Man, I’m glad I waited a couple of days to do this. Or maybe I’m not. I can’t recall an election when so many things were coalescing on any one point in time:
#1. South Carolina and the Kennedy Bounce. We have to group these two things together because they occurred so closely together in time. My perception is that Obama has gotten a South Carolina bounce of about 5 points nationally, plus a bonus Kennedy bounce that ranges from as much as 10 points on the coasts, to almost nothing in the heartland. This is based on a whole number of things. The Gallup tracking poll, obviously, is one of them. So are all the favorable state polls that have come out in the past 24-36 hours. Obama beat his polling averages by 5 points in Florida, and the actual margin was probably more like 7-8 points once you exclude the early voting results. Google Trends suggests that Obama has had a sustained period of momentum of at least three days, rather than just a one-day spike.
The Rasmussen national tracker appears not to have shown much momentum BUT I have an algorithm I use to try and pull individual day’s results from a tracking poll, and that suggests that Hillary’s results are being boosted by a very strong showing on Saturday – the last day before the results of South Carolina were known. If I’m right, we’ll see the Rasmussen tracker consolidate to somewhere between 2-6 points when it comes out later today.
#2. A Florida bounce? Frankly, I don’t expect to see one. Florida did not affect the national narrative, Hillary’s press coverage was generally very skeptical, and any headlines she got were usurped immediately by John Edwards’ withdraw. Speaking of which:
#3. The withdraw of John Edwards. When I analyzed this issue before, my conclusion was that an Edwards withdraw would yield a net gain of a couple of percentage points to Obama. That was based on an analysis of New Hampshire exit polls. When I looked at Florida and South Carolina exit polls instead, I saw no statistically significant differences in Edwards’ voters perceptions of Hillary Clinton and Barack Obama.
So my default assumption is that Edwards’ withdraw is probably about neutral in the South, and probably helpful on balance to Obama everywhere else. It certainly isn’t as simple as saying "Edwards supporters are working-class white folks, and working class white folks like Hillary". This is because the most important feature of Edwards voters is not that they’re working-class white folks (which, by the way, they aren’t – they tilt slightly upward on the income scale) but that they’re working-class white folks who didn’t have Hillary Clinton as their first choice. Nor, of course, did they have Barack Obama as their first choice. So they might like Clinton or Obama, but they didn’t like them enough to make them their first choice. This is very important.
In fact, the best way to look at Edwards supports is that most of them are undecided. Some of them undoubtedly prefer Clinton or Obama – but those are probably fairly weak preferences.
Who benefits from an injection of undecided voters into the race? On balance, Obama does, because Obama is currently behind. The more time he has to catch up, the better off he is. And in politics, undecided voters are your clock. The election begins when nobody has any idea whom they’re voting for, and it ends when everybody has decided on a candidate. So increasing the number of undecided voters has the same effect as buying Obama more time – which is helpful when you’re coming from behind.
#4. The endorsement of John Edwards (and Al Gore, Bill Richardson, et. al.) I have not attempted to "price in" any endorsements. Certainly, I agree with the conventional wisdom that John Edwards is more likely to endorse Barack Obama or to endorse nobody, and is less likely to endorse Hillary Clinton. But I haven’t made any assumptions about this. If Edwards does endorse, I suspect it’s worth something on the order of a 1-3% swing, depending on how actively he campaigns as a surrogate. I think an Al Gore endorsement would be worth slightly less than that. And I think a Bill Richardson endorsement would not move the needle much nationally, but might deliver the state of New Mexico, which is still pretty significant.
#5. The Snub, and other State of the Union related developments. Believe it or not, I suspect The Snub might be worth 1-2% points for Hillary among female voters. It’s gotten a fair amount of attention in places that Kossacks might not visit on a daily basis, but a fair number of voters will. I actually think that The Snub is more important than Florida, which mostly tells you how little I think of Florida. With that said, Barack Obama otherwise won the State of the Union news cycle, so all The Snub might be doing is offsetting those gains.
