Tuesday's BOCC contest: A race about race

It's taken a bit longer than I hoped to get some more thoughts about Tuesday's BOCC race up here on the blog, due to my travel commitments. In the interim, Ray Gronberg -- who raised some interesting questions here in the comments section yesterday -- has a very good analysis of the race in today's Herald-Sun.

Gronberg makes two central arguments that will be the source of discussion for weeks to come among local politicos: that the election was heavily racially polarized, and that the People's Alliance seemed to have a hard time getting energy behind their slate.

As the H-S story notes, the presence of early voting sites has made it a bit trickier to project turnout by race based on voter appearances at their geographic precinct. However, we do know that in this election, based on state-wide data, that the presidential race was highly stratified by ethnicity, with black voters supporting Obama more strongly than white voters. In Durham, there were 11 precincts where Obama won over 90% of the votes; 12 more where he garnered more than 80%. Only in three -- Bahama, Rougemont, and Neal Middle School, which incorporates rural Eastern Durham County -- did Obama actually lose to Clinton.

If we use the strength of support for Obama vs. Clinton as a proxy for the racial makeup of each precinct in Tuesday's election, and then reassign percentages based on what we'll call "top racial affinity" (precincts that likely strongly matched the BOCC candidates' own race) to "low racial affinity" (precincts that were likely comprised largely of voters not of the candidates' own race), you get the following interesting picture:

(Click the image above for a full-size version in a new window.)

In the graph above, the range or separation between the four data points for each candidate shows the relative level of racial disparity likely present in these candidates' support, though again since race is present only through correlation with Obama-Clinton votes, this is a rough guess at best.

Still, the data support the trends seen in other measures, like geography. Michael Page did well across the board, coming in first even in likely majority-white districts. Becky Heron and Ellen Reckhow trailed Page in majority-black districts, but still did notably better in those areas than Don Moffitt, Josh Parker or Doug Wright did.

But there were a lot of likely majority-black districts, and they turned out votes disproportionately to majority-white districts, it appears. 17% of all BOCC votes were cast in precincts where Obama earned 90%+ of the presidential vote, and an additional 39% of BOCC votes came in locations where Obama earned 80-90%. Just 9% of all BOCC votes, on the other hand, came in precincts where Obama scored 60% of the vote or less.

Witness, to that end, Joe Bowser's divided victory. In likely heavily-black precincts, Bowser was a close #2 to Michael Page. Yet he fell off relatively quickly in the remaining precincts, doing no better than and often worse than Heron and Reckhow in districts that were not as solidly majority-minority in turnout. In fact, it's instructive to compare Fred Foster's graph with Bowser's. Both were endorsed by the Durham Committee, yet Foster fell to a sixth place finish.

The key difference? Neither polled all that differently in districts that were less likely to be Committee strongholds, as measured by the yellow, blue and red lines on the graph. Yet Foster trailed in the strongly-black precincts by a couple of percentage points, and that was enough, it seems, to push him to sixth.

If there could be a unity candidate in this election besides Page, it's Brenda Howerton, whose graph converges tightly around the 10% mark across districts. Howerton trailed Victoria Peterson slightly in the strongest-black precincts, but bested her in other areas, notably those with low and medium likely ethnic affinity. We suspect Howerton is the candidate who benefitted most from the Indy and PA endorsements; Page was clearly safe without them, but that broader, cross-racial support appears to have helped Howerton to a crucial fifth seat on this year's BOCC, narrowly edging Foster out of a seat.

So what happened to Don Moffitt? As the blue line demonstrates, voters in largely-black districts didn't back Moffitt at all, a problem shared by fellow challengers Josh Parker and Doug Wright. But Moffitt also didn't benefit from notably strong support in less strongly minority precincts, either. Of course, even strength in those precincts this year likely could not have overcome the weakness in the core Obama/Committee precincts, given the massively disproportionate turnout there -- but it still appears to be a sign that Moffitt had a harder time attracting voters that should have been in his core than Reckhow and Heron did. (Compare their points on the red, yellow and green lines to see what I mean.)

