Noah Millman may not have persuaded me that Bob Barr will be a big factor in the '08 race, but his latest post has me convinced that neither Karl Rove's consulting firm not Howard Dean's pollster should be trusted when it comes to Barack Obama's electoral-college strengths and weaknesses.
« Doubting the Barr Factor | Main | Terminal Narcissism »
Strange Maps
21 May 2008 05:12 pm
TrackBack
TrackBack URL for this entry:
http://rossdouthat.theatlantic.com/cgi-bin/mt/mt-tb.cgi/21620
Comments (23)
"The whole election will come down to Ohio, Nevada, New Mexico, and Colorado. If Obama wins Ohio, he wins; if he wins all three southwestern states, he wins; if he wins Colorado and either New Mexico or Nevada, it's an EC tie and he wins in the House because Democrats control the majority of delegations."
I'm not sure I agree with that, for several reasons. This essentially cedes Michigan to Obama (likely, I think, but very premature at this point) and Virginia to McCain (ditto, I think, unless Webb is the nominee). It also removes the possibility of an Obama upset in New Hampshire, one of the High Plains states or Alaska.
The way I see it is this:
The first test is the Mid Atlantic/Rust Belt. Michigan, Virginia and Ohio are on the table (I'm assuming Pennsylvania, North Carolina and Florida are safe at this point). If McCain wins three of those states, he wins. If Obama wins two (or three) of those states, he wins. If Obama only wins one of those states, we move on to the tie breaker -- New Mexico and Nevada (Obama's consistently been polling 5-10 points ahead in Colorado, so I'm giving him the state).
If Obama wins both states, he wins. If McCain wins both, he wins. If they split the states, we move on to the second tiebreaker -- which candidate wins more "upset" states (probably NH, ND, AK or NE for Obama; probably IA, WI, or CO for McCain). If neither candidate wins an upset (or the upsets balance out), we go to the House.
Why would anyone trust Karl Rove about anything? He's the American Martin Bormann.
A bunch of Millman's reasoning is specious.
The issue here is who prevails in a 50/50 popular vote election, and I think that McCain would have a clear edge in an election like that.
If Obama wins 54/44 nationally, of course he's going to carry a bunch of states currently in Republican column.
But he's definitely going to run behind his national numbers in FL and OH, and he's far more likely to run behind his national numbers in PA and MI than Clinton.
That's why he's a weak electoral college nominee.
Re: This essentially cedes Michigan to Obama (likely, I think, but very premature at this point)
Michigan has been trending strongly Democratic since the early 90s, voting Clinton, Clinton, Gore, Kerry—while sending two Democrats to the Senate and, in 2002, a Democrat to the governor’s chair. Even the legislature has seen a Democratic comeback, after a very long time of total GOP dominance. The Michigan economy is in hideous shape. The people of Michigan are mightily upset-- and not at a senator from Illinois. And yes, there’s plenty of Black votes to be had out of Detroit and its metro area. Any idea that Michigan will go Republican this year is a pure fantasy.
pollster.com currently has Obama running about 1% ahead of McCain in the aggregate of national polls. If it really does turn out to be that close, then it might actually be worth obsessing over where the electoral votes are going to come from.
But in all likelihood it won't matter. There's no law of politics that says that the current margins are going to hold over the next 6 months. As long as the popular vote winner wins by more than about 1%, he's going to win the electoral college. A popular vote / electoral college split is something that only happens about once a century, so it's a bit early to worry too much about it now, when we have no idea what the race will look like in October.
Obama's only up by an average of 1.2% in Michigan (and down 1 in the most recent poll), so he could lose the state, but I tend to think he should at least be able to hold the Kerry states. McCain will probably try to make an issue of how Obama killed the revote, but I think Obama's pseudo-protectionist message should play well there - definitely better than McCain's "the jobs that left Michigan are never coming back" line, which I'm sure will work its way into an ad or two. Of course, an upswing in the economy or events in Iraq could change everything, but at the moment I think it'll really come down to Ohio and the southwestern states.
