The Demographics Of The Dem Contest
By Big Tent Democrat
Speaking for me only
Josh Marshall writes about the demographics of the Dem race:
The problem is in states with substantial but not particularly large African-American populations in which you have a deep-seated and pre-existing racial politics that ends up playing in Clinton's favor. This, if the theory is right, would explain why Obama does well in the Mountain West and the South but has a harder time in states like Pennsylvania and Ohio. . . . But I don't think it adequately deals with all the admittedly small set of data we have. How, for instance, does it account for Obama's victories in Missouri, Maryland, Virginia, Connecticut among other states? . .
This is an issue we have explored in great detail here at Talk Left. For example here. Josh seems unfamiliar with the exit poll data. I am surprised he is pointing to Missouri as the counterpoint, since it actually is in line with the data in other bigger states. Wisconsin (with Virginia) was Obama's best results with white voters in bigger states. More . . .
I earlier wrote:
By my math, Hillary Clinton won about 64% of the non-African American vote in Ohio. African Americans were slightly less than 4% of her total votes. By contrast, Barack Obama won about one third of the non-African American vote. Even then, non white voters amounted to two thirds of Obama's vote. (Compare that to Mississippi where two thirds of Obama's votes came from African Americans.) Indeed, Obama garnered 26% of the white vote in Mississippi as compared to 34% of the white vote in Ohio (the difference largely stems from Clinton winning white Republicans in Mississippi and splitting them in Ohio.) Not that significant a difference, particularly among white Democrats, which were 70-23 in MS and 70-27 in Ohio. The demographics of Mississippi are what led to Obama's win.
It seems clear that if the Ohio result predicts Pennsylvania, Obama will get trounced. Indeed that is what SUSA predicts currently, Obama losing by 19 points, losing non-African Americans 61-29. If we adjust the SUSA number to match the Ohio exit polls (the adjustment primarily would be to up Obama's percentage of the A-A vote), the result would be about 60-40 Clinton. Thus, it seems fair imo to expect a Clinton win in Pennsylvania of at least 15 points.
Another possibility is a return to the Wisconsin model. In Wisconsin, Obama won among all whites 54-45 and African Americans 91-8. If this occurs, Obama wins Pennsylvania handily. The question is will Wisconsin or Ohio be a better model for Pennsylvania? Can Obama recapture his Wisconsin magic?
So I looked at another state, Missouri. Obama captured 39% of the white vote in Missouri and 84% of the African American vote. What can we learn from these three results?
Let's look more closely. In Ohio, Clinton won white Democrats handily but tied among Republicans, who are 99% white. Thus, Obama lost white Democrats in Ohio by 70-27, Mississippi numbers. In Missouri, Obama lost white Democrats by 62-35, won white Republicans by 75-21 and white independents by 59-37. To wit, white non-Democrats were nearly a third of Obama's vote in Missouri.
In Wisconsin, white Democrats voted for Hillary by 51-48, while white independents (26% of the vote) went for Obama 62-35 and white Republicans went for Obama by 72-28. Thus, nearly over a third of Obama's vote in Wisconsin was white non-Democrats.
As we all know, Pennsylvania is a closed primary. Only Democrats can vote in it. If Obama could somehow capture 48% of white Democrats, as he did in Wisconsin, he can narrowly win Pennsylvania. If he loses white Democrats 70-27, as he did in Ohio, he will lose Pennsylvania by 20 points.
Marshall confuses two issues. The number of African- Americans in a state can be dispositive for Obama - as he will carry at least 80% of the vote in almost every state. In Mississippi, that translated into a big win. In Ohio, it translated into a big loss. The demographic breakdowns in Ohio and Mississippi were almost identical. Yet Clinton won Ohio by 10 and lost Mississippi by 22.
In Missouri, Obama won white Republicans and that gave him a better showing. He now seems to be losing white Republicans (a redundancy of course. There are no non-white Republicans.) Luckily for Obama (before it would have been lucky for Clinton) there are no white Republicans in the Pennsylvania primary. Nonetheless, the demographics work strongly against Obama, because 82-85% of the Pennsylvania Dem primary vote will be white. He has to win at least 35-40% of the white vote to win Pennsylvania. That would be a striking improvement for Obama.
Outside of the big contested states, Obama has done much better with white voters. Yes, this does point to the West as fertile ground for Obama. But these very trends do make Obama a bigger risk in Ohio, Michigan, Pennsylvania and Florida. This is not news for anyone who has been paying attention.
| < Open Thread | Evening Open Thread > |





