And So It Begins…
It was an odd way to falling in love with politics. It was 1998 and I was a junior in high school with most of my time spent being a jock and chasing girls. I had also recently joined my first nerd club, the high school debate team. Everyone told me that my true love was argumentation and I really didn’t have the physique to play sports much longer — all 5'10" and 110 pounds of me. I was becoming enthralled with the Clinton Impeachment trial. But not just with the politics of it, and the fallout it created from a policy standpoint, but it was my first introduction to approval ratings, public opinion polls, exit polls, and how events like this can affect the upcoming election cycle.
The 2000 Election, my freshman year in college, was the first Presidential Election I followed. In my home state of Minnesota, as an 18 year old, the first major race I was ever exposed to was my generation coming out en masse and electing a professional wrestler to be governor (Full Disclosure: I voted for Norm Coleman that year, Jesse wasn’t my fault). But then two years later, I experienced the most dramatic election we’ve ever had for the White House. I pulled an all-nighter and skipped classes the next day as I watched every pundit, every exit poll, and the insanely thin margins of how an election can be decided.
Fast forward 20 years later, and there is no hobby I spend more time on than evaluating and (trying to) analyze election numbers and demographic stats. I am still building my model for the 2020 election cycle, but we have seen a shift in so many areas of the American electorate in the last 20 years, that I think it’s getting more difficult to accurate predict electoral outcomes with the traditional binary models that have existed.
So, I will be diving into some deep demographics in future articles. However, for right now, we need an initial jumping off point of where we stand as we go into what will be the most unique and odd general election of our lives. The potential for no rallies, traditional fundraising events, and other community events will make even down ballot races interesting.
To gather a baseline to start from this election, I am basing my initial ranking on current polling, plus the basic understanding of demographic trends in each individual state over the last decade. Not all polls are created equal and numerous high level poll aggregators give polling firms and universities grades based on their accuracy in election data.
Polls typically same four different types of groups: All Adults (A), All Voters (V), Registered Voters (RV), and Likely Voters (LV). While they all have value for different reasons; in general, the accuracy of these polling models ranges from least accurate to most accurate. Just polling adults gives us no indication if they even know what a voting booth is and just polling registered voters could mean you’re sample someone that hasn’t voted in decades. This is why most election data analysts put more weight toward polls that same a high number of likely voters.
You’ll notice a few stylistic differences from my map than most other predictors you’ll see. I do not straddle the fence with toss up states. This often means that I might flip a state more often than other professional psephologists.
As of right now, here are half of the states that I see being the most competitive during this season and what will likely sway them demographically. There are so many external events — from COVID, to unemployment, to major third party bids — that could play factor which are unpredictable right now. I am only speaking to how demographics groups are acting right now:
Michigan (Biden, +6%)
Along with Pennsylvania, this was the biggest surprise from the 2016 election. Internal leaks from the Trump campaign report they are already about to write this state off, likely due to numerous uphill battles. Polling has been very static for Biden, with Trump only winning one nonpartisan poll head-to-head against him in 2020.
Trump performed five points higher among independents than he did nationally, won the 30–39 age group 46%-45% (which Clinton won national 51%–39%), and he won college graduates in Michigan (another category Clinton won nationally). Both candidates won 99% of the support from voters that viewed them favorable, but among voters that viewed them both unfavorably, Trump won this group 50%-29%. This was about five points higher than the split on this question nationally.
In the primary, Clinton only won 4 counties in Michigan, a precursor to potential under performance in this state. Biden won every single county with a higher turnout among democrats than four years earlier. Especially after some of the gains made among these groups in 2018, it is unlikely he will be able to maintain these level during a referendum election.
However, and this will be repeated numerous times. Nationally, the turnout split among urban-suburban-rural was 34%/49%/17% with Clinton winning the urban vote 60%, Trump winning the rural vote with 61%, and the suburban vote being roughly a wash. In Michigan, the rural vote was 27% while the urban vote was just 23%. Clinton ran up a 71% win among urban voters, but this massive unprecedented turnout among rural voters is what catapulted him. He can win these states again if he turns out that base.
Virginia (Biden, +10%)
Trump has not led Biden is a single head-to-head poll during the 2020 cycle. However, before Obama in 2008, Virginia had only voted for a Democrat once since 1950, and the Republicans were winning the state by nearly 10 points in both 2000 and 2004. So why such a huge shift and why is still on a watch list?
The major shift in Virginia has been away from rural voters. In 2004, 25% of Virginians voted from the sticks with Bush winning that vote with 56%. In 2016, it was down to 19%. Furthermore, Virginia has increase 7 points, to 30% of the electorate, in urban areas of the state.
Over half the voters in Virginia are college educated, a five point advantage over the rest of the country and from where they were slightly over a decade ago. Lastly, in 2004, 72% of Virginian voters were white (Bush, 58%) while in 2016 it was down to 67% (Trump, 59%).
