democrats outvoted republicans by seven points. But as we’ve seen, the mid-term election is probably a better predictor of turnout than the past presidential election. And in 2010, republicans and democrats voted in fairly equal numbers. So you would think that pollsters would weight their polls according to expected participation. But they do not. Many of the national polls have been sampling democrats by anywhere from 6-19 points higher than republicans. And they’re not weighting the results. With democratic enthusiasm down, it’s unlikely that democrats will outvote republicans by a significant margin. If at all. These polls are not likely to be indicative of how people will actually vote.

This is also going on in the state polls. For instance, a recent CBS/NYT poll showed Obama with a six point lead over Romney in Florida. In 2008, democrats outvoted republicans by 3 points in Florida. In 2010, they were an even split. So what’s a realistic expectation for 2012? Probably a slight advantage for democrats — maybe a point or two. But the CBS/NYT’s poll oversampled democrats by nine points. And they still only have Obama up by six.

The Future’s Not Bright
In just about every poll, Americans are pessimistic about the future. Economic confidence is down. People are pessimistic about jobs. The majority feel that the country is headed in the wrong direction. Similar to what Carter faced in 1980, this is a difficult atmosphere for an incumbent president.

Pethokoulis And More Stats
If you really want to read some more, check out this article from James Pethokoukis. He explains five stats and how Obama stacks up. All five show Obama with a tough mountain to climb.

Parting Thoughts
My prediction at this point is that the presidential election will feel similar to the recent Wisconsin recall election. Polls will be close till the end. Exit polling will show the election too close to call. Then in the end, it won’t be very close at all. I’m predicting a Romney win 52-47.

Tags:

13 Responses to “Can Obama Win?”

  1. Bobby Says:

    I hope you don’t think Romney is the next Reagan. But I do love the pick of Paul Ryan for VP!

  2. Lee Eldridge Says:

    It would be unfair to compare Romney to Reagan. At least unfair to compare him to Reagan in 1984. But Reagan in 1980? The left mocked him as an actor turned politician. He had been governor of a liberal state, and many would say not that successful as governor. Polls had him down to Carter just a couple months before the election. Reagan didn’t become “REAGAN” until probably 1984, when he won in a landslide.

  3. John 313 Says:

    So, you still think the rest of the country will join you and bobby and elect Mitt?

  4. Lee Eldridge Says:

    The rest of the country? Nope. Just predicting that 52% will vote for him.

  5. John 313 Says:

    Do you think Mitt will win the electoral college since the general election vote/poll is meaningless?

  6. Lee Eldridge Says:

    Would be tough to win 52% and not win the electoral college. I suppose that’s possible, but it’s not likely. And yes, I do think he’ll win the EC. Just about every state poll is sampling dems at higher rates than they voted for Obama in 2008, which means that the races are closer than shown in the polls. I don’t have the time or resources to go through every poll, but every poll that I HAVE gone through is sampling at rates equal to 2008 or higher. Dems outvoted Reps by 7 points in 2008. But in 2010 they vote was even. Yet polls are showing results with Dems outvoting Reps from 7-19 pts. Those numbers may be “accurate” to the poll results, but are not indicative of how the numbers will come out on election day.

    I think it’s much more likely that nationally the dems will outvote the reps by 2-3 pts. And even that relies on people being truthful during the exit polls, which if Wisconsin is any indication, they will not be truthful. Exit polls showed a very tight race for the recent recall election, yet Walker won by (I think) 5 points. I think we might see that exact same scenario play out in November.

    John, what is your prediction? For the popular vote and the EC?

  7. John 313 Says:

    You seem to suggest that the large volume of pollsters whose frequent polling results are reported on in the media are not conducting professional, scientific, proper political polls. I’ve heard your oversampling argument before and its a popular theory among right wing circles and those who have a vested interest in downplaying Obama momentum. I get that.

    I remember Bush polling ahead of Kerry by election day and by god he won the election. These things do correlate.

