The Landslide that Fell Uphill: The Decline and Fall of Artur Davis

Everyone's had time to hit the beach, or the lake, or the saloon, since the Democratic primary on June 1. The attention of the chattering classes at the Newhouse Division of the Republican Party has been preoccupied with the James-Bentley struggle for a runoff spot. I suppose that "breaking news" sells more newspapers than thoughtful contemplation, even if not as many as the most recent rumor concerning the college preference of a cornerback at Coffeeville High (for the unlearned, with an average daily attendance of 50, the smallest high school in Alabama that plays football). That gives us the opportunity to look at the landslide that fell uphill - the Artur Davis landslide win that turned into the Ron Sparks landslide win.

So, what do we make of the real story of Primary Day? That story, of course, is the wholesale repudiation of the most serious black gubernatorial candidate in Alabama history, Artur Davis, by a substantial portion, if not a majority, of the black vote. The answers to all the questions that issue raises will tell us a lot about how the black vote will act in future Democratic primaries, and may give us some lines of inquiry into how, someday, this state might actually have a black governor. A lot of factors come into play in this analysis, including Davis's own strategy, the role of historical black political organizations - mainly the ADC, New South Coalition, and the ("New") JCCC, and the selection of basic metrics. There weren't any significant exit polls on the ground on Primary Day - at least not that have surfaced - so we will have to look largely at the available returns.

But before prescription, comes diagnosis. Before diagnosis, come the symptoms. In that regard, first, let's look at the gubernatorial results in the seven counties with the largest percentage of black population:

County	Davis	Sparks	Sparks%	Black%
Macon	2,028	2,311	53.3%	83.9%
Greene	1,196	2,556	68.1%	79.0%
Sumter	2,424	2,282	48.5%	72.9%
Wilcox	1,132	3,489	75.5%	72.5%
Bullock	1,464	2,159	59.6%	70.5%
Lowndes	1,373	2,510	64.6%	70.2%
Perry	855	2,202	72.0%	68.5%

Those figures create a strong anecdotal suspicion that Davis badly lost the black vote. There is simply no way for Sparks to get 68% in Greene County without a majority of the black vote. On a somewhat more abstract level, consider the correlation of the black percentage of population, with the percentage of the vote received by Sparks, in all 67 counties, which would include the larger urban areas not in the above table:



This chart tells us a few things, mathematically. First, the overall correlation between black percentage of population, and the percentage of the overall vote Sparks got, shows a negative correlation, confirming anecdotal belief that Davis did better among black than white voters. The correlation, however, is somewhat weak; confirming that Sparks made significant inroads into the black vote, as exemplified in the counties in the table above. The strongest actual correlation is shown in the upper left corner, but that represents the whitest counties. No news there; Davis's white vote was negligible, even among a moderate Democratic primary electorate. (Imagine, if you dare, how bad it would have been among the much less enlightened general election cohort.) This trend toward randomness increases as you move rightward on the chart; that is, as the electorate represented by each data point becomes more black. This is consistent with the known data that Davis carried Mobile and Montgomery Counties, where the Democratic primary electorate is overwhelmingly black, and came closer to Sparks in Jefferson County.

Any number of explanations can be put forward for these numbers, but the one that best satisfies Occam's Razor is that since the variations can be seen so readily at the county level, local variables are significant in the determination of the result. Or, as Tip O'Neill would have said, "All politics is local."

But that observation still doesn't tell us what the local variations were, or what caused them. A defensible conclusion awaits more research, but I will give it a stab anyhow. The two least-depressing spots on what was a dismal night for Davis, were easily Montgomery and Mobile Counties. Both counties have factors that likely accounted for Davis's stronger showings there. In Montgomery, Davis is a native son, and his wife and mother are also locals. This means there are going to be people who grew up with them, went through school with them, and are related to them. Even beyond that circle, he would be known as "local." While these factors aren't going to be enough alone to carry an urban county like Montgomery, they will make a difference in a tight race, and Davis only carried Montgomery by 50.5%-49.5%.

In Mobile County, Davis was the beneficiary of a pre-implosion endorsement by popular - and black - Mayor Sam Jones. Jones never recanted the endorsement, and word on the street is, he considered himself obligated to follow through on his pledge of support. Whatever happened in Mobile County, it was Sparks's weakest county in the state; he only took 39.9% of the vote there. Since Mobile County is only 35.7% black, other factors may have been at play. Dauphin Island is as far as you can get from Sparks's home in Fort Payne and still be in Alabama, and Mobilians have an instinctive dislike for anyone from places in North Alabama, like Evergreen and Wetumpka. (In a semi-related aside, after this display of political muscle, is Mayor Jones now the de facto leading spokesman among black elected officials in Alabama, as former Birmingham Mayor Richard Arrington once claimed to be?)