In summary there seem to be a few things going on here:

• While early voting skews the data, the Obama correlation suggests that black voter turnout was tremendous.
• The Committee appears to have been largely successful in getting its slate out in strongly minority districts and in its three-candidate strategy, as Gronberg noted in today's paper.
• There'll be a lot of head-scratching over how the well-financed Parker and Moffitt campaigns failed to draw more support than they did, even in districts it seems they should have won.
• Will PACs respond to what appears to be a challenge to their influence? Not that that challenge is, I believe, as large as Gronberg implied in the H-S -- the Committee was only really key at the margins last fall with Farad Ali, who also enjoyedsigni ficant support outside the Committee in traditional PA strongholds. And this Obama factor is not to be expected to recur in most election cycles. Still, expect some soul-searching by the PA and Friends in the months to come.

We'll repeat the analysis above for the DA race and school board race later this week.

Comments

[Elizabeth T.]

Kevin,
Do you (or others) have any data that shows the relative number of voters in each precinct in this primary as opposed to other elections? I think it would be interesting to see where the largest jump in votes came from.

[Michael Bacon]

Great data, Kevin, but I have one quibble -- it sounds (I can't quite tell) if you're using percentage for Obama as a measure of racial composition. That seems pretty sketchy to me, unless you're using category variables and sectioning off districts where Obama got, say, 90% or more. If you're using it as a continuous variable, I'd say that risks invalidating the whole thing.

If you want a racial breakdown based on census numbers for every voting precinct in Durham, if one doesn't already exist, I can try to work one up. (I've got ArcGIS on my work computer again. Yay!)

I haven't read Ray's article, but pace the conversation about the PA in the other topic, I'll hopefully have a blog post up about that in the next couple of days.

[Mike]

I hope Josh Parker will remember to pick up all of his signs that were put everywhere. UGH, that really turned me off.

[Frank Hyman]

Good work on the graph Kevin.

( It might help some readers to know that the green line refers roughly to white precincts when looking at a white candidate and black precincts/black candidate and the blue line refers roughly to the reverse: white precincts/black candidate and black precincts/white candidates , etc. Is that clear as mud?)

Your chart indicates several things worth noting:

* the white candidates endorsed by the Friends--Reckhow, Heron, Moffit and Parker--have divergent totals in white precincts (green line) which comes from lots of Friends voters doing a double shot for Reckhow and Heron for fear of losing them in this high turnout election. The several innercity, white precincts that Gronberg notes as equally supporting Reckhow, Heron and Moffitt are PA strongholds with pockets of older, Friends voters largely accounting for the falloff in votes for Page and Howerton.

* the black candidate endorsed by the Friends--Page--would have had much higher totals in white precincts except that about half the Friends voters honored their heritage of racial discrimination by giving a vote to the only white candidate not endorsed by the Friends--Wright (absolutely not implying that Wright is racist--just that his total benefits somewhat from that dynamic, just as other non-racist white candidates have benefited when the Friends endorse a black candidate in the past)

* Parker scores higher than Moffitt in black precincts (blue line) likely because he recruited young african americans to hand out his flyer with a big bold NCCU on the flyer--walking into the grey area of implying some kind of endorsement by the University.

* Moffit outpolls Parker in white precincts (green line) because of the PA/Indy/Friends endorsements--Parker only had the Friends, some Traction voters and a small boost from African Americans. In a "normal" election year (as 2012 might be) Moffitt's endorsements and hard campaigning (even without the money) would have put him in the winner's circle.

* Reckhow and Heron also polled better than Moffitt and Parker in black precincts (blue line) because of their incumbency--meaning that lots of black voters did not follow the triple shot the Committee recommended and backed some combination of Reckhow, Heron, Howerton and Peterson.