    I tend to believe the polls, since they all correlate with each other at this point, more than the Lee index found on your website. It doesnt mean that everyone called up on the phone (who are picked with precision based on voting history and likely voting demographics) will end up voting. It doesnt mean there couldnt be a change in the dynamics of the election should Mitt start to show he is willing to share a single plan more than just tax cuts for the wealthy.
    I believe if the election were held today your boy Mitt would get beat pretty handily.
    If i have to defer to someone who is very well respected statistician, does quite a bit of analysis and doesnt go on any one poll, who doesnt get too high or too low, I believe in Nate Silver and his predictions on the 538 blog. He is predicting an Obama EC blowout at this point, with an 80% probability. Im not pollster but i dont pretend to be so smart as to disbelieve those who do it for a living. And i definitely am not smart enough to go against a consistent universal trend by virtually all of the polls now which, for the moment, foretell of a electoral asskicking of your boy Mitt. I pray for the progress of the country that they are right.

  8. John 313 Says:

    I just saw that FOX NEWS has Obama up by 7 % in Ohio. Early voting starts next Tuesday. Do you think the Fox poll is inaccurate and over represented Democrats? The fox poll was in line with many other polls of Ohio I wasnt sure if you knew that no Republican has ever won the whitehouse without winning Ohio?

  9. John 313 Says:

    Look like I agree with Wolf Blitzer and you agree with Karl Rove. I guess well find out Nov 7th..regardless im sure well both be watching Fox News.
    http://www.huffingtonpost.com/2012/09/28/wolf-blitzer-polls-fox-news_n_1922393.html.

  10. Lee Eldridge Says:

    So according to Blitzer, the flaw in the argument that some polls are skewed is that the critics aren’t also citing the FoxNews polls as skewed? That’s a silly argument.

    First of all, I have seen people mention the same sampling issues with the FoxNews polls. So he’s wrong on this point. Though it’s probably not mentioned as often as the polls from the other media sources. But that doesn’t prove or disprove if the polls are skewed.

    Secondly, FoxNews, like ALL of the major media outlets (including CNN), hires out polling to companies like Marist, Opinion Dynamics and Quinnipiac. Blitzer is making an indirect assertion that a FoxNews poll would be skewed towards the GOP, and since the FN poll results are similar to the others, then they must all be right. He should know better. I’ve watched the polls for years. The polling company used by FN is often slightly to the left of many of the other polls.

    The reason that some are skeptical of the polls is two fold.

    1 – The pollsters are weighting the polls according to the voter demographics of the 2008 election. This heavily weights the polling towards Obama. For instance, younger voters typically make up about 12% of the electorate, but in 2008 it made up 18% of the electorate. Skeptics do not believe that Obama will get the same turnout among key demographics that he got in 2008. So if the pollsters weight young voters at 18% of the electorate, and they come in less than this closer to their traditional number of 12%, then the polls are skewed. This is also true of the black vote and the Hispanic vote, which turned out in much larger numbers in 2008 than they traditionally do in presidential elections.

    2 – The pollsters do NOT weight according to party ID, and have been sampling democrats at higher rates than 2008 voter turnout.

    Bottom line is we’ll find out in a few weeks. I don’t believe voters are undecided. And I don’t believe that voters will be changing their minds in the next few weeks. If Obama wins by 6-7 points, then I’ll have been wrong about the polls. And if Romney wins, it won’t be because people’s votes changed in the last few weeks heading into the election.

  11. John 313 Says:

    An interesting article in the lame street media today about addressing the issues you bring up which are floating around the right wing in light of Mitts bad ‘polling data’. It takes on some of your key accusations. I find it compelling and believable.

    http://www.huffingtonpost.com/mark-blumenthal/unskewed-polls_b_1924293.html

  12. Lee Eldridge Says:

    Hey John, I appreciate the civil debate.

    If Romney wins will you admit that I was right? :)

  13. John 313 Says:

    I’ll def have to admit that the polls where wrong, depending how he loses and where. Lets see where the final polls end up vs the vote. I don’t mean exit polls of course.

    Ill also have to admit that this country isn’t that bright either. After Bush winning twice we already learned that though.