#6. Boratgate. Way too early to call. It’s obviously a negative for the Clintons, but I don’t know whether it’s an 0.1% negative, or a 1% negative, or a 10% negative. Extremely early indications (e.g. the lack of coverage on Morning Joe) are that it will be closer to the 0.1% side of the spectrum. But, it does provide some tactical advantages to Obama. He is probably insulated from any oppo research dump for the rest of the cycle (including Rezko), and there are a number of ways that it could come up unfavorably for Hillary in tonight’s debate.
#7. The Debate. Don’t ask. It’s important. Really important. Especially with all the Edwards supporters tuning in. It could determine our next president. No pressure.
#8. Ground Games and Late Deciders. I looked at this yesterday and concluded that Obama has outperformed his polls by an average of 3-4 points so far on election day. There are some reasons (the "incumbent rule", the difficulty of reaching cellphone-only voters) to believe this phenomenon may be persistent. I have accounted for a couple of extra points for Obama as a result, especially in caucus states.
Looking at this entire landscape of these trends, I see a very close election. Very, very close. Closer than the media seems to recognize. Moreover, I see a highly volatile election. If a whole bunch of things break toward Obama – Edwards endorses, he wins the debate, Boratgate blows up – he could win Super Tuesday by a decisive enough margin to win the nomination. I did not believe that was possible a week ago, even as a best-case scenario, but I believe it’s possible now. Naturally, the same is also true for Hillary. If she wins the debate, and gets an enthusiastic endorsement from Bill Richardson, and Obama makes some gaffe that creates a Youtube Moment, she could win decisively. But the most likely probability, of course, is that the election battle is going to continue for some time. Let’s go to the individual states.
We had a late, post-SC poll break here, which shows Obama ahead by 5 percent. That’s much more consistent with my impressions about Alabama than the Rasmussen poll from a week ago. While Alabama is somewhat whiter than South Carolina and perhaps the deepest state in the Deep South, I can’t imagine Obama winning South Carolina by 29 points and not being favored in Alabama. He’s had a presence on the ground here for a long time, and Edwards hadn’t polled as strongly here as he had in some other Southern states, so I don’t know that he’ll swing things either way.
Should be solid Obama for the demographic and organizational reasons I discussed in the last installment. The question is whether he picks up a 1-point, 3-point, or 5-point delegate margin. Given how well Dennis Kucinich did here in 2004, there’s an outside chance that Mike Gravel gets a delegate, but I’m not going to worry about that.
Early voting state in which both candidates have been advertising fairly heavily. We have two polls, both from before South Carolina, which show Hillary with an 8-point and 21-point lead respectively. I’m inclined to believe the margin will wind up closer to the 8 than the 21, based on a combination of Obama’s national momentum, the level of engagement in the state (which suggests both sides think it’s close), and the fact that the Cronkite/ASU poll used a fairly restrictive voting screen that tended to filter out new voters. Outside chance it could be a surprise state for Obama, but between the closed primary and the early voting, he’ll most likely only win it if he’s having a very good night nationally.
Counterintuitively, I think Edwards dropping works to Obama’s benefit here, since if you weren’t for the Clintons in Arkansas the first time around, you’re unlikely to be with them now. Still, Obama was late to the ground here, and has no institutional support. I would not be surprised by anything from a 15-point to a 50-point win, but either way, it’s going heavily for Clinton.
So much going on here. But let’s start with the basic lay of the land. Survey USA suggests that roughly one-quarter of California voters had mailed in their ballot before the South Carolina primary. If we assume that Hillary had a 15-point advantage in those ballots – roughly her lead in the polling averages before South Carolina, that means that Obama would have to win the balloting by 5 points from here on out to take the state.
