NEW REPORT: The Race Chasm & the Clinton Firewall

by: David Sirota

Mon Mar 31, 2008 at 13:03

In my newspaper column on Friday, I touched on a little-explored phenomenon in the 2008 presidential race. Amid all the punditry and intricate television graphics showing delegate counts and precinct voting trends, almost no one has bothered to look at the overwhelming dynamic that deals with race. Specifically, while Barack Obama has won states with both almost no black population and and very large black populations, he has had trouble winning states with a modestly sized black populations. How pronounced is this trend? I answer that question in a new In These Times investigation about what I call the Race Chasm - a trend that has been almost completely missed by the media. You can find the piece here, or at the backup location here (In These Times servers crashed from traffic to the piece this morning). This chasm is the key pillar of Hillary Clinton's much-vaunted "firewall."

As you can see from the key graph included with the piece, the states Hillary Clinton has won have been mostly those with black populations above 6 percent and below 17 percent of the states' population. In all, 75 percent of Clinton states have this demography. Obama has been only able to eke out three victories in these Race Chasm states (as an FYI - the graph omits 9 states for very obvious reasons elucidated in the In These Times article).

Why is this trend so pronounced? There's no scientific answer to that question, but as I say in the In These Times piece, it probably has something to do with the state of black-white racial politics.

David Sirota :: NEW REPORT: The Race Chasm & the Clinton Firewall
In super-white states, black-white racial politics barely exists, and therefore racism or subtle race-coded messages are not all that devastating a weapon against a black candidate - especially in a Democratic primary. In states with large black populations, black-white politics is very intense, but in a Democratic primary, the black vote can offset a racially motivated white vote. But in the states in between, black-white politics is equally as intense, but the black population is not big enough to offset a racially motivated white vote.

This isn't to say that race is the only factor, nor that every white person voting against Barack Obama and for Hillary Clinton is a racist or racially motivated - not at all. However, over the course of 33 separate elections, a trend like this is significant - and probably explains why the Clinton campaign has been working hard to keep the race issue at the forefront of the campaign. Barely a week goes by without some Clinton surrogate - or, in last week's case, Clinton herself - reminding the electorate of Obama's ethnicity. That's not an accident - that is an effort to maximize the Race Chasm, especially with Pennsylvania, Indiana and Kentucky coming up.

Last week, New York Times columnist Bill Kristol said, "The last thing we need now is a heated national conversation about race." But really, the last thing we need are more wealthy white pundits sitting in the comfortable confines of their plush Washington, D.C. offices telling us that race shouldn't be talked about. As the Race Chasm shows, now is precisely the time we need a national conversation about the divisions that still afflict our society and culture.

Read the whole In These Times article here, or at the backup location here. As I said, this phenomenon has not yet been reported on in any comprehensive way. But I think you will agree that the graph is a troubling image showing just how powerful the race issue has become in the 2008 election.

UPDATE: I appeared on Jay Marvin's AM 760 drive-time radio show here in Colorado to discuss the Race Chasm. You can listen here.

Tags: , (All Tags)
Print Friendly View Send As Email

too broad (4.00 / 1)
Treating states as a unit is too broad a measure. To extract behavior one would need to examine patterns at a finer grain. This might be legislative districts or even zip codes.

For example a state with a middle range of blacks might have them so distributed that their voting power is diluted and, thus the state appears to be solidly "white". The inverse might also be the case. The blacks are all gerrymandered into a few majority minority districts and thus their power as a proportion of the voters is diluted.

Those who create house districts use data like this all the time, so the effect of racial grouping is well understood by those who make it their business to do so.

Policies not Politics

SD's in Texas (0.00 / 0)
I think that the experience of Senate Districts in Texas might bear out this thesis. But, I think that class may be a bigger factor than race.



[ Parent ]
this is a really interesting stat, and the analysis makes sense (0.00 / 0)
but I'm reluctant to make this kind of generalization. we haven't had nearly enough states, and there are plenty of other factors that could be at play. I mean, RI, NJ, and MA are all close to New York, where Clinton represents. You dropped AR, NY, and IL from the analysis but not those states.  FL is in there, but that could easily be explained by Obama not campaigning there, and not really having much of a bump to carry him there.

I agree that we'd need to probe a bit deeper to know if this really is racial politics at play.  

33 separate elections (4.00 / 1)
This graph documents 33 separate elections. I'd say that's a pretty huge amount of elections.

[ Parent ]
within a single primary season though. (0.00 / 0)
you're testing far more variables than just the race of the candidates, obviously.

