By David de la Fuente
Normally Third Way is all about persuasion and swing voters. But the Alabama race will come down to turnout, and specifically the behavior of three types of normally rock solid Republican voters. I call them the Trumpers, Shelbys, and Flakes.
Let’s start with the Trumpers. The president may have supported Moore’s opponent in the primary, but he’s all in for Moore in the general. Voters like these would support Moore even if he was caught cruising the Riverchase Galleria Mall in Hoover. There are a lot of Trumpers in Alabama, like most in-state Republican leaders, including third-place primary finisher Rep. Mo Brooks who endorsed Moore. Next are the Shelbys, named for Alabama Republican Sen. Richard Shelby who said he voted for a Republican write-in. The Shelbys are Republicans who will either stay home or cast a protest vote against Moore, but not for Jones. There will be more everyday Republican Shelbys in this camp than elected officials as party officials tend to circle the wagons around a flawed candidate. Data indicates that the race is as close as it is because of the Shelbys. And last are the Flakes, an homage to Republican Arizona Sen. Jeff Flake who flipped and actually wrote a check to Doug Jones. I had to cite an out-of-state party official for this category, because, as shown in the data below, there doesn’t seem to be many Alabama Republican Flakes — those willing to proactively show up for an election just to vote for one Democrat.
So the big questions will be how these groups of Republicans voters will affect turnout, and will that create and opening for Democrat Doug Jones on Tuesday to pull off the upset.
Moore’s Bleed
The analysts at FiveThirtyEight have put together a helpful comparison of polls conducted by the same pollsters before and after the allegations against Moore. They found that the margin between the two candidates shifted nine points in Jones’ favor during this time period. That demonstrates real movement away from Moore and hints at a causal relationship, as nine points is outside the margin of error.
One recent poll points to Republican voters potentially staying home as a key driver for Moore’s erosion of support. In a Fox News poll that showed Jones leading Moore by a 50% to 42% margin, it also found that Barack Obama had a favorability rating of 52% — in Alabama. A 2015 Gallup poll pegged Obama’s approval rating in Alabama at 35%. That latter number is more in line with how Democrats have performed in the state in recent elections; gubernatorial nominee Parker Griffith got 36% of the vote in 2014 and Hillary Clinton got 34% in 2016. This doesn’t mean the Fox News poll is bad, but rather that since the poll only included likely voters, a lot of Republicans told the pollster that they will stay home.
However, there is precedent for pollsters being wrong in this type of situation. In a 2013 special election for U.S. House in South Carolina, the Republican nominee was scandal-plagued Mark Sanford and his opponent was an uncontroversial Democrat. Polling heading into the election favored the Democrat, but in the end, Sanford won by nine points. The most logical explanation for this would be social desirability bias, meaning that people are not willing to tell a pollster they’re going to vote for someone that might bring them shame. Nevertheless, they were still willing to turn out on a day when that election was the only thing on the ballot and privately cast their vote for Sanford. And just like Sanford, Moore has appeared to have some bounce back in the polls a week from the election, which points to many Republican voters coming around to the Trump/Brooks way of thinking.
Turnout is King
That means the big question will be turnout. Looking at past elections can provide a baseline for what turnout could be expected in this election.
Alabama had 64% turnout among registered voters in the 2016 general election and 40% turnout in 2014. The 2016 primary turnout was 42%, and the 2014 primary turnout was 21%.
The special primary and runoff in 2017 had a combined turnout of 19%, but the raw numbers were larger than the 2014 primary turnout thanks to increased voter registration numbers. If the upcoming special election is consistent with the primary-to-general election turnout arc detailed above, turnout could be in the 40% range.
But that should be taken as a ceiling. Alabama’s Secretary of State is predicting turnout as low as 20% in the December special election, and that changes the calculus of how many voters Moore can lose before his numbers sink below Jones’ ceiling of possible support.
