By the Numbers: Look at Alabama!
And what would it take to replicate that victory in CD4?
Just for a moment, as an exercise, let’s call our congressional district, CD4, “the Alabama of Colorado.”
The parallels are imperfect. Alabama is 70% white. CD4 is 92% white, but the fact that Latino/Hispanic persons are often not accorded “white privilege” regardless of racial heritage complicates predictions based on race. Roughly 22% of CD4 residents self-identify as Latino or Hispanic. A future article in this series will examine how race and ethnicity can be used to predict and affect voter turnout here in Colorado.
The four largest cities in Alabama account for about 50% of its population; in Colorado, that’s 42%. The median household income in Alabama is about $47,000 per year; CD4 is much richer, with a median family income closing in on $68,000. And we are more Libertarian and our voters are less consumed by, um, relitigating the Civil War. The metaphor is not perfect.
But when you reduce the world’s complexity to just two parties, you discover that CD4 votes quite similarly to Alabama. Trump won Alabama by 28 points in 2016; Trump won CD4 by 22. The same year, Richard Shelby won his Senate seat by a 28 point margin, while Ken Buck won here by 32. The physical mountains in Colorado are larger, but the metaphorical mountain that blocks a Democrat from taking office looks just as insurmountable.
So, if we want to break Ken Buck’s grip on CD4, the most logical election to analyze is this very odd, very southern show. To understand CD4, we start in Alabama. And because we are going to go a bit into the weeds here, I will start by telling you what you’re about to read.
The usual advice to winning an election is to turn out the base, and that makes sense when you have a close election like our recent presidential one. And turnout was a huge factor in Alabama, though on its own not enough to account for Jones’s victory. To win in such a conservative district, we also need to fundamentally restructure the conversation, and convince both independents and a few, brave conservatives to jump ship. This can happen — Alabama is proof. It’s up to us to learn from their successes.
Alabama just swung 30 points. Let’s strip away the narrative about Commandments and disbarment and ephebophilia (seriously, that’s a word). When there is a 30 point flip, there should be a pretty strong signature of change. What, exactly, happened with all those votes?
Roy Moore was a miraculously bad candidate. Moore’s political performance was a brush fire that made an evangelical state question whether God Himself was commanding them to vote Democratic. There are, thankfully, few other candidates like Moore, and even the perennial foot-in-mouth Ken Buck is unlikely to invoke such a Biblical pestilence upon the land.
So Democrats were highly engaged, much more so than in 2016. Independents started to look around for other options. And even some conservatives rejected the taint of sin that surrounded Moore and his campaign.
What were the main factors in flipping Alabama to a Democratic senator?
Hypothesis #1: Turnout won it
In 2016, Richard Shelby walked away with the election as Alabama’s senator, winning 64–36% and netting 1.33 million votes. In 2017, Doug Jones won in a tight race for Senate, netting 51%, and a total of 674,000 votes. Richard Shelby’s vote total was enough to defeat Doug Jones and Roy Moore combined.
Clearly turnout — the composition of who showed up for an off-year election — must have made a huge difference. Perhaps Democrats came to the polls, but Republicans didn’t? At first glance, the totals below look like lots of Republicans just stayed home.
Republicans generally have an edge in mid-term elections — their base is older, and older people are, for a variety of reasons, more likely to avail themselves of Democracy. This alone gives GOP candidates a roughly 3 point advantage nationally in a normal mid-term election. In fact, in the last 40 years Democrats have never held a mid-term turnout advantage.
Unfortunately, Alabama does not register voters by party, so we cannot definitively compare turnout against voter registration rolls to see if Democrats exceeded expectations. But we can use other measures: Outside polling from 2014 showed a 49–35% Republican lead in the state, yet in this special election Republicans had only a 43–37% advantage. Something changed.
This eight point shift might reflect a turnout advantage for Democrats. Yet it Is also neatly in line with a nationwide 5 point drop in Republican identification in the era of Trump. It could be that, never having had to declare a party affiliation publicly, some Republicans have recast themselves as independents — we saw 20% of voters identify as independents in this election, up from 16% in past polls. Did people switch their stated party preference in reaction to Trump, or did Jones we attract a different set of people to the polls in this highly contested special election?