As much as Jones might like to take credit for "delivering" Mobile County to Davis, and as much more as the leaders of the traditional black groups want to take credit for delivering, well, everyplace else to Sparks, the reality is probably a bit more complex. Davis's position until the night of the Primary was that leaders such as Dr. Joe Reed and Senator Hank Sanders couldn't "deliver" the black vote anymore. There may have been a kernel of truth to that, but it's equally likely their support of Sparks was indirectly necessary to his win. For that, I am going to humbly offer a working hypothesis for what happened with the black vote in this contest.

Any black voter in Alabama was looking at several competing claims on his or her vote in this election. Certainly, there was a "pull" by virtue of Davis's race. Davis seems to have presumed this pull would overwhelm all other factors. Related to this would have been the Obama-esque prospect of creating a "first." However, if Obama the concept was Davis's friend, Obama the actuality was his undoing. Specifically, Obama's act of tying himself to the mast of health care reform, and Davis's willingness to pander to the Business Council and Tea Party crowds to the point of twice voting against it, played an immeasurable part in Davis's undoing. (To be sure, Davis cast other votes in an effort to pander to the center-right: against the GM/Chrysler rescue, for Patriot Act renewal - but none of these were as visible as the health care vote. Sparks showed his sharp political instincts by taking full advantage of the mistake.)

Did the health care votes irretrievably doom Davis? Not necessarily. It is important to remember that voting behavior is influenced by a multitude of factors, in ways that are far from totally understood. Certainly, candidates' stands on high-visibility issues play a well-established role in predicting voter behavior. But issues can easily be trumped. Obama ran, even in the 2008 primaries, as something of a moderate among Democrats. (Remember those warm, fuzzy Obama statements about Ronald Reagan?) Other candidates, including, arguably, Hillary Clinton, were probably much closer to the ideological preferences of black voters than Obama. The South Carolina primary showed how overwhelmingly racial identification, and the "first" impulse, overwhelmed those concerns among black voters.

So what was different in Alabama on June 1? First, probably, Davis's ideological break to the right was more marked than Obama's, and as it could be encapsulated in the health care reform votes, more visible. Davis didn't help himself by beginning his fall campaign early, and highlighting his "moderate" positions. Perhaps that is explanation enough, but it seems more had to happen than that. And that is where the historical black organizations came in. Their full-throated condemnation of Davis's votes, and support of Sparks, gave a black voter internal intellectual cover for voting against a prospective historic "first." If you like, they told him or her "it's OK to vote against Artur, you won't be the only one on your block." The buzz they provided along those lines in churches, barbershops, and elsewhere in the black community, allowed black voters to express their issue-based anger with Davis without feeling a sense of group betrayal. This model comports with the empirical data of the returns. Where this "buzz" was not present, or attenuated, - as in Mobile - Davis did well. Where it was present, Sparks handily won the black vote. In this vein, Davis didn't help himself with his very public snub of these organizations' screening processes. Even black voters who don't take their cue from these organizations are still mindful of their pioneering roles in securing basic civil rights and political influence in Alabama. In sum, these organizations - while their street operations do still deliver some votes, if at considerable cost to the candidate - probably could not have provided Sparks with a majority of the black vote against John Lewis. But while Congressman Lewis could, and did, endorse Davis, he couldn't save Davis from himself.

In addition to the historical black political organizations, one factor in Davis's loss of black support bears mention. I am referring to the numerous black members of the Legislature who supported Sparks. Some did so visibly, like Sen. Bobby Singleton of Greensboro, and, notably, Sen. Hank Sanders of Selma, whose open letter to Artur Davis telegraphed his eventual support for Sparks. Other black legislators gave their support quietly, either by off-the-record conversations with community leaders, or by putting the word out to their substantial personal GOTV personnel on the day of the election. This support was understandable, and predictable, on a purely pragmatic basis. As these seasoned politicians understood, a Davis nomination would have likely resulted in significant losses for Democrats in downballot races throughout the predominantly white areas of the state. This could easily have left the Legislature in Republican hands, and left for naught the decades of struggle to put legislative power in the hands of black leaders such as House Finance & Taxation Chairman John Knight, House Speaker Pro Tem Demetrius Newton, Senate Finance & Taxation Chairman Hank Sanders, and Senate President Pro Tem Rodger Smitherman. (For clarity's sake, I am not implying that any of these leaders supported Sparks, beyond their own public statements; I merely note their leadership posts would likely have vanished after a Davis nomination.) While this phenomenon was foreseeable months ago, and became more evident as the primary approached, it was almost totally unreported in the media, especially in the Newhouse outlets. Of course, in their case, it would not have fit their "inevitable Artur nomination" story line.

It was the prospect of this position by numerous black legislators that led me to predict, some time ago, that the black vote for Davis would not be monolithic. That it broke as heavily as it did for Sparks, I admit, caught even me by surprise (and perhaps Sparks, though I know better than to ask). I was figuring Sparks would win, but I had it figured at about 52-48. 55-45 if every single variable broke his way. I never bought the media-poll line that Davis was getting 40% of the white vote, and a near-unanimous black vote. Of course, I never had their Davis-nomination agenda, either. Sometimes, it's fun to be wrong.