These shifts may not seem like much. But when you see that over the course of 16 years, white voters in Virginia voted for two very different Republican candidates at an almost identical clip shows how far demographic shifts go in explaining an election.
Texas (Trump, +4%)
Truth in advertising: I don’t think there’s a chance that Texas goes blue. That being said, significant money will have to be spent here for the first time in more than a generation. The trend lines that almost caused a candidate like Beto O’Rourke to beat Ted Cruz highlights some of the future problems for the current path of the Republican Party.
The reason why is race. In 2008, 63% of Texan voters were white and they trended for McCain at 73%. Two cycles later, that number is down to 57% (for the Beto/Cruz race it was 56%). The black vote has stayed consist at 12%, supporting Obama at 98% before dropped to 84% support for Clinton. Beto received 90% support just two years ago.
The story here is the Latino vote. Representing just less than 20% of the vote in 2008, it has risen to just over 26%, with 2/3 of those voters consistently supporting Democratic candidates.
The gender difference is interesting in Texas. Clinton under performed her national line by 5 points, barely edging out Trump. In 2008, McCain beat Obama among women 52–47. In 2018, Beto won women with 54% with the gender split 50/50. If Biden were able to keep up with Beto at 54% and the gender split returns to 53% female electorate in Texas, where it has been for every Presidential contest since 2000, this state could be in play.
Ohio (Trump, +1%)
I still put Ohio high on the list based almost entirely out of tradition, but until I am shown otherwise I don’t believe Ohio is winnable for the current Democratic party. The shift in the last decade toward appealing toward college education, suburban, and female voters have created major gains in areas we’ve already covered. It has cost them dearly in Ohio.
Ohio was the only swing state that Trump won every education group, including postgraduates (49–46). He won every cross section of white voters and he won the 18–44 age group. Though, this all pales into comparison to one demo in Ohio that paints the entire picture: Trump won union households in Ohio, 54–41%, by 13 percentage points.
Trump did comparatively well in the urban areas of Ohio, won by 20 points in the suburbs (which didn’t happen anywhere else among swing states), and completely slaughtered in the rural areas. Just 44% of their voters are college education and after handing their state to Obama in 2008, there was a severe backlash.
It may only equate to 23% of the electorate, but it is a huge base of the Democratic party and without it, there is no chance for a democrat to win the state. Biden has spent a career putting blue-collar union voters on his back so it will be interesting to see if that helps. But until I see strong evidence of it, Ohio is gone for the Democratic party.
North Carolina (Biden, +2%)
North Carolina is another example of the general under performance that Hillary had across the board. Pundits have tried explaining this is many forms, many of which may be true, but the reality is that someone with her experience should have delivered in North Carolina.
Representing the same split of the electorate (46/54), McCain and Trump both got 56% of the male vote, but Hillary performed five points worse. This could be expected, as much like Obama’s exploded support among the black vote, all she had to do was pad the margin among women.
Instead, she performed 4 points worse than Obama did in 2008 among women, while Trump delivered the same percentage of women as McCain.
Meanwhile, black turnout decreased as well as support and Trump had some of his best support among the Latino population in North Carolina. The state has seen a huge increase in college education voters, but Democrats have yet to make inroads with them as both Trump won college educated voters (by a slightly lesser margin than McCain).
Romney flipped North Carolina in 2012, based mostly on a great turnout in the urban region, increase among younger and older voters, as well as women. Trump didn’t do as well as Romney in replicating those results, but combined with Hillary’s inability to replicate Obama’s numbers in the state, it was just enough.
Wisconsin (Biden, +3%)
There is a lot of evidence has become the new bellwether state, surpassing Ohio as it becomes more monolithic. I won’t dive into their last three cycles of Supreme Court races, but combined with their other statewide contests, they don’t seem to be able to have an election where one party runs away with it outside of any reasonable margin of error.
Compared to nationally, democrats in Wisconsin seem to be swimming uphill: college graduates represent a less percentage of the electorate, they are one of the least diverse among the swing states, and it easily has one of the biggest rural populations (and smallest suburban areas) as well, making the needle that needs to be thread for Biden very thin.
In 2008, Obama swept every age category, he won the male vote, he won the white vote, he won every geographic region (winning the rural vote for 10 points), and won all education level categories. McCain won Republicans and those making over $75,000.
In contrast, Trust won the 18–24 age group (albeit only 9% of the vote) and Wisconsin was the only state for Trump won that age group and didn’t sweep the entire category in a state. Trump almost won white women (49–47), college graduates (49–45), tied for the 2nd-lowest income group, and he won independents by 10 points (Obama won them by 20 points).
When people talk about the Obama coalition and Obama voters that abandoned him to vote for Trump, there were many states where this happened that I’ve outlined above. But there is so state where this is more stark than Wisconsin. In an election where he won’t have the winds of issue voters at his back, will some of them come back into the fold?
If Wisconsin doesn’t turn blue in 2020, I don’t see many scenarios where Biden can take the White House.
Next time I will cover the next set of states that will fill up the election cycle as well as pivot toward some issue polling.