One way to separate the groups is to look at ideology, which should remain constant even if a voter shifts from GOP to independent. According to exit polls, special election voters were 6 points more liberal than 2016, and 3 points less likely to be white Evangelicals. The fraction of voters identifying as African-American was up 4 points (29%, compared with 25% in a ‘normal’ mid-term), which means that whites fell by 4. Less reliable Democrats were replaced with more reliable ones, and some conservatives disappeared altogether. Overall, it does look like turnout alone could have accounted for 8–10 points towards Jones’s victory.
This result is entirely historically unprecedented. Getting out the vote yielded real, tangible results for Jones, far in excess of even the Tea Party advantages of 2010.
Republicans should be terrified.
Conclusion: Turnout boosted Democrats from -3 underdogs in a generic midterm, to +10 in this special election. A generic shift from GOP identification to independent identification (or independent to Democratic) could have contributed 2–4 points more.
Hypothesis #2: Republicans crossed over to vote for Jones
It is entirely possible that Republicans fled in droves from the charlatanism of Roy Moore.
It’s also possible that Republicans chose not to believe these slanders, choosing instead to stand by their man.
In fact, while 90% of all voters stuck with their self-professed party, self-identified Republicans were 4X more likely to cross over than Democrats. And crossover matters — the GOP’s nominal 43–37 party ID lead vanishes here. Taking into account this minor mutiny, our expected outcome slides to a 40–40 tie.
The focus in normal elections is to turn out the base, and to worry less about persuading the opposition. But with a candidate like Moore on the table, there were enough unhappy Republicans in play to boost the Democrat, contributing almost as much as turnout provided.
Conclusion: Turncoat Republicans gave the Democrat 6 points in Alabama.
Hypothesis 3: It was the independents who done it
Over 20% of Alabamans self-declare as ‘Independent’ or ‘Other’, and Jones took them 51–43. This advantage put Jones over the top.
How much of a shift was this from a normal year?
Alabama does not keep voter registration statistics, and because elections in Alabama are rarely competitive, there are few exit polls for comparison. But we can do a little bit of educated guessing.
In 2016, let’s assume Republicans had a 49–35 party identification advantage (as per polling from 2014). Shelby won 64–36, which means that independents provided the necessary margin, 15–1 in favor of the GOP. To make these numbers work, independents must have gone over 90% for the Republican candidate.
In reality, some Democrats probably bolted to Shelby as well. So let’s look at the problem another way: If we assume that the demographic profile of Alabama independent voters is much like the national norm, then the independent electorate is expected to be 77% white, 20% African American, and 3% from other races. Assuming African Americans voted 90% Democratic in 2016 and white Alabamans voted 85% Republican, we’d get a roughly 70–30% GOP lean for Alabama independents.
Something big clearly changed for this voting bloc. A 70–30 GOP lean turned 51–43 in favor of the Democrats. If Moore had won independents at the expected rate, they would have provided an 8 point boost to his campaign, instead of a 2 point loss.
Yes, Virginia, there really are swing voters. And this year, in sweet Alabama, they swung Democratic.
Conclusion: Independents swung +10 towards Jones, sealing his victory.
What does this mean?
In the next installment, we will apply these lessons to voters in Colorado’s CD4. For now, we can offer the following summary of what mattered most to Alabama in this stunning Democratic win.
- +10%: Turnout was key. The base really did show, and the opposition did not. Yet we must keep in mind that overall turnout for this special election may not reflect what we will see in a regular, November mid-term.
- +2%: The nationwide loss of support by the GOP meant a lower number of voters self-identified as Republican. As independents, their odds of voting for Democrats rise substantially. This factor will influence races across Colorado and the country.
- +6%: A small fraction of self-identified Republicans broke away from Moore. It may seem like our political discourse takes place in entirely non-overlapping spheres, but that’s not entirely true. It is possible for proud Republicans to get fed up.
- +10%: The swing of independent voters from R to D was a huge factor in Jones’s victory. Independents actually lived up to their name, and the election swung on their actions.
- +2%: Some conservative votes went to a write-in candidate instead of Moore/
Alabama witnessed a stunning +30 point swing in party voting in the span of one year. In the world of Trump, things that were previously considered impossible are now merely improbable. And things that were previously improbable now casually stroll through the door.
In this singular, very special election, we have learned that even the very conservative people of Alabama can vote for sanity. The center can hold.
In the next installment of this series, we’ll talk about how hard the rest of us will have to work to build on the lessons of Doug Jones and Alabama. And what the numbers might look like in Colorado.
