Weekly Polling Report: Winning in the South
by CA Pol Junkie
The South has become the anchor of Republican presidential politics, as Reagan and Bush II embraced cultural conservatism. Bill Clinton won a number of Southern states in 1992 and 1996, in part because of his charisma and Arkansas roots and in part because neither Bush I nor Dole aligned themselves strongly with the South.
Winning in the South will not generally be a matter of swaying undecided voters. Although Obama's performance in the primaries shows potential for doing well among young and well-educated whites, the key to winning is based on who votes and who doesn't. Exit poll data shows that there are four important voting blocks: African-Americans, Hispanics (FL, TX), young whites, and "older" (30+) white voters. Barack Obama will do very poorly among older Southern whites, but he will get nearly unanimous enthusiastic support from African-Americans. Except for Oklahoma, young whites in the South are voting 5-43 points more Democratic than their parents. John McCain will still win a very high percentage of white Southerners' votes, but turnout might be depressed because nobody is enthusiastic about him and conservatives don't trust him.
Polling in the South will be of limited use because so much depends on the likely voter screen used by each pollster. By looking at reweighted 2004 exit polls, we can analyze the potential effect of shifts in turnout patterns and generational change. A few assumptions were made for this analysis: African-Americans will vote 95-5 for Obama and increase their turnout rate by 10%; young minority voters vote like their parents; half of 30-34 year old white voters will retain their voting pattern from the 2004 election while the other half will vote like their parents.
Listed below are the Southern and border states most likely to go blue this year according to my analysis. Also shown are the percent chance of Obama winning each state as according to FiveThirtyEight.
Florida (2004 exit poll) 538 Percent Obama: 27%
In 2004, Floridians from 18-29 voted for Kerry by 17 points while he lost the state overall by 5. Even if young voters don't increase as a share of the population, the voting pattern from 2004 will give Obama a 0.2% win in the state. There are two additional factors in Obama's favor. First, Florida had lifetime voting bans for all ex-felons, but Governor Crist was able to restore voting rights for most of them. Since those with their rights restored are disproportionately African-American, Obama will net extra votes. Secondly, indications are that the Cuban-American population is continuing its Democratic shift, from 80% Republican to 70-76% Republican in 2004. The crosstabs of current polls show Obama leading among Florida Hispanic voters by about 6%, while Kerry lost by 12% in this group.
UPDATE: DailyKos has a nice report on voter registration in Miami-Dade County rapidly shifting Democratic among whites, African-Americans, and Hispanics.
Virginia (2004 exit poll) 538 Percent Obama: 49%
Obama's most promising polling in the South has been from Virginia, where he has had a tiny but consistent lead. The primary cause is likely the realignment of northern Virginia, which is well educated and not so culturally Southern. Young voters here are strongly more Democratic than older voters, so partial maintenance of party loytalty into their 30's will help. Without a shift in older white voters, the 2004 pattern would extrapolate to a loss by 4.7% in November. If Obama can reduce McCain's margin among 35+ year old whites to 66-33 (Kerry lost by 41), he can pull of the win here without increased 18-29 turnout.
Missouri (2004 exit poll) 538 Percent Obama: 24%
18-29 year olds favored Kerry by 3 even as he lost by 7. The part of the state which pronounces it "mizzurah" is culturally Southern, while the part where it's pronounced "mizzuree" is more Midwestern. Since the voting isn't so polarized here, this analysis has its limitations. Under 2004 voting patterns, Obama would reduce Kerry's loss margin from 7 to 5. To win, Obama would need to reduce his loss margin among whites by 8, which is a challenge without a Virginia style realignment.
North Carolina (2004 exit poll) 538 Percent Obama: 24%
John Kerry lost North Carolina by 12 points, but it shows very promising demographic change. Voters 18-29 preferred Kerry by 13%. The Research Triangle area is known for its large number of well educated white voters. A 44 point loss among white voters (as opposed to Kerry's 52 point margin) combined with 5% lower turnout rate among whites would be needed to turn this one blue.
Arkansas (2004 exit poll) 538 Percent Obama: 6%
Democratic presidential candidates perform relatively well here (compared to other Southern states) among older white voters: Kerry lost this group by "only" 30 points. Still, the African-American and youth populations aren't high enough to bring this state in for Obama: he would need a stronger youth vote combined with older white turnout depressed by 10% to win here.
South Carolina (2004 exit poll) 538 Percent Obama: 8%
30% of South Carolina voters are African-American, and only an overwhelming turnout combined with depressed white turnout would give Obama a chance. Young whites favored Bush over Kerry by 36%, an improvement over their parents' 80-20 margin but not enough to give Obama much hope.