There is nothing else on the ballot on December 12th, so this election is very different from another recent election that has drawn comparisons by commentators. That’s the 2012 Senate race between Claire McCaskill and Todd Akin, which was upended by Akin’s infamous comments about ‘legitimate rape’. In that election, turnout was primarily driven by the presidential race, and McCaskill ended up winning around 260K more votes than Obama did, which indicates that many people voted for Mitt Romney and then McCaskill. Without another race on the ballot, these types of voters will probably just not show up, so Jones will not likely get an assist from these kinds of Jeff Flake ticket-splitters.
Breaking Down the Math
Using numbers from progressive data firm Catalist, we created a potential universe of voters and then narrowed it down to find Moore’s loss number — the concept that Moore could get so few people to turnout that he’ll be below what Jones can realistically get. This should not be conflated with a likely universe — one that we are saying will probably vote. Rather, this is a list of the most reliable voters, and are thus the most likely to possibly show up for a December special.
We looked at Alabama voters who are registered and voted in at least two of the last three big elections: the 2016 general, 2016 primary, and 2014 general (data is not currently available on people who voted in the 2017 primary). That created a potential pool of 1.47 million (M) voters, which would be a turnout of 45%. Turnout will almost certainly not be that high, but this is a good way to narrow down the 3.24M Alabamian registered voters in the system. Also, it should be noted that this group of voters will lean more towards the Republicans, because 2014 was a good midterm year for that party, and the presidential primary was more heavily contested on the Republican side in Alabama.
Of these 1.47M voters, applying a partisan model reveals that 374K are listed as likely Democrats, 769K are likely Republicans, and 303K are likely Independents. Alabama is a heavily partisan state thanks to its large African-American population and the fact that Whites in the Deep South are starting to vote Republican by wide margins, even as big as 80-to-20, so the number of Independents is lower than in many places across the country.
Layering on data from the Cooperative Congressional Election Study about the 2016 election helps to show that Catalist’s likely Democrats and likely Republicans will almost certainly stay with their party in this election. In Alabama, 97% of self-identified Democrats supported Clinton and 97% of self-identified Republicans supported Trump, which shows very little propensity for crossover or ticket splitting. Likewise, Catalist data showed very low numbers of potential ticket-splitters amongst those coded as likely Democrats and likely Republicans. Also, Alabama Independents supported Trump by a 69%-to-19% margin — a breakdown which is consistent with the partisan lean of the Independents in the Catalist data (which are coded as more likely to be Democratic or Republican leaning).
If the special election were to hit midterm turnout levels, Jones probably has a ceiling of around 475K votes when combining Democrats and Independents who might vote Democratic. Therefore, if Moore can get that many Republicans or Republican-leaners to show up, he should be able to get across the finish line. Anything below that, and he is in danger of hitting his ‘loss number’ unless Jones suffers from below-average Democratic turnout. Hitting that loss number would require about half of the reliable Republican voters who could typically be expected to turn out in a midterm or big Republican primary to not show up on December 12th, but Moore might just be abhorrent enough for that kind of result to materialize. Democrats can win if they have gangbuster turnout and Republicans suffer from apathy.
There have been some reports that turnout with African Americans (any Democrat’s base in Alabama) could be low. Unsurprisingly in a state where partisanship tracks so closely along racial lines, Jones will need African American turnout to be higher than White turnout comparative to their total population numbers to put himself in a position to take advantage of Moore’s shortcomings. If African Americans make up a share of the electorate that crosses into the mid-30s, Jones could well be in for a good night. To put that in perspective, African-Americans made up 24% of the electorate in both the 2014 and 2016 general elections. This late in the game, campaign resources would likely be best spent convincing that community to come out in midterm-like numbers for this special election, rather than going after White persuadable voters. Regardless of the outcome, this type of work would strengthen the Democratic political infrastructure in the state — a goal many leaders in the state, like State House Democratic Leader Anthony Daniels, have been working on long before this election.
Conclusion
Generally, we tout the necessity of persuasion. But in this election, it’s all about turnout.
A Jones victory is predicated on the hope that Republican voters will be Shelby or even Flake Republicans and not Trump/Brooks Republicans. This is a rare case in which low turnout could be beneficial to the Democratic Party, and Jones’ best hope for victory is that Moore hits his loss number.