I think there's support for your conclusion. I'm just constantly skeptical of everything, that's all.

[ Parent ]
Too many states left out (4.00 / 2)
CA, TX, AZ, NM, NY, AR, IL, MI, and on and on.

I understand your rationale, but that's a large population and one large region (Southwest) that has been left out.  And I agree that class is as important.  At most there has been a dog-whistle campaign that she hoped would give her 50% + 1.  

John McCain--He's not who you think he is.

[ Parent ]
Why? (4.00 / 1)
I don't understand why those aren't included in the graph.

I support John McCain because children are too healthy anyway.

[ Parent ]
He spells out the reasoning in detail... (0.00 / 0)
... at In These Times:

To date, 42 states and the District of Columbia have voted in primaries or caucuses. Factor out the two senators' home states (Illinois, New York and Arkansas), the two states where Edwards was a major factor (New Hampshire and Iowa) and the one state where only Clinton was on the ballot (Michigan) and you are left with 37 elections where the head-to-head Clinton-Obama matchup has been most clear. Subtract the Latino factor (a hugely important but wholly separate influence on the election) by removing the four states whose Hispanic population is over 25 percent (California, New Mexico, Texas and Arizona), and you are left with 33 elections that best represent how the black-white split has impacted the campaign.


[ Parent ]
Interesting... (4.00 / 4)
Judging by the demographics alone, Clinton should thus win Pennsylvania, Indiana, and Kentucky, while Obama should win North Carolina, Oregon, West Virginia (!) South Dakota, and Montana.  Puerto Rico is anyone's guess, since it's 8% black, but another 12% of the population is of partial African ancestry.  

But I think your model is too simple: your glib dismissal of West Virginia shows this honestly.  As does looking at the county by county breakdown of votes in Virginia, Maryland, Tennessee, or Ohio.  In all cases, Obama does worse in the Appalachian counties than would be expected by the statewide totals among white voters.  Even in Mississippi, where Obama lost the white vote overall, he seems to have done far worse in Northeast Mississippi (which is Appalachian-like), than he did in Southeast Mississippi.  

These are not areas with significant African-American populations.  These are areas with virtually no African Americans.  However, they are areas that have had "black-white racial politics" at least insofar as there is still deep-seated racism in large portions of this region.  

Furthermore, one would expect if this were the case that Clinton would have done well in the "collar counties" surrounding Milwaukee and Minneapolis St. Paul, as these are clearly areas racial politics have played a role in the past.  Minneapolis shows no such association, and Milwaukee only perhaps a weak one.  Also, in Missouri, where Obama essentially only won the urban areas, he lost the immediate collar counties by a much smaller margin than in the rural parts of the state.  

So, one could easily posit that Obama does badly among Appalachian whites, and more broadly culturally-southern whites in rural areas.  This alone is probably enough to account for part of his losing Ohio, Tennessee, and Oklahoma, and Florida, and certainly played a role in why Missouri was so close.  Once you take those out, you're left with a narrow loss in Nevada, and a handful of Northeastern states, all of which have fairly strong party machines which could have played a decisive role as well.  

i have to say (4.00 / 1)
This is the most interesting bit of analysis I've seen in a long time; that trend seems quite real, actually; it's rare for things to line up that well. Well spotted!

What about %white vote for Obama as a function (0.00 / 0)
of %black population.  You could break this down by congressional district to get smooth data.

You could then get a nifty function to predict the total vote % for Obama:

Vote(Obama) = 0.9 * (% population that is black) + White Vote

Where White Vote = Function(%population that is black) as determined above.

Patterns in random data (4.00 / 1)
The human mind tends to find patterns in more-or-less random information.  I'm not saying that the argument is wrong, or that there isn't a pattern, but the data is not as strong as it looks.

Omitting 9 states for the various reasons given seems like Clinton trying to discount caucus states, or small states, or states with fewer electoral votes, etc.  Omitting Michigan obviously makes sense, and maybe the candidates' home states, but I don't see a sound reason for omitting the early states where Edwards was still in or omitting states with a large Latino population (after all, aren't race issues in those states even more pronounced?)

So, it really seems like finding what may be a pattern, or may just be coincidence, then fudging the data to make that pattern more pronounced.

Conduct your own interview of Sarah Palin!

dont think so. he used a fairly large sample here. n/t (0.00 / 0)

End this war. Stop John McCain. Cindy McCain is filthy rich.