* The concentration of most candidates votes in specific parts of town is a strong indication of the degree to which endorsements matter in Durham. Howerton's ability to draw some votes from all over town indicates effective campaigning that reached beyond the usual boundaries of the endorsements as well as the fact that she was on the right side of the groundswell of angry black women who were ticked off that the Durham Committee did not endorse any women candidates for BOCC.

* In the interest of full disclosure I belong to a high-powered supper club with Don Moffitt composed of some of the local organic food industry elite :-) and I managed Brenda Howerton's campaign.

[Kevin Davis]

Thanks, all, for the comments and useful insights. This is one of the valuable things (to me) about a blog -- garnering the collective received wisdom of folks seeing this from lots of different angles.

To the question of historic turnout -- I'm gathering data and working on that.

Michael: Yeah, I'm using category variables for this. 90%+, 80-90%, 60-80%, and below 60% correspond to the 4 categories.

[KH]

All of this data is great but it almost makes Dur'mites seem stupid. What are Moffit's issues that should resonate with Blacks in the city? I really think it comes down to TRUST more than anything. Blacks don't rubber stamp every candidate of the same color. Obviously, whites don't rubber stamp candidates either.

IMO David Harris and Moffitt struggled in portraying a strong message of what they were about. You can't just ride your community positions to a win where the majority of the people (outside of the inner circles) don't know anything about you.

Even blind support (i.e. Victoria Peterson) will not get you into office. I really only wanted to vote for four commissioners...I almost flipped a coin for the last pick.

Personally, I want people who are about moving this city forward and who can bring solid yet fresh ideas to the table. I like candidates who are able to build consensus around controversial issues creating positive solutions and not more problems.

[Ray Gronberg]

Don't count on 2012 being any more "normal" than this cycle. Obama clearly intends to be on the ballot then too.

That said, I'm not convinced that there was a significant, race-based turnout differential. Late yesterday the Board of Elections updated their report and started aggregating one-stop voters into precincts. I played with those numbers late night after filing my story and it looks like there was an increase of ~15k voters in majority-black precincts over the 2004 primary. But there was an increase of ~33k voters in majority-white precincts.

Now, the majority white/majority black distinction is crude, but it's the best we have available given that the BoE doesn't report voting (as opposed to registration) by race. To go beyond the simple categorization you have to starting making some assumptions.

Census mapping of the precincts should help, but bear in mind that the data is now 8 years old. In a fast-growing community like ours, that's an eternity.

[Tar Heelz]

In 2008, given the influx of new blood into Durham politics over the past twenty years or so, I think some of us who are consumed by local politics tend to vastly overestimate the impact of the endorsements of the local PACs.

Without any evidence to support it, I would guess that the majority of white, Durham voters would have no answer for you if quizzed as to whom the Friends or PA endorsed this election.

The Committee's impact plays out somewhat differently as the Committee's endorsements are targeted by race in a City where race is important. The Committee's importance over the past few cycles has been commented on as in decline. (Truth is all the PACs' importance is in decline.) However, when the country has the first truly viable black presidential candidate in US history on the ballot, the sheer turnout in the African American community was enough to make the Committee's targeted message absolutely critical.

I suspect by the next election, the Committee, Friends, and PA will be right back to their normal status as critically important -- to members of the Committee, Friends, and PA.

[Ray Gronberg]

Also, like Michael I'm not sold on the idea of using Obama support as a proxy for the ethnic breakdown of the vote. The bottom line is he won pretty much everywhere in Durham County by large margins. Now, I did find a negative correlation between his percentage support in a precinct and the percentage of non-black registration (> -.90) but that just indicates the direction of the trend, not its strength. I think it just says very few blacks voted for Hillary, and that some whites did. You'd have to do an ANOVA or something to figure out exactly how strong the relationship is because race is a merely a nominal variable. I didn't bother yesterday because it seemed that voting patterns in local races were much more fractured. I doubt that any winning candidate in the commissioners race was named on as many as half the total number of ballots cast. And clearly no one got majority support in the DA or school board races.

BTW, good luck analyzing the DA's race. That one really is all over the place.