"pollster.com currently has Obama running about 1% ahead of McCain in the aggregate of national polls. If it really does turn out to be that close, then it might actually be worth obsessing over where the electoral votes are going to come from. But in all likelihood it won't matter. There's no law of politics that says that the current margins are going to hold over the next 6 months."
The problem is that the race could very well break to the point where Obama wins nationally by 3% or 4% and still loses the electoral vote.
Obama is going to run considerably behind his national numbers in both OH and FL, and losses in those two states make the math very difficult. Obama basically has to run the entire table at that point to get to 270.
If Obama gets a 5% or 6% popular vote advantage, then yes, counting states will become irrelevant. But short a blowout victory, he's got some real problems with the electoral math.
It's worth noting that the Appalachian region, where Obama struggles so badly, is precisely where Presidential elections have been decided for the past 40 years.
"Obama is going to run considerably behind his national numbers in both OH and FL, and losses in those two states make the math very difficult. Obama basically has to run the entire table at that point to get to 270."
Petey, this simply isn't true. If Obama wins MI and loses OH and FL, he can still win the election by winning (this all assumes he wins the Kerry states minus NH plus IA and CO, where he's up by an average of 5 in RCP):
1) New Mexico (269 EV sends it to the Democratic House);
2) Nevada (same);
3) NM & NV (274 EV);
4) NM or NV and any other state (NH being the most likely culprit);
5) NH and any other state (AK or ND maybe?);
6) Missouri (275 EV);
7) Virginia (277 EV);
8) North Carolina (283 EV).
The fact is, the Kerry states plus IA and CO are a big chunk of votes. And the extra 16 EV goes a long way towards closing the gap left by not winning FL or OH.
"It's worth noting that the Appalachian region, where Obama struggles so badly, is precisely where Presidential elections have been decided for the past 40 years."
Again, this just isn't close to being true. The only post-war elections that have been close enough so that one region would be decisive are: 1960 (decided by Illinois and Texas), 1968 (decided by a bunch of close states that fell Nixon's way--most prominently CA, NJ and MO), 1976 (decided by OH, MS and WI), 2000 (Florida) and 2004 (Ohio). Only 2004 really involves Appalachia, and even that takes a huge view of "Appalachian region" -- only a tiny part of OH (geographically and population-wise) can be considered "Appalachia."
What Joe said.
But also: Petey's assumption is that the geography of campaigning doesn't affect the aggregate vote totals. And I don't think that's true. If McCain, with fewer resources than Obama, financially and organizationally, has to fight in the Mountain West, the High Plains, the Upper South, and other regions he'd like to have locked down, while still fighting in Ohio and Florida, then he's going to have a harder time keeping this close. Whereas if the whole contest is fought as trench warfare in Ohio, McCain's in a better position to keep the aggregate numbers close, and hence to win. Basically, Petey's assuming that Obama's going to run a campaign that racks up huge totals in safe Democratic states and then fights like hell to win an uphill battle in the swing states of 2000 and 2004, and thereby could lose the Electoral College with a 3% national margin. But why would he run that kind of campaign? If he's got a better shot at making a race of states in the High Plains, Mountain West and Upper South than previous Democrats have, why not campaign there? And why wouldn't such a campaign have an impact on the aggregate national numbers as well as the Electoral College?
That's my view, anyway.
Also: the question asked was not "which map looks stronger based on today's polls: Clinton's or Obama's" but "which map looks stronger based on not merely today's polls but the shape of their coalitions: McCain's or Obama's." Petey's comparing Obama and Clinton, a contest that, he has admitted, is over. I'm comparing Obama and McCain.
Incidentally, if anyone is wondering, the current state delegation breakdown in the House is 27-23 in favor of the Democrats. I'm not sure whether the 110th or 111th would end up voting, though I'll note that it is unlikely that Republicans would gain any states other than perhaps Mississippi. Democrats look likely to pick up New Mexico.
Quick questions: (1) What happens if it is 25-25, and (2) what happens if a state has an even number of reps and they split?