[ Parent ]
Are you serious? (4.00 / 3)
Thirty-some points is not a large sample; it's tiny.  For statistical samples, what matters is its absolute size, not the percentage of the population sampled. To make matters worse, more than 20% of the data was dropped. In a sample this small questionable data should be labeled as such (possibly being plotted in a different color,) rather than simply thrown away.

Also, the X-axis is in arbitrary units. In other words each state is plotted by its relative position in the series rather than its actual percentage of AA citizens. A finer grained data set (county-level data, perhaps) plotted on non-arbitrary axes would be far more convincing.

[ Parent ]
but the trends shown here are so unambigious they cannot be mistaken n/t (0.00 / 0)

End this war. Stop John McCain. Cindy McCain is filthy rich.

[ Parent ]
Aaaargh! (0.00 / 0)
The problem is not the pattern is not apparent.  The problem is that the apparent pattern may not be a real, meaningful pattern.  The posted data does not convince me: I know very little about statistics, but I know impoverished data when I see it.

Further analysis is warranted, as I suggested above, but no conclusions should be drawn from this graph.  

[ Parent ]
you hit the nail right on the head again David. n/t (0.00 / 0)

End this war. Stop John McCain. Cindy McCain is filthy rich.

quick thought... (4.00 / 1)
doesn't this graph also kinda prove that were Obama not a black candidate or were nobody racist that he would easily defeat Clinton, thus making him the far superior candidate?

I suppose it could maybe be used as an argument against him in the general election.

But this sure as hell flies in the face of what Geraldine Ferraro was saying about his color being a boost for him!

End this war. Stop John McCain. Cindy McCain is filthy rich.

Quick response (0.00 / 0)

If nobody were what you call "racist" (I prefer racially biased when I talk of democratic voters), Obama would only get around 45-55% of the black vote and Hillary would easilly defeat him.

[ Parent ]
even though he'd be getting FAR MORE of the white vote? n/t (0.00 / 0)

End this war. Stop John McCain. Cindy McCain is filthy rich.

[ Parent ]
If Obama wasn't getting (4.00 / 1)
the black vote "Hillary would easily defeat him"?

That argument is rather spurious.  Consider the point you make about the reasoning for African Americans giving their support to Obama.  If you want to make that argument you also must make the same argument for Clinton.  The largest base of support for Clinton is the gender factor: women overwhelmingly support Clinton.  I think the argument can be made that both candidates get support from sectors of the voting public that are easily identified by race/gender biases.  

The really interesting thing about this is that Obama continues to do well, and is beating Clinton even though the majority of voters are women!

[ Parent ]
The Social Theory behind the observations... (0.00 / 0)
The pattern is not all that different from what was measured back in the 60's and 70's when we studied successful and unsuccessful strategies for integrating housing, and developing the data to back up policy and in some cases Court Cases in the field of fair housing.  The same pattern also occurred in school integration efforts.

White populations will buy into integration, and shared social space (in housing, neighborhoods, and schools) when they are not really challenged for dominance in number, or social place in the power and influence hierarchy.  Thus, it is much easier to intergrate an upper income neighborhood or apartment/condo complex, if price, or let's say co-op building rules limit the probability that the number of minority race or excluded ethnic/cultural persons will follow the first to move in, buy in, whatever.  There is what was always called the "Tipping Point" -- that point at which either number or matters of social/cultural dominance seems to cause the majority population to reject the idea of integration of social space, and move, change schools, whatever is necessary to reject and leave the social space. The point at which racial and/or ethnic conflict is most likely to occur is after the Tipping Point has been reached, as people fear a loss of status or dominance of the social control system, but have not yet been able to resettle, and recapture control of their circumstances. (You can illustrate this by recalling that during the urban riots beginning in 1965, and ending or at least diminishing in the 70's, Urban Police Forces remained almost entirely white, (Chicago, Newark, Detroit) while "white Flight" was the dominant pattern in housing and education.)  While some of this may have changed since the 70's, I would suggest that the voting patterns are simply a continuation of the underlying social and cultural pattern.  While it is asserted that the younger generation have different attitudes and attidudenal/behavor patterns from parents or grandparents, I would like to see serious empirical research to back this up -- and I don't see all that much of it.