[Michael Bacon]

Kevin,

I think doing categories that narrow is going to be inherently flawed in this case. I'd have to go back over the numbers, but I'll bet you I can find a largely white district somewhere that went 80%+ for Obama. I was thinking a dummy binary variable in the regression, broken down as either +90 or -90, might do the trick. But four categories that close together makes me awfully squeamish.

Ray,

True dat about the 2000 Census numbers, but we have some limited results from the American Community Survey and the Census Bureau's new rolling methodology that's replacing the decennial long form. I was thinking we had finer scale data for Durham already, but unfortunately, it doesn't look like it's out yet -- I'm hoping it comes out with the 2007 set later this year. Until then, you're right, I'm stuck with 2000 numbers. :( But I still say those numbers are better than extrapolating off of Obama's performance.

[chris]

Yo Michael - True dat homey!

[Ray Gronberg]

I agree with Michael about census v. Obama extrapolation. And so this doesn't look like "pile on Kevin" day, I will say that in other places, the Obama vote might indeed be a good stalking horse for ethnicity. But Durham was basically Obama Central in this primary and his vote here clearly crossed racial lines.

[Kevin Davis]

Hi Michael, Ray--

Not a pile on at all. Like I said in the post, the Obama percentage is by no means a perfect proxy for white/black voters; I'd label it a weak one at best. However, there are still two key reasons I think it has some value to look at this this way:

1) First, the results it brings (as noted in the graph above) meshes well with what we're all seeing from other sources and reporting on the election. Page did well across districts; Heron and Reckhow did well everywhere except in a few strong-minority districts; Howerton's steady performance; Bowser's disproportionate strength in Committee strongholds. Of course, a model's explanatory power isn't judged well by it providing results you expect -- but in this case, it resembles other sources well enough to be useful, I suspect.

2) I think the Obama vote % is useful because even though there are majority-white districts in Durham, it's possible that blacks turned out disproportionately to whites within those precincts -- something that as Ray notes we can't learn from the BOE.

I think the Bowser example is instructive: I doubt Joe Bowser, given his history, would earn many votes from white voters, a trend that was clear among the 60+ white voters I polled during early voting.

Here are the precincts that made up the 90%+ Obama support:

BURTON ELEMENTARY SCHOOL
C C SPAULDING SCHOOL
DURHAM COUNTY MAIN LIBRAR
I R HOLMES SR. REC. CTR A
JAMES E SHEPHARD MAGNET S
MONUMENT OF FAITH CHURCH
NCCU - HOLY CROSS
ONE STOP NCCU
VFW POST 2740
WEAVER STREET CENTER
WHITE ROCK BAPTIST CHURCH

And the 80-90% ones below. Interestingly, many of these precincts are not typical Committee strongholds, but Bowser still beat Moffitt across these by about 4,000 votes, which tells me something about intra-precinct turnout:

AMERICAN LEGION POST #7
CHRIST THE KING MORAVIAN
DURHAM SCHOOL OF THE ARTS
EDISON JOHNSON COMMUNITY
EVANGEL ASSEMBLY OF GOD
FIRST UNITED ANTIOCH BAPT
GEORGE WATTS ELEMENTARY S
HOPE VALLEY BAPTIST CHURC
LOWES GROVE BAPTIST CHURCH
MOREHEAD MONTESSORI MAGNE
MT CALVARY LIGHTHOUSE OF
ONE STOP BOE
PEARSONTOWN ELEMENTARY SC
ROGERS-HERR MIDDLE SCHOOL
SOUTHERN HIGH SCHOOL
THE RIVER CHURCH
W I PATTERSON RECREATION
W J BROGDEN MIDDLE SCHOOL
WATERFORD VILLAGE APARTME
Y E SMITH SCHOOL

Not a perfect measure by any means, but I still think we can learn from the data.

[concerneddurhamcitizen]

As an African-American female who lives in Durham, I did not vote for Joe Bowser. I think it is a disgrace that he was re-elected to the BOCC. Durham needs elected officals who will bring cohesiveness to the Durham community, not anger and hostility. Joe Bowser in my opinion will create and has created this kind of energy.