I think Petey is right, he just misdescribed the region; its the Upland South region of the Midwest that has been determinative, that's a line from Missouri, Indiana, Kentucky, Ohio, West Virginia and Pennsylvania (yeah, the last two aren't Midwest, sue me). But its a political/cultural border that one would expect to be disproportionately influential.
Incidentally incidentally, the above was slightly off. It's actually 27-21-2 in the Democrats favor -- though I note that one split is Arizona, and you would suspect that one Democratic rep would probably abstain to let the state vote for McCain.
For what it's worth, the Appalacia region's decisiveness is quite debatable, Indiana is not a swing state. And WV was quite solidly Democratic, and looks solidly GOP.
It's true that there are those voters in PA and Ohio. But, as Alec points out in "The Trail," Democrats have beaten blue collar defections through improved performances in the philly burbs. Indeed, Kerry didn't lose ground to Gore in the rural areas, he lost ground in the citie, particularly. Some members of that cultural line are certainly important, but they're hardly the only key voters in Ohio and PA.
Incidentally, Survey USA has Obama beating McCain by a surprising margin today:
http://www.surveyusa.com/client/PollReport.aspx?g=9901f8fc-034e-4a1d-ab36-f6e5c918614e
The more I think about it, the worse the electoral map looks for McCain if he can't win PA or MI. If Obama wins those states plus IA and CO (where he is significantly ahead), all of the sudden McCain must win ALL of:
NM, NV, VA, NC (and he's currently behind in two of those)
and all but one of
NH, AK, ND, MT.
Folks, THAT'S threading the needle.
"Petey, this simply isn't true. If Obama wins MI and loses OH and FL, he can still win the election by winning (this all assumes he wins the Kerry states minus NH plus IA and CO..."
So, Obama would need to win each and every contest in MI, WI, MN, NH, IA and CO, all of which are likely to be quite close if the national popular vote is close. And that still would only get him to 265. Then you start talking about needing states like ND and NC and MO that Obama would almost definitely lose if the national popular vote is close....
I believe that is precisely what I mean by saying that Obama basically has to run the entire table if he loses OH & FL to get to 270. If McCain manages to poach absolutely anything - say MN by picking Pawlenty - the whole thing falls apart.
Obama can win the election if he loses OH & FL, but if the popular vote margin is within a couple of points, it's not particularly likely that he will win the election if he loses OH & FL.
-----
And, of course, from today's Quinnipiac:
Florida:
McCain 45, Obama 41; Clinton 48, McCain 41.
Ohio:
McCain 44, Obama 40; Clinton 48, McCain 41.
Ok, Petey, so...what exactly is your point? You've admitted the nomination is over; is it at all beneficial to anyone at this point that six months from the election Clinton holds an edge outside the margin of error in a couple swing states?
(Which, of course, doesn't account for the factor of Clinton's supporters knowing their only hope of her nomination is to make her look more electable by claiming not to vote for Obama when we all know they will, as soon as she concedes and the healing can begin.)
So, are you just concern trolling or what? Do you *really* think polls six months from the election have anything to do with the actual result? Do we really need to bring up Clinton's numbers from six months before every primary she lost? Bill's May 1992 numbers? Or are you still making these flimsy points to try to keep your intrade holdings on life support?
"Do you *really* think polls six months from the election have anything to do with the actual result?"
I think that the polls six months away from the election do indeed tell us that Clinton would be significantly stronger than Obama in FL & OH.
I think that the polls six months away from the election do indeed tell us that Obama will very likely run well behind his national numbers in FL & OH.
"Ok, Petey, so...what exactly is your point?"
Exactly what I said my point is: that Obama is an extraordinarily weak electoral college nominee because of how he underperforms among the downscale whites east of Mississippi, who have an outsized voice in electoral college politics.
If the nationwide popular vote is close, Obama likely loses the election.
What's your point? That we should pretend this dynamic doesn't exist because you personally don't find it "beneficial"?
Petey, you arguably have a fair point on Clinton doing better in Ohio and Florida.