This remains relatively unlikely. Obama might win California – but it will probably be on the strength of a substantial nationwide surge in which it’s just the icing on the cake. It’s states like Massachusetts, or Connecticut, or Missouri, that are more likely to behave exogenously from the rest of the country, and provide Obama with a "surprise" victory. Still, I don’t think we can completely rule it out. In addition to the Rasmussen poll – and I assume that Rasmussen is smart enough to account for early voters in its tabulations – we also have a private tracking poll that evidently showed Obama pulling ahead after South Carolina. Al Giordano at The Field has also parsed through the California internals, and concluded that they may be underestimating Obama’s support. And there’s a lot happening on the ground. Obama has 18 field offices open. He’s finally picked up some big name newspaper and legislative endorsements (though they’re still no match for Hillary’s). Both sides are buying up a lot of airtime. And a lot more late votes came up for grabs when John Edwards pulled out of the election.
If Obama can slingshot to a last-minute victory, it will probably require making substantial inroads with Latino voters, as Latino voters tend to wait until election day to vote. But I actually think this is possible. Remember that it took Obama a long time to pick up African-American support. And frankly, it was only recently that the Obama campaign recognized the importance of the Latino vote. The endorsements from Ted Kennedy and Xavier Becerra should help, and the rest is up to the ground game. Obama actually picked up a higher percentage of votes from Latinos in Florida (30%) than in Nevada, in spite of performing much worse in the state overall; he also picked up a higher percentage of Latino votes than white votes. If he can get that Latino number up to 40% in California, the state becomes too close to call.
We’ve always had this state in Obama’s column, and he had a 2-point advantage in the only recent (though pre-SC) poll, that coming from a pretty good agency in Mason-Dixon. Several factors point to the fact that he might be able to run up the score a little bit. It’s a caucus state in which he’s better organized, he gave a pretty important speech in Denver yesterday, and the Colorado breed of liberalism seems like a good fit for him, with pockets of progressivism and libertarianism, but not as many of the mainline Democrats that tend to favor Hillary. His organization may also help him to pick up the bulk of the Edwards support, which was not insubstantial here.
Well ... I know that the Obama folks are excited about the Rasmussen poll that shows it dead even, and you can sort of understand why it’s a good state for Obama since there are lots and lots of $100K+ voters in Connecticut, a fact which was manifested in his strong fundraising here. But I’m still keeping it narrowly in Hillary’s column based on her overall strength in the region, the closed primary status, and the fact that we don’t have a lot of confirmatory evidence of his bounce here.
Delaware looks like an 8/7 state, and it’s just a matter of where that 8th delegate goes. I’m giving it to Obama for the most trivial of reasons, which is that Michelle Obama is visiting the state tomorrow, and in a tiny little state that never gets any political attention, that figures to swing a few votes. As I noted last time, Delaware has a fairly substantial African-American population (perhaps as much as 30%-35% of the electorate in what I can infer from the Farleigh-Dickinson poll).
Georgia now clearly seems to have differentiated itself from Alabama as being a more favorable state for Obama, and that shouldn’t be surprising as it’s more urban and has a somewhat larger African-American population. There was one poll showing Clinton with a 1-point lead, but Obama led in the averages even before South Carolina, and he’s now invested in seven more offices in the state and visited it twice since the first of the year, while picking up the endorsement of the Atlanta Journal-Constitution. Even if a plurality of Edwards’ support breaks to Hillary here, if Obama can’t win this one something’s gone seriously wrong.
Four field offices? Obama’s really going for the kill in a state that’s always looked pretty good for him; he’s also scheduled to visit there later this week. While there aren’t a lot of Democratic endorsements to go around in Idaho, Obama also has substantial support among the state legislature. Clinton is not really contesting the state and it all seems to be adding up to the sort of wide margin that Mitt Romney won in the Nevada Republican caucus.
I’ve actually shifted a delegate or two to Clinton here. Although another poll came out showing a broad lead for Obama, this is sort of the reverse of the Arkasnas case, in that Obama is effectively running at the incumbent/better known commodity, and so Clinton might pick off some chunk of Edwards’ support downstate, where Bill was campaigning today. Still, the situation is not analogous to New York, where Obama may actually win some New York City CDs – odds are that Obama gains more net delegates in Illinois than Clinton does in New York, even though the Illinois pie is slightly smaller.