I voted for Brenda Howerton. I voted for her not because she was a Black female but because one of her good friends, who happens to be a friend of mine, was on her campaign committee and said she was a strong leader and would bring people together. This good friend of mine and Brenda's happens to be a White female.

Don Moffitt, in my opinion, did not do a good job of making himself well known. I did not know who he was until I went to early voting and he handed out one of his flyers. After I voted, I began to research him to learn more about where he stood on the issues.

Also, I read about him and other candidates in the Durham News, but unfortunately it was after I voted. After I read where he stood on the issues, I wished I voted for him.

Finally I do not and never did vote based on the flyer the Durham Committee on the Affairs of Black people hand out. I guess I am the exception, not the norm.

Basically, I voted for Page, Heron, Reckhow, Howerton and Parker. I felt the BOCC needed some young blood.

One thing I did learn from this local election is to be better informed about the candidates and where they stand on issues.

[Michael Bacon]

Kevin,

I don't mean to be a badger about this, and maybe I'm trying to turn a simple quick-reaction blog post into too much of a peer-reviewed article, but just looking down your 80-90 list, there's some of the whitest precincts in whitedom down there. Just having George Watts Elem. and DSA in there doesn't show something about ethnic turnout, in my opinion, but rather something about Durham County for Obama leader Faulkner Fox's persistence in her own neighborhood. I also find it hard to believe that Hope Valley Baptist could be construed a heavily black precinct.

Obviously, there's going to be a very, very noisy but still significant relationship between percentage black voters and Obama support, so I guess some use of that as an indirect correlation of racial composition isn't THAT terrible for aggregated vote totals. I guess the mushiness of it is still setting off alarms in my head, but that's probably from reading too many multiple regression papers in grad school.

[Ray Gronberg]

Here's something to chuckle over, from Michael Barone of U.S. News & World Report:

"Obama's relative success in the May 6 primaries can, I think, be chalked up to his performance in three metropolitan areas which have had robust or relatively robust economic and demographic growth and where urban politics has not been focused on racially polarized city elections—Charlotte and Raleigh-Durham, N.C., and Indianapolis. I haven't calculated the metro area totals yet, but it's apparent from these data that Obama ran very well among whites in their largest counties."

Love that. Durham "has not been focused on racially polarized city elections." Well, we have been talking about a county election, but still, you gotta love the national press corps.

That said, Barone's right in saying Obama ran well amongst whites here.

The full piece is at http://www.usnews.com/blogs/barone/2008/5/8/clinton-and-obamas-super-tuesday-in-indiana-and-north-carolina.html#read_more

[slowbyrne]

Wow - After reading these posts, this is so much inside baseball, I feel like a rubber core wrapped in 100 yards of string.
Just a thought - maybe some of these BOCC candidates would break out of the race rip-tide if they clearly communicated positions on issues effecting residents RIGHT NOW.

EXAMPLE: SCHOOLS AND WATER

1) Why should existing residents pay for NEW schools when the NEW schools are needed because of NEW sub-divisions??? Impact fees, anybody? They're the law of the land for the last 30 years in California!

2) Why are existing residents forced to curb water use as the municipal resevoirs are literally overflowing AND new sub-divisions are approved AND allowed to plant more and more sod. At this point, DON'T EVEN CALL IT A DROUGHT. The water shortage is not about water, it's about a lack of water storage and water capacity, which goes back to municipal leaders not planning and paying for growth.

Sure, BOCC may not directly control these issues. But they have influence to steer the debate. They can deny annexations. They can deny a school bond, and hammer the General Assembly to allow school impact fees. But they are doing neither.

As long as our local leaders are not responding, I will not follow the water rules and I will not vote for school bonds. THESE are local issues. God help Durham if our local PAC's are endorsing people basically because of skin-color. That sounds like the MO of at least one of them.