The day of reckoning us upon us, as Alabama votes in its Senate special election between Democrat Doug Jones and Republican Roy Moore. Even the experts have literally no idea what the result will be, with the most recent round of polls showing results anywhere between a 10-point Jones lead and a 9-point Moore advantage. Several pollsters are even offering choose-your-own-adventure results where different turnout models lead to different candidates winning. But despite the huge spread, an average of all the polls shows a very tight race, with Moore ahead by less than 2 points.
To help you with your results-watching tonight, Daily Kos Elections is offering our usual election night tool: benchmarks for the most important counties, which will give target numbers that Jones or Moore will have to meet in order to eke out a win statewide. The way that this works is that we use the 2016 presidential results, both statewide and in each county, as a baseline, then adjust upward or downward in each county by the same factor in a manner that would point to a narrow Jones win. (You can find a county map and more data on each county on Wikipedia.)
In this case, we're modeling toward a 49-47 Jones win. (Keep in mind that I’m not necessarily predicting a 49-47 Jones win—though that does seem like a potentially credible outcome—just using that as a bare minimum victory threshold.) That reflects the fact that in 2016, 4 percent of Alabama’s votes went to third parties and the potential that a similar number of people will choose to cast write-in votes this year. Last year, Donald Trump beat Hillary Clinton by a 62-34 margin, so getting to 49-47 win for Jones means adding 15 points to each county’s 2016 results for Clinton and subtracting 15 from each county’s 2016 results for Trump.
The first column in this table gives the name of each of the state’s counties, and the second column shows what percentage of the total number of votes statewide came from each county. (Note that, as usual, we’re restricting this to the counties that cast 2 percent or more of the total statewide votes. These 12 metropolitan counties reflect not quite two-thirds of the state’s total number of votes.) The third column shows each county’s 2016 breakdown, and the fourth column shows the adjusted number that, for instance, Jones would need to hit in each county en route to 49 percent statewide.
||% of 2016 vote
||2016 results (d/R)
||what jones needs
I’m also trying something new here with my fifth column, which describes each county’s level of diversity (based on the percentage of non-Hispanic whites) and education (based on percentage of college-educated persons over age 25), broken into quintiles (with the quintiles reflecting national averages, not averages for Alabama on its own). If this formulation seems vaguely familiar, recall that we developed it in a recent piece demonstrating that 2016 results showed that race and education were both highly predictive, more so than ever, of which way counties broke that year.
And this detail may be especially important in this unusual election, in that we may see the most highly educated parts of the state break disproportionately toward Jones. We’ve been hearing, for instance, lots of anecdotal reports of unusual levels of support for Jones in the usually country-club Republican areas in the affluent suburbs south of Birmingham, some of which are in Jefferson County but many of which are in Shelby County.
And that may extend to Huntsville’s Madison County, where a large NASA facility has created a critical mass of well-educated voters. We could, in fact, see a map where Jones, say, hits 60 percent in Madison County and ekes out a win in Shelby County and thereby scraping out a statewide win while still falling short of his benchmarks in similarly white, but much less-educated, small-town places like Calhoun County.
Finally, as with our Virginia benchmarks from last month, I want to add a couple of additional counties from elsewhere in the state that, individually, contribute less than 2 percent of the state’s total votes but are reflective of large rural swaths of the state that, collectively, could tip the race one way or the other even if the candidates hit their benchmarks in the larger counties. This, you may recall, was the big functional problem with our benchmarks in 2016: Clinton mostly hit her benchmarks in the large counties, but lost the race via death of a thousand cuts in the rural counties too small to make the list.
||% OF 2016 vote
||2016 RESULTS (d/R)
||WHAT JONES NEEDS
||Very Low/Very Low
||Very High/Very Low
If you’ve ever looked at a red-and-blue map of Alabama, you may be familiar with the dark blue band that slices through the state’s middle south of Birmingham. This is known as the Black Belt, a swath of counties with an African-American majority, though the name reputedly originated with the dark color of the region’s soil. (As an aside, a fascinating argument from several years ago claimed that the ancient shoreline from tens of millions of years ago shaped the region’s political geography, in that the unique soil formed in the Black Belt was especially suited to the cultivation of cotton, the development of a slave-based plantation economy, and thus formation of a largely black population.)
Dallas County is where Selma—possibly ground zero for the civil rights movement—is located; it’s the state’s second-largest county with a black majority (after Montgomery County, site of the state capitol). It’s worth keeping an eye on the Black Belt counties, because we’ve also been seeing lots of anecdotal reports on how the African-American community isn’t tuned in to the race. So an unusually poor performance for Jones in Dallas County could be indicative of broader problems among black voters statewide … or, conversely, an especially strong performance here, at 2008 or 2012 presidential levels, could be a clue that Jones is on track to win.
Less known than Alabama’s Black Belt is the state’s White Belt. Okay, it doesn’t really exist: I’m coining the term myself to reflect the way that there’s a broad swath of exurban and rural counties in the area north of Birmingham but south of Huntsville that—contrary to what you might expect about Alabama—are almost entirely white. Cullman County is a prime example: It’s 93 percent white, with most of the remainder Latino rather than black. Also consider that these counties are some of the most heavily evangelical parts of the entire country: Cullman County, for example, is around 68 percent evangelical. These counties form the core of Alabama’s mostly-rural 4th Congressional District, perennially one of the reddest districts in the country, and Cullman is the cornerstone of that, with Clinton getting a whopping 10 percent of the vote here in 2016.
So it’s worth keeping an eye on whether Moore can retain most of his evangelical supporters and overperform his benchmarks in Cullman and counties like it; if he’s pulling down 80+ percent of the vote in Cullman and its neighbors like St. Clair and Walker Counties, he might be able to eke out a win even if Jones hits his benchmarks elsewhere. On the other hand, if there’s substantial Republican abandonment of Moore, not just in the well-off suburbs but also in evangelical strongholds, Jones is likely on track for a good night.
Finally, there’s one other county that’s demographically similar to Cullman County and is also found in the 4th District: Etowah County, where the city Gadsden is located. (Etowah is on the upper list, since it’s responsible for 2.1 percent of the state’s votes.) If Gadsden sounds familiar, it’s because you’re thinking of its mall: Moore is from Gadsden. Now, under normal circumstances, you’d think, “Well, he’s the hometown hero; he should overperform there, and Jones will have to make up votes elsewhere.”
But in 2012, when Moore won a statewide election for the state Supreme Court by a shockingly narrow 52-48 margin over a Democrat, Etowah County gave Moore his second-largest underperformance of any county in the state compared to Mitt Romney’s performance in that same election (following only Shelby County, home to so many country club Republicans). That’s pretty remarkable that the people who ostensibly know Moore the best are the ones most disproportionately likely to say “No, thanks.”