Articles and Analysis


43 Days to Go and the Pendulum Swings Back

Obama won the earned media battle last week for the first time in more than a month and he did it without any particular tactical or strategic gambit. Instead, the political environment swamped the McCain campaign - another reminder that this is a Democratic year and this election remains Obama's to lose.

That is not to say that the McCain camp didn't make any tactical mistakes last week - they did. The candidate continues to search for its voice on economic issues (saying the economy's fundamentals remained strong was perhaps McCain's worst moment of the campaign) and his surrogates hurt him (see Carly Fiorina). The economic crisis also put President Bush back in the news, a development that doesn't help McCain. After a few weeks on offense, the McCain campaign is back on defense and playing on Obama's turf.

Poll numbers began moving toward Obama last week and we are sure to see some additional movement this week. Our sense is that Obama is ahead nationally by approximately 2-3 points. Note how easy it was for Obama to erase a small McCain lead (note, too, that it had taken McCain months of attack ads and a successful GOP convention to build up even a small lead). All of this reinforces the notion that the environment favors Obama decidedly and his campaign will benefit when the spotlight is on the current economic situation.

We want to caution everyone, however, that this is movement at the margins (primarily among swing and true undecided voters). This is, of course, where this race will be won or lost, and we will see this micro-pendulum swing back and forth several times in the next few weeks. The week ahead has some potential for some swing-inducing events, as well:

  • While the financial crisis will dominate the early part of this week, it may not be front-and-center by the week's end. In fact, we are taking a somewhat contrarian view that perhaps the timing of this financial meltdown is good for the McCain team (though obviously it would better for McCain if this hadn't happened at all). If today was October 22nd, our sense is that the election would be over. Instead, there are five weeks until the election and anything can (and will) happen between now and then. Chris Dodd said this morning that Congress will move quickly to pass the $700 billion package. Assuming this happens and the economy and financial sector stabilize a bit, the issue could move off of the front pages and the campaign could pivot on other issues. Until then, however, it is all about the economy.
  • There will be a debate on Friday night. This one is on foreign policy. If you are McCain, could there be a better time to change the subject? To this point, Obama hasn't been a very good debater. He was just adequate during his myriad of primary debates and not much better at the recent forums at Saddleback and Columbia University. Most voters haven't seen him in this type of format before; if Obama can avoid his professorial tendencies this is a real chance for him to show voters that he is capable of serving as commander-in-chief. For McCain, a "draw" won't cut it: he needs to win this first debate.

Current Issue Environment and Obama's Underperformance

It is all about the economy and jobs. The below snapshot from a recent CBS/NYT survey shows that the economy dominates as the most important issue when considering who voters would like to see as the next President. The below chart is startling for three reasons:

  1. This Presidential contest is on the verge of becoming a single- issue election. Nearly 50% of voters say the economy is the most important issue.
  2. We are engaged in a war that has gone on for five years, drained the treasury, cost the lives of more than 4,000 soldiers and has been at the center of our national dialogue for years...and yet only eight percent of voters in this poll say that Iraq is the most important issue when thinking about the next President. Think about it: the war in Iraq is the 5th-most important issue in the country right now - that is stunning.
  3. Obama has failed to convert the nation's deep economic anxiety into an electoral advantage (yet). The same CBS-NYT poll gave Obama a five-point lead (48% to 43%). While this is a modest lead it is surprisingly shallow given the Democratic lead in party identification and on the generic ballot. It certainly means (without benefit of the cross-tabs) that a large chunk of voters who think the economy is the most important issue are voting for McCain.
issue importance 9 22.png Historical Perspective

We decided to take a quick look back at Gallup polls conducted among registered voters in 2000 and 2004 so that we could compare them to today.