This has sort of become Obama’s adopted home, and between his recent visit and the Sebelius endorsement, he could rack up a pretty decent margin in a caucus state.
Well ... we’ve sort of come full circle on Massachusetts since the last update. If you’ll remember last time around, I was insisting that Massachusetts was an Obama state, but had no polling data to back up that assertion. In the interim, there were a couple of Survey USA polls showing enormous margins for Hillary. Now, I think Survey USA is an average-to-decent pollster, but one thing about them is that they push learners very hard. Actually, what happens is that their automated call script will pause for several seconds before saying ‘Press 9 for undecided’, which is enough to discourage many people from selecting that option. So Survey USA polls tend to pick up a lot of soft support ... consider that they had Clinton up by 33 in Florida barely a week ago, for example, up 37 in California in mid-October ... their margins for candidates with name recognition tend to be magnified. Note also that there are other polls out that were in the field before South Carolina showing more reasonable 16- and 28-point margins for Clinton.
Nevertheless, there’s evidently been pretty substantial movement post-Kennedy, post-South Carolina. And since the Rasmussen poll suggested that most of Edwards’ support is inclined toward Obama, and since it’s an open primary with no early voting, I think Obama could absolutely pull off the upset. But I don’t quite have the guts to call it for him; 6 points isn’t much of a lead, but it’s still a lead for Hillary. Big state as far as narrative goes.
We finally got a Minnesota poll!. But we have to be careful about how we interpret it because it’s a poll of registered Democrats. This presents two problems: #1, Minnesota has an open caucus. If, as in Iowa, about 20% of the caucus participants are independents, and those votes break 2:1 for Obama, that’s worth a swing of 7-8 points to him. #2, a poll in a caucus state is pretty useless without a likely voter screen, and Obama appears to be making a late organizational push in the state. So I’m taking this poll as an indication that Minnesota has tightened – and giving the benefit of the doubt to Obama, as in other caucus states.
Along with Massachusetts, Missouri and Tennessee are the two states that I appear to have miscalled for Obama before. Culturally, Missouri is probably as much a Southern state as it is a Midwestern state, and a Southern state with just 12% African-American population spells trouble for Obama. Clinton is also fighting for this one, having made three visits since the first of the year. Still, this is the one state where the distribution of Edwards’ support could have the most effect on the outcome, and an actual endorsement by Edwards might be enough to swing the momentum to Obama.
I’m mostly just going with the math on this one. Obama trailed by an average of 15 points in the last four pre-SC polls, which actually represents a fair amount of improvement from a month ago. Add in the South Carolina bounce, and we’re probably talking about an outcome somewhere in the range of 10 points. But I don’t see much more than that because Clinton has played good defense by visiting the state several times. This state is also sort of a mini-bellwether; if the networks can’t call it for Clinton fairly quickly after polls close, she’s likely in for a long night (the analogous state on the Obama side is probably Georgia).
I just can’t see this being a good state for Obama, between the very heavily Hispanic population and a closed primary with early voting. But he’s making a major late push, opening no fewer than seven field offices and with a couple of events scheduled on Friday. If one of those Friday events involves a Bill Richardson endorsement, he should be competitive; otherwise I think the underlying demographics are metrics are too friendly to Hillary to overcome.
Obama has quietly been picking up ground here faster than any state in the country. Couple that with the fact that Kennedy bounce appears to be highest along the Eastern Seaboard, and suddenly it looks like he might limit Hillary’s margin to 30-50 delegates. Still, I’m guessing the PPP poll that shows Obama -12 will be something of an outlier.
There’s also a lot of very late engagement on the ground in New York, including an Obama ad buy in New York City, which probably reflects the delegate math. Most CDs in New York have 5 delegates. What that means is that the winner of a CD in New York City gets in effect a bonus delegate, and New York City is presently very competitive. Upstate and in the suburbs, meanwhile, Obama gets 2 out of 5 delegates if he hits 30% of the vote in a CD, a threshold which is also very close. In any event, the popular vote won’t tell the whole story here.
