Brad DeLong: No, I Am Not Miss Informed: Tuesday Hoisted from the Archives from November 2011 with Subsequent Reflections: The Puzzle of Bill Galston Weblogging
Back in November 2011, I wrote that Richard Just was not doing a good job as editor of the New Republic: Am I Miss Informed?: American Elections "Department of 'Huh?!'" Department. I was annoyed by his publishing Bill Galston's insistence that:
Of the states that Obama won in 2008, he is certain to lose Indiana, he will be hard-put to reproduce his razor-thin edge in North Carolina, and his chances of prevailing in Florida appear well short of 50-50. Those three states alone accounted for 53 of Obama’s 365 electoral votes in 2008. Given all this, it would political malpractice for the Obama campaign not to go all-out in Ohio. At the same time, they should focus on fortifying the president’s standing in Pennsylvania, a state that traditionally has given Democratic presidential candidates a share of the popular vote about two percentage points higher than their national average. Winning Pennsylvania is a necessary condition of Democratic victory; winning Ohio is a sufficient condition…
This seemed to me--and to everybody else I know who actually looked at electoral demographics and the demographic map--wrong.
So I went back through the internet, looking for any explanation from Richard Just as to why he had chosen Bill Galston rather than, say, Nate Silver or Sam Wang, as his electoral strategy guru. I have found nothing. I went back through the internet, looking for any explanation or apology from Bill Galston for talking trash about issues he was underbriefed on. I have found nothing.
Can anybody help me out here--either with an explanation of why Galston believed then what he believed, or with a link to his reflections and updating-of-beliefs in the wake of 2012?
For Galston was wrong. Galston was wrong because:
- If Pennsylvania were the state in play, Ohio, Virginia and Colorado would already have been lost with high probability, and winning Pennsylvania would only get Obama to 263.
- If Ohio were the state in play, Pennsylvania, Virginia, and Colorado would already have been won with high probability, and even with losing Ohio Obama would have 303.
There is an argument for spreading effort among all seven credible possible swing states--New Hampshire, Iowa, Pennsylvania, Colorado, Virginia, Ohio, Florida. There is an argument for concentrating effort on Colorado and Virginia. There is no argument for prioritizing Ohio and Pennsylvania over the key two, or as more important than the rest of the key seven, is there?
Here is my original piece:
Why oh why can't we have a better press corps?
And here comes Richard Just's called third strike as editor of the New Republic...
What has he done? He publishes Bill Galston, rather than asking Bill Galston to go back and explain his--rather strange--reasoning.
I, you see, had thought that the states most worth cultivating for Obama were Iowa and Colorado (and Virginia and New Hampshire) rather than Ohio and Pennsylvania. If Ohio were in play, it would be very likely that the 2012 election was already won for Obama. If Pennsyvania were in play, it would be very likely that the 2012 election was already lost for Obama.
If you think the election is going to be very close, you are much better putting resources into Iowa and Colorado (and Virginia and New Hampshire) than into Ohio. And if you think you have to put resources into Pennsylvania the election is probably lost anyway: better to hope things break your way, in which case resources should be put into Iowa and Colorado (and Virginia and New Hampshire)
Am I wrong?
Bill Galston says that I am. But I don't think he has done his arithmetic:
Forget the 2008 Map: A New Poll Shows Why Obama’s Re-Election Is Riding On Ohio: As the headline in Thursday’s Politico boldly touted (“Ohio back on Obama’s dance card”), the Obama campaign is suddenly refocusing on the Buckeye state. There’s a positive reason for this reported shift in the Obama campaign’s thinking: Coupled with the rebuke Ohio swing voters administered on Tuesday to an overreaching Republican governor, Mitt Romney’s lack of populist appeal makes Ohio a more tempting target than it appeared just a few months ago. But there’s a negative reason as well: “Virginia and North Carolina, key to Obama’s victories in 2008,” the article continues, “are becoming more and more uncertain.” Indeed, if Obama hopes to win reelection, he needs to double down on Ohio, Pennsylvania, and the rest of the heartland—and acknowledge that his “new majority” coalition—upscale professionals, single women, young adults, and minorities—won’t be enough to get the job done….
Of the states that Obama won in 2008, he is certain to lose Indiana, he will be hard-put to reproduce his razor-thin edge in North Carolina, and his chances of prevailing in Florida appear well short of 50-50. Those three states alone accounted for 53 of Obama’s 365 electoral votes in 2008.
That would leave him with 312--42 more than 270.
Given all this, it would political malpractice for the Obama campaign not to go all-out in Ohio.
But why not "go all-out in Colorado"--which, if 2008 is your guide, is 4.4% points easier for Obama to win than Ohio? Why not Virginia--which, if 2008 is your guide, is 4.4% points easier for Obama to win than Ohio? If Obama holds Pennsylvania, New Hampshire, and Iowa, either Colorado or Virginia would do without Ohio.
At the same time, they should focus on fortifying the president’s standing in Pennsylvania, a state that traditionally has given Democratic presidential candidates a share of the popular vote about two percentage points higher than their national average. Winning Pennsylvania is a necessary condition of Democratic victory; winning Ohio is a sufficient condition.
And why Galston's focus on Pennsylvania? Pennsylvania is very close to necessary, indeed. But Minnesota and Wisconsin are very close to necessary as well, and if Pennsylvania cannot be assumed to be in the Democratic column why aren't Minnesota and Wisconsin shaky as well?
As I’ve argued before, the president’s path to victory in 2012 runs through the heartland, not around it. And that means addressing the concerns of heartland voters, who mirror the demography of the U.S. population more closely than they do Obama’s “new majority.”
The overwhelming impression is simply that Galston did not do the electoral arithmetic, and Richard Just did not tell him that he really needs to do the electoral arithmetic if he is going to write this kind of piece.
UPDATE: A correspondent writes:
No, you are not wrong.
If the states maintain their 2008 rank ordering of relative two-party votes, the swing state is Colorado. If Colorado goes Democratic, then the Democrats have won with 272 electoral votes. If Colorado goes Republican, the Republicans have won with 275 electoral votes.
If Colorado and Iowa switch their places, then the election is tied if the Democrats take Colorado and the Republicans take Iowa.
On the other side, the state in which Obama's margin was smaller than Colorado's was Virginia. If Virginia and Colorado switch their rank order and the Republicans take Colorado, the Democrats win with 276 electoral votes.
If the states maintain their 2008 rank orderings, then if Ohio is in play the Democrats have already won with 285 electoral votes. If Pennsylvania is in play then the Republicans have already won with 285 electoral votes. Pennsylvania and Ohio are barely swing states. The more likely swing rates are New Hampshire, Iowa, Colorado, and Virginia.