Sept 16 - 18 2008: Obama 48%, McCain 44%

Sept 13 - 15 2004: Bush 52%, Kerry 44%

Sept 16 - 18 2000: Gore 49%, Bush 41%

In 2000 the debates and intervening campaign events turned the tide; in 2004 they did not. McCain needs to do well in the debates or it will be very difficult to buck the current environment.

LCG EV Map: Colorado is the New Florida?

map 9 22.png

In 2004 Bush beat Kerry in Colorado by 52% to 47%. He beat Gore by ten points in 2000 (Nader had a sizable impact, garnering five percent of the vote in the state).
There has been a lot of talk about Colorado being the new Florida. Several smart political commentators have suggested that the race may come down to this one state. It is an interesting analysis. Certainly the state has moved from a "safe red" to a toss-up. Here are some demographics to consider:

  • Only 41% native to the state, so the electorate is shifting with new arrivals
  • 75% white, 4% black, 17% Hispanic
  • 14% military vets is high, not to mention lots on active duty, especially in Colorado Springs
  • Thought of as a rural state, but population is actually 85% urban, mostly in Denver and its suburbs
  • Colorado Springs is the state's second-largest city
  • 33% of state is college graduates, a bit below the national average
  • 21% blue collar, 65% white collar, 15% gray collar, so, again, it defies traditional perception as blue-collar/rural

Note below that our LCG average of the most recent polls shows that Obama's lead increased slightly last week, to 47% - 44%.

colorado 9 22.png

Yes, Colorado is important and is truly a toss-up in this election. But it is no more important than Ohio, Virginia or Florida. Given the recent historical presidential voter pattern, our sense is that if this state moves into Obama's column it will mean an electoral blowout for the Illinois Senator. Virginia is more likely to be the state that settles this race.


Vicente Duque:

Mr Lombardo :

You are mentioning Colorado and Virgina.

I visited INTRADE and got these numbers for bets on the presidential election :

Bids :
Colorado : Obama 56.2 McCain 39.2
Virginia : Obama 50.0 McCain 46.0

So according to bettors Obama still has a chance in these two states.

Vicente Duque


Gary Kilbride:

Vicente, the Intrade odds on Virginia have been all over the place recently. Obama has traded above 60 and below 40. I know one speculator who bet both sides at optimum prices guaranteeing a profit regardless who wins. Stuff like that goes on all the time on Intrade but it's seldom mentioned by analysts of the Intrade prices. Often it's not betting on who you think will win, as much as anticipation of a major line move. The convention period is prime for that type of speculation, since there's guaranteed to be a bounce, impacting the national price and also the individual states.

Anyway, those are the two states with blue demographic trends, and it's a matter of which state flips first. I don't claim to be certain of the correct answer. The factors certainly differ. In Colorado it's all about white women. They make up nearly 50% of the electorate in presidential years. I'm surprised we haven't had much focus on the white female voting trends during Obama's current poll surge, since it was a major topic when McCain/Palin were on the uptick.

Colorado is a strange state because the voters self-identify as Republicans but not as conservatives. The gap in the '04 exit poll was 29/38 in favor of Republicans but only 22/35 in liberal/conservative. That's very unique. Normally a 29/38 would equate to more than a 20+% difference in liberals and conservatives. Virginia was more typical, 35/39 in favor of Republican party ID, and 17/38 in liberal/conservative. By 2006, in a pro-Democratic second term midterm, Virginia was 36/39 and 21/35. There wasn't a Colorado exit poll in 2006, at least not one I've seen, but I can guarantee it would have been tighter than 21/35 in liberal/conservative.

Those numbers, particularly the 22% liberals already established four years ago, lead me to believe Obama has somewhat better opportunity in Colorado than Virginia. But the lack of black voters in the Colorado electorate, compared to a very significant amount in Virginia, make me hedge my guess. In sum, Obama might get across the finish line via massive African-American turnout in Virginia, while in Colorado he's at the mercy of preference among white women.