Mississippi (2004 exit poll) 538 Percent Obama: 5%
Mississippi is a wild card because the 2004 exit poll data showed it to be the most polarized state of all. According to the poll processed with the assumptions above, whites over 30 voted for Bush over Kerry by a margin of 94-6. The exit poll showed a huge disparity with the young, as white voters 18-29 only chose Bush by 3 points over Kerry. That extent of polarity is hard to believe, but how far off can it be? Mississippi also has the highest African-American concentration of any state, though (thanks in part to ex-felon disenfranchisement) in 2004 African-Americans didn't vote at quite as high a rate as they occur in the population. If African-Americans and the young were to vote at a high rate while 10% of bored McCain voters stay home, the state would go to Obama, but that's probably too much to ask. Sean at 538 gives an even more bleak assessment.
Louisiana (2004 exit poll) 538 Percent Obama: 5%
Young whites in Louisiana are almost as Republican as their parents, having voted for Bush by about 71-29 over Kerry. Between that and Hurricane Katrina scattering the most Democratic voters in the state, there is very little chance at all for Obama.
Georgia (2004 exit poll) 538 Percent Obama: 11%
In 2004, African-American turnout was about 10% lower than their population. Obama has put alot of resources into organizing here, which will hopefully reverse this trend and help on down ballot races. Even with a massive increase in African-American turnout (to overperform relative to population by 10%) and depressed white turnout, it is very difficult to see how Georgia could turn blue.
West Virginia (2004 exit poll) 538 Percent Obama: 18%
West Virginia is 95% white, so high African-American turnout won't do it for Obama. The good news is that 18-29 year old voters chose Kerry by 4 points while he was losing the state by 13. The bad news is that Obama does very poorly in the Appalachian region, so it's hard to see him winning over enough white voters over 30 to win.
Texas (2004 exit poll) 538 Percent Obama: 6%
This state is a demographic time bomb. Everyone knows its days as a red state are limited, but nobody knows when it will go blue. The primary change will be from Hispanics becoming a high enough percentage of the electorate to influence the outcome of statewide elections. Another uncertainty for this year is how much of Bush's success came from being a home state candidate versus the generally Republican character of the state. Even discounting the Republican margin among whites by 6 points and giving Obama the Hispanic vote by a 70-30 margin, Obama is still going to lose by about 8 points.
Tennessee (2004 exit poll) 538 Percent Obama: 3%
Young white voters in Tennessee still favored Bush over Kerry by 21%, only 16% of the state is African-American, and part of the state is in the famed Appalachian region where Obama tanks. Not going to happen.
Alabama (2004 exit poll) 538 Percent Obama: 0%
The racial polarization of voting here is second only to Mississippi, but young voters are maintaining more of their Republican roots and there are fewer African-Americans. Losing by 20 points here is a reasonable expectation.
The poll averages used are from pollster.com. A 10 point lead puts a state in the base for each candidate, which means it's not going to be the state to put one candidate or the other over 270. Since those states are boring, their status is not updated here.
The current pollster.com prediction would be Obama 312, McCain 226 with Ohio as the state that puts Obama over the top and Colorado as the closest state. This week, Minnesota moves into the competitive zone. Although most polls showed a shift toward McCain, polls in Oregon and Alaska moved in Obama's direction. The latter is now a blue state according to the most recent poll.
Obama Base (190 EV): California, Washington, Hawaii, Wisconsin, Illinois, DC, Maryland, New Jersey, New York, Connecticut, Rhode Island, Massachusetts, Vermont, Maine
Competitive states, cumulative electoral votes, and new polls:
Minnesota (Obama +9.4) 200 EV
Rasmussen 8/13 (7/22)
Obama 49 (51)
McCain 45 (39)
Franklin & Marshall College 8/4-10 (2/13-18):
Obama 46 (43)
McCain 41 (44)
Rasmussen 8/7 (7/14)
Obama 52 (49)
McCain 42 (40)
Rasumssen 8/7 (7/10)
Obama 49 (51)
McCain 44 (41)
Public Policy Polling 8/12-14 (7/17-20)
Obama 45 (48)
McCain 45 (40)
Hays Research 8/6-7
Virginia(Obama +0.9) 303 EV
Colorado(Obama +0.2) 312 EV
Rasmussen 8/13 (7/21)
McCain 49 (47)
Obama 48 (50)
Insider Advantage 8/11
Rasmussen 8/11 (7/16)
McCain 48 (45)
Obama 45 (47)
McCain Base (82 EV): Arizona, Utah, Idaho, Wyoming, Nebraska, Kansas, Oklahoma, Louisiana, Alabama, Tennessee, Kentucky, West Virginia