But, I think you reach a pretty sweeping conclusion on a dubious basis. Yes, working class whites are important in some swing states like Ohio, Pennsylvania, and Florida. They're not the only swing voters or states out there.
Virginia and Colorado are worth as much as Ohio is.
So, let's suppose Obama loses New Hampshire, but wins Iowa. (Also, the former proposition is hardly a given, as he was leading for much of the cycle. Indeed, several of the polls that have him down came when the Wright stuff was prominent in the news, and may have died down since).
That gets you to 255.
He could win with a combination of Colorado and Virginia or Colorado and New Mexico and Nevada.
The idea of him being an extraordinarily weak candidate in the electoral college ignores the fact of cost-benefit analysis. It's true that having more of a white collar appeal hurts you in Ohio, but it can help in Colorado and Virginia.
Furthermore, for what it's worth Quinnipiac's national polls had Obama performing as well among white voters without a college degree as Clinton, and Washington Post only gave her a small edge. Indeed in both of these surveys, Obama did about as well among these voters as congressional Democrats did in 2006 and better than Gore and Kerry.
"Virginia and Colorado are worth as much as Ohio is."
Sure. But look at the numbers we've seen in the past couple of days.
According to Zogby, Obama is up by 10 points nationally.
And according to state polls at the same time, he's up by 7 points in VA, 6 points in CO, and 6 points in PA.
Notice the trend here? Virginia and Colorado - where you think Obama can replace Ohio's electoral votes - are states where Obama is running behind his national numbers.
When Kerry was up nationally in summer 2004, he was running ahead of Bush in Missouri. But it was always obvious that Kerry was going to run behind his national numbers in MO, and when the national numbers closed, his lead in MO disappeared.
Maybe Obama is going to win nationally in November by 10 points. In that case, he'll win VA, CO, and PA, and electoral college maps will be utterly irrelevant.
But maybe the national race is going to tighten, and if it does, Obama is going to have terrific difficulties getting to 270 because of the nature of his coalition and the dynamics of the electoral college map.
The whole point of looking at the electoral college map is seeing what'll happen in a close race, not in a blowout.
"for what it's worth Quinnipiac's national polls had Obama performing as well among white voters without a college degree as Clinton"
Obama's difficulty is cultural, not racial.
If you look at the two statewide races Obama ran in Illinois, you'll see that he has a problem in the Upland South region of his state. First, in his Senate race against Alan Keyes in 2004 and second in the Democratic primary race against Hillary Clinton in 2008, his worst performances were in the Little Egypt region of the state. Is that a Republican region? No. The area went strongly for the Democratic governor in 2004.
That said, I think Obama is going to win with a Southwestern strategy, which makes me wonder if Romney and Richardson aren't the top considerations for VP right now.
Petey, you're entitled to make an argument, but you shouldn't cherrypick polls to do it.
The Zogby poll does show him up double digits, but it's certainly not representative of all polls.Indeed, the Gallup tracking has him up 3. So, from that standpoint he's not running behind his national standing in Virginia or Colorado. The Quinnipiac had him up 7 nationally, and Battleground has him up 2.
PD Shaw, I don't fully buy into your argument. Michael Crowley did a very fascinating analysis of John Kerry's blue collar appeal based on MA and the primaries in early 2004. In the fall, Kerry performed as well as Gore among college-educated voters, but worse among blue collar voters.
| 
Grand New Party: How Republicans Can Win the Working Class and Save the American Dream
Privilege: Harvard and the Education of the Ruling Class
He doesn't convince me. See Fivethirtyeight.com - Obama doesn't have a real chance in the states Noah's calling tossups. Obama has the Kerry states (minus NH, plus IA), and there are only a few states that he has a real chance at picking up. The whole election will come down to Ohio, Nevada, New Mexico, and Colorado. If Obama wins Ohio, he wins; if he wins all three southwestern states, he wins; if he wins Colorado and either New Mexico or Nevada, it's an EC tie and he wins in the House because Democrats control the majority of delegations.
Posted by Asher | May 21, 2008 5:47 PM