I had this leaning Hillary before on the assumption that Obama wouldn’t do well with rural voters, but he somewhat challenged that assumption in Nevada. Both camps now seem to be acknowledging Obama’s superior organization in the caucus states, and I’m calling those states for Obama unless I have a good reason not to.
For whatever reason, the demographics in Oklahoma break poorly for Obama; he had yet to break 19 percent in any public poll. His performance has been poor enough, in fact, that I assume Hillary will pick up the majority of Edwards’ strong cache of support in this state, although Obama has opened an office recently here in what might be considered a symbolic gesture. Pretty big delegate grab for Hillary.
We do have a post-SC poll, and it has Hillary +11, which isn’t terrible for Obama but suggests that victory is an uphill climb, especially as he lacks any strong surrogates in Tennessee, and as Hillary paid a recent visit here to shore up her advantage. Remember, the Memphis area is practically in Arkansas. This is another state where an endorsement from John Edwards could have some of its most tangible benefits – not to mention Al Gore.
With the help of a commenter, I’ve tracked down a couple of Utah polls, but they only serve to confuse the situation as they produce wildly divergent results, apparently because they’re based on extremely small sample sizes (~100 people). The numbers that jump off the page to me are the extremely young population and Obama’s threefold advantage in fundraising. That plus Obama’s strength in rural Nevada are why I’m still giving him the extra delegate.
Obviously, we’d be in for a hell of a fight if Tuesday night turned out this way. Eleven states for Clinton and 11 for Obama; 845 delegates for Clinton and 833 for Obama. We might actually be looking to the results of Democrats Abroad and American Samoa to break the tie.
By the way, I do think it’s going to be the delegate narrative that’s the key. One positive side effect from all the squabbling about Florida and Nevada is that it’s put the media’s focus squarely on delegates – that’s how the outcome of Super Tuesday is liable to be judged.
More broadly, there are five basic scenarios that could emerge from Super Tuesday. I’ll provide those scenarios, with a hypothetical example of what each might entail.
#1. Clinton effectively ends the race on Super Tuesday (10%). Clinton limits Obama to winning Illinois, Georgia, and 2-3 smaller caucus states. She wins a couple of states that Obama was supposed to win, like Alabama and Colorado. She wins California by at least 12 points, and New York by at least 20-25 points.
#2. Clinton has the lead, but it’s anybody’s nomination (25%). Obama wins Illinois, Alabama, Colorado, Georgia, Idaho, Utah, and Kansas, and keeps the California margin to some reasonable number, but Hillary wins all of the states she’s supposed to win, plus sweeps most or all of the swing states like Minnesota and Connecticut.
#3. The race is truly too close to call (35%). Something like the base case scenario I’ve outlined above.
#4. Obama has the lead, but it’s anybody’s nomination (25%). It might not take a whole heck of a lot to go from #3 to #4, because the media will be quick to anoint Obama with the momentum if he beats expectations, and his expectations remain lower. Also, as we discussed last time, the rest of February is probably the most favorable part of the schedule for Obama, so he could consolidate this momentum by winning the Beltway states on February 12th. If Obama wins California by any margin, he will probably be regarded as the frontrunner, even if he has a couple of small stumbles elsewhere. He might also be regarded as the frontrunner if he loses California narrowly, but he wins at least two states from the following group: Arizona, New Mexico, Missouri, Tennessee, Connecticut, Massachusetts. By the way, another underrated advantage for Obama is that if John McCain effectively ends the Republican nomination on February 5th, Obama will get the bulk of independent support in any remaining open primary states.
#5. Obama effectively ends the race on Super Tuesday (5%). Obama wins California, plus at least four out of the six states in that "leans Hillary" group above. If, for example, Hillary were limited to wins in New York, New Jersey, Oklahoma, Arkansas, New Mexico, and Tennessee, that would probably cause enough consolidation around Obama to effectively end the race.