As the RNC convention comes to a close, the biggest question is whether it can give Pres. Trump the bounce he needs to catch up to former Vice President Joe Biden. Trump came into the convention trailing the Democratic nominee by 8-9 points in national polling. Thus far, the RNC has portrayed the president as the man standing in the breach — protecting the country from the dangers of socialism and urban violence, while also painting a gauzy picture of a country with a rapidly improving economy where the threat from the coronavirus is receding.
But, while many are looking for signs that Trump is filling in cracks in his base, Trump's real problem is much deeper. Trump is trailing, not because he's losing his 2016 base, but because he has never expanded beyond it.
This week I dug into the most recent national poll from Pew research (7/27-8/2) and compared it with the results of their 2016 validated vote survey (basically, a post-election exit poll that uses official voting records). What you find is that Trump is hitting his 2016 share of the vote among most demographic groups. But, he's not grown beyond those voters. Instead, it's Biden who has improved markedly on Clinton's 2016 performance.
Now, the all-important caveat. The voter validated survey is of people who actually voted in 2016, while the July-August survey is of registered voters. In other words, some of the people in the July-August survey may not vote, while everyone in the 2016 survey did. But, it does help give us some perspective on how Trump is performing with key demographic groups compared to how he did with them in 2016.
You can see that in almost every demographic category, Trump's share of the vote mirrors what he got in 2016. In other words, there's no evidence that he's slipped significantly with his base, or gained support from anyone who didn't support him in 2016.
For example, for all the talk of Trump's cratering in the suburbs, the recent Pew poll finds that the president isn't doing any worse today among white, college-educated voters than he did in 2016. In 2016, Trump took just 38 percent of the vote from this group. Today, he's still sitting at 38 percent. But, Biden has improved on Hillary Clinton's 55 percent showing by 6 points to 61 percent. Trump hasn't lost support from his core white, non-college base either. The July/August poll found him taking 64 percent with this group — the same percentage he got in 2016. But, Biden has improved on Clinton's anemic 28 percent showing by 6 points. Most important, Trump has made no gains among independent voters, while Biden has improved on Clinton's showing by 14 points.
One bright spot for Trump is an increase in support from Latinos. He took 35 percent of the vote among Hispanics in July/August, a 7 point increase from his 2016 showing. His support among Evangelical Protestants has also improved — from 77 percent in 2016, to 83 percent.
So, how can this work? If Trump isn't really losing support from his 2016 base, but Biden is gaining on Clinton's performance, where are those extra votes coming from?
Answer: a lot is coming from voters who supported third-party/other candidates in 2016. According to the Pew July survey, voters who didn't support either major party candidate last election are now breaking decidedly for Biden — 55 percent to 39 percent. This group of non-Trump/non-Clinton voters doesn't get the attention of Obama-Trump voters or suburban moms, but they are a not-insignificant portion of the electorate.
In the 2016 election, the non-Clinton/Trump vote was 6 percent. The Pew validated voter survey put it at 7 percent — so a touch higher than the popular vote. In a race decided on the margins, these voters can shift the race a couple of points. Or, as Pew's Director of Political Research put it to me in an email, "If nothing else, what this shows is how much 3rd and 4th parties cost Clinton (even though, as I said, our estimate was higher than popular vote)."
There's only so far we can take this thought exercise. After all, this is just one survey. But, it does confirm what we have already seen for the last three years — that Trump has failed to broaden his base. Instead, his ability to win this election will come down to whether he can deepen his base.
That means turning out a higher percentage of his base vote and hoping that Biden and Democrats fail to meet that same level of enthusiasm with their core constituencies.
Analysis from my colleague David Wasserman earlier this year revealed a huge number of non-voting, non-college whites in the Upper Midwest. "In 2016, non-college white voters made up 50 percent of the electorate in Michigan, 49 percent in Pennsylvania and 63 percent in Wisconsin. But among non-voters, whites without college degrees were 60 percent in Michigan, 64 percent in Pennsylvania and 64 percent in Wisconsin." In other words, there's a large pool of potential voters for Trump to mine in key swing states.
But, Trump's path to building on his 2016 showing isn't looking as clear and easy as some might expect. An analysis of the most recent Marquette University Law School survey by Craig Gilbert, the pre-eminent expert of voting trends in the state, and a reporter/columnist at the Milwaukee Journal-Sentinel, finds signs of weakening for Trump in parts of the state he dominated in 2016.
Gilbert found that Biden's strength in the Milwaukee and Madison media markets is similar to Clinton's, but that the former Vice President "is battling Trump to a draw in the combined media markets of Green Bay, Wausau, La Crosse, Duluth and the Twin Cities" — media markets that encompass the kind of small-town, rural areas where Trump beat Clinton by 15 points in 2016.
This could be due to a number of factors, writes Gilbert, including Biden's "own political strengths," an "erosion of support for Trump," and/or the fact that these are parts of the electorate where Clinton underperformed in 2016 and Biden is simply hitting 'normal' Democratic performance.
Either way, it suggests that just trying to run up the score by bringing in low-propensity voters from these parts of the state, may not be enough for Trump to win here in 2020. Of course, we don't know how the events of the last 60+ days of this election may impact this turnout. Will images of burning cities — especially ones in their own state — be enough to motivate more non-Trump-like voters to the polls this November? Will Trump be able to do something the campaign has thus been unable to accomplish: drive up Biden's negatives to Hillary Clinton's level? What happens if the coronavirus comes back with a vengeance this fall?
Trump's path to Electoral College victory has always been a narrow one. But, the latest Pew poll suggests that turning out infrequent base voters is going to be more important than ever.