If you're interested in analysis of Intrade state-by-state data, maybe you should look at http://politicalsimulation.blogspot.com , where I post the results of some simulations I run daily based on Intrade.com state-by-state data. The results generally contradict the main Obama vs. McCain to win odds on Intrade.


Vicente Duque:

Answer to Mr Gary Kilbride :

You said :

"I know one speculator who bet both sides at optimum prices guaranteeing a profit regardless who wins."

Excellent strategy to make money. I guess that some people at Wall Street use that strategy of buying little lows and selling little highs day after day. I don't know the name of the strategy but it may be "microtendencies".

Even if INTRADE is very distorted by speculation, I see that COLORADO, NEW MEXICO, NEW HAMPSHIRE and IOWA have had favorability to OBAMA in that betting house and for a good amount of time.

On the other hand, VIRGINIA and OHIO seem very even for both candidates in INTRADE. INDIANA seems very difficult for OBAMA.

If Obama has money he should be investing heavily in these very crucial states.

Answer to Heywhoah :

I visited your site :


and was delighted by it. Very valuable site. And I understand that you are tossing coins with the odds that INTRADE has. I also agree that the data of bettors is pure gold, because they are risking their money, even if there is speculation, speculators have to retire when November 4 approaches, or refine their bets according to reality. Also money is painful, pollsters can make mistakes but bettors are ruined by them.


The fun is not going to end on November 5, we can continue analysis of results for many months.

I want to study the demographies of blacks and latinos in the toss up states, the demographies of youth and gender, specially for the very young voters.

I guess that some of that data could be useful for marketers.

Thanks a lot

Vicente Duque


Vicente Duque:

A wonderful article in CNN about Latinos "the sleeping giant" helping Obama win COLORADO :

"Latino vote emerges as swing vote in Colorado"


CNN says :
"Nationally, Barack Obama leads John McCain 65-30 percent among Latinos in the latest CNN/Opinion Research Corp. polling."

also CNN :

"Latino vote could prove pivotal in Colorado, New Mexico, Nevada, Florida and perhaps even North Carolina, which ranks 11th nationally in its percentage of Latino residents."

Vicente Duque


Colorado is interesting in the polls vs prediction markets comparison with both tracking each other fairly consistently (unlike most other swing states). The Intrade implied probability for the Dems in Colorado has been bouncing around the mid 50's to mid 60's for months (with the occasional blip happening), which is close the the average probability that the national market has been producing for a 1 to 4 point Democrat polling margin in the national polls - especially in the months leading up to the convention. After the convention, even though the national markets went berko, Colorado remained stable in the Intrade markets with a decent trading volume.

Perhaps Colorado is this year's Bellwether State?

For anyone that likes their markets with a dose of nerdiness, I've been running daily and weekly simulations on the Intrade markets that adjust for State non-independence and the longshot bias in the probability tails for my Australian audience for a while, as well as tracking those results through time.


The simulated State market results are turning out to be a pretty good leading indicator of the national headline Intrade markets and certainly seem to be acting as an error correction mechanism for the national headline markets.

Also interesting is the way the State markets do seem to contain much richer information - often moving a day or two before a big poll result in a given State is released. If it happened once or twice you'd put it down to chance, but happening as regularly as it does it smells like the electoral equivalent of insider trading.


Vicente Duque:

Answer to Possum Comitatus :

I visited your website :


Excellent approach. I agree with you on the importance of INTRADE.

I like your many considerations and agree with them.

I think that state bettors living in a state function as "insider trading" or "priviliged traders" with more knowledge. And those that trade bets on a state that they know very little are risking their money in an imprudent way.

There are more ways of knowing the future than just looking at polls. People have intuitions about their neighbors or nearby towns in the same state, they know what is going on and the traditions and mentality of near folks.



Post a comment

Please be patient while your comment posts - sometimes it takes a minute or two. To check your comment, please wait 60 seconds and click your browser's refresh button. Note that comments with three or more hyperlinks will be held for approval.