For instance, I live in Keith Ellison's District in Minnesota, and strongly supported his candidacy in 2006.  But I also know well that given this district, which is about 90% White, he lost about 15 points of the DFL vote he should have received, given the past voting patterns.  When I look, precinct by precinct for where the greatest loss occurred, it was in those where one could describe a "tipping point" and easily illustrate it.  The previous two DFL Representatives, Don Fraser (62-78) and Martin Sabo (78-06) normally carried the district by 70+ percentage of the vote, but Ellison won with 56%.  I expect much of that difference to disappear in 2008 -- largely because Ellison has tended to the interests of the whole district.  Just the same, one would be a fool not to comprehend that this phenonema exists.  I strongly support Obama, but I also believe to serve his candidacy well, we have to execute a campaign that overcomes these realities. Argueing
racial dynamics don't exist (or still exist) is just foolish. There is a huge body of social theory about Racism and the way it conditions behavior, and one of the ironies of it all, is that much of it rests on Sociological Research done over the years at the University of Chicago.  


Tammy Lee? (0.00 / 0)
"I expect much of that difference to disappear in 2008 -- largely because Ellison has tended to the interests of the whole district."

And....because Tammy Lee of the Independence Party is unlikely to run.  While I've only lived in this district for seven years, I do not recall Sabo ever having as substantive a challenger as Lee.

I see Ellison's win in 2006 as a fine example of how race was not a major issue in the race - he took 56% of the vote against two white candidates in a district that is 90% white.  Lee was not just a "cop-out" vote for whites that couldn't bring themselves to vote for Ellison - she was a strong candidate that did very well in debates and the media.  

"It sounds wrong...
     ...but its right."

[ Parent ]
Look at the actual vote, comparatively, precinct by precinct (0.00 / 0)
Lee was a Green Candidate, and normally in a CD district wide contest might get 3-5% of the vote.  Greens had gained automatic ballot access and then lost it after 2004 because they no longer were achieving that 5% cut-off.  

In 2006, Lee got about 19% of the vote, so to understand it you have to look at those precincts that had lower than expected DFL/Ellison margins, and where Lee gained votes way over what would be expected of a Green endorsee.  I would suggest that 15% of the vote -- or all but 4% of Lee's support was a response to Ellison being both Black and Muslim.  I would agree it will disappear in the future, (election 2008) because the fears of 2006 will have been dissipated by Ellison's performance in office.  But I think it was real in 2006, and while one may not share in the fears, any election strategy worth a few cents needs to be realistic about the problem.  

As to Lee -- I think one also has to deal with that campaign by realizing that the DLC through their founding member in Minnesota, Former State Senator Ember Reichgott Junge, brought outside money into that campaign, much of which was used for TV and Radio advertising.  That support was hardly from Green Party Members in Minnesota.  Unless we understand that the Democratic Party, and even the DFL have factions that will play these kinds of politics (Race and Religion), we will be caught making assumptions not grounded in reality.  

Yes, Minneapolis is better off now that it was in the late 1960's when the City turned down the DFL Nominee for Mayor, Harry Davis, (Davis was the first black member of the School Board, first elected about 1962), and instead in the Primary nominated a loud mouth law and order cop for Mayor.  The story of Charlie Stenvig is indeed part of our "Liberal" city's history.    

[ Parent ]
Lee was not a Green (0.00 / 0)
She was from the Independence Party.  Part of Team Minnesota.

"It sounds wrong...
     ...but its right."

[ Parent ]
industrialization (0.00 / 0)
Industrialization is another feature of the states you place in the "race chasm." But if it makes you feel better to think that race politics is driving Democratic voters away from Obama, knock yourself out.  

Averting the eyes (4.00 / 1)
Inevitably, there will always be those who will do anything to avert their eyes from the topics that make them uncomfortable.

[ Parent ]
You are fooling yourself if (4.00 / 1)
you don't believe race will be an issue.  I am shocked by the people who "admit" they won't vote for him because he is black.  I always respond with, what about his white half?  Will you vote for that side?  IF Obama cannot obliterate the race divide, who the hell can.  I would not want an Al Sharpton or a Rev. Wright to be my President; but I would be proud as hell to have Obama in the WH.  Some of his policies, well...  That another discussion.  

[ Parent ]
President Sharpton? (4.00 / 2)
Well, it would not be a boring administration, that's for sure.  Might even start looking forward to the State of the Union Address.

"It sounds wrong...
     ...but its right."

[ Parent ]
bad links (0.00 / 0)
The links in your first paragraph don't link to the In These Times article, they link to a piece on Rev. Wright here:

Don't know what to think about this.  I may try to do some number crunching as suggested in some of the other comments and see how it looks.


Open Left Campaigns



Advanced Search

Powered by: SoapBlox