Judging by my social-media feed, a lot of people were shocked by the results of a newly released opinion poll from the New York Times and Siena College, which showed that Donald Trump remains “highly competitive in the battleground states likeliest to decide the election”—Arizona, Florida, Michigan, North Carolina, Pennsylvania, and Wisconsin. The surprise is a bit surprising. Presidential elections are nearly always close, battleground states have earned the label for a reason, and Trump’s approval rating has remained pretty steady since he was inaugurated. (So far, the impeachment process has only knocked it down a couple of points.) On the basis of these factors, you would expect the 2020 election to be tightly fought, particularly in swing states where, in 2016, Trump eked out a victory in the Electoral College despite losing the popular vote. The Times poll merely confirms this supposition.
The poll has also drawn attention because of what it shows about head-to-head matchups between Trump and the three leading Democratic candidates: Joe Biden, Elizabeth Warren, and Bernie Sanders. The poll showed Biden leading Trump among registered voters in four of the six states: Arizona, Florida, Pennsylvania, and Wisconsin. It showed Sanders ahead in three states: Michigan, Pennsylvania, and Wisconsin. Warren was leading Trump in just one state: Arizona. However, she was running level with him in two others: Pennsylvania and Wisconsin.
These findings confirmed what other polls have found: in head-to-head polling against the President, Biden fares the best of the Democrats. Sanders is popular in the Midwest. Warren has work to do. But the numbers were too bunched up to support any definitive conclusions or predictions about next year’s election. In fifteen of the eighteen matchups—three for each battleground state—Trump was either leading or trailing his Democratic opponent, among registered voters, by three percentage points or less. Taking the six states together, Trump was losing to Biden by two points, running even with Sanders, and beating Warren by two points. All of these results were within the poll’s margin of error.
The key takeaway isn’t that one Democrat is likely to win and another is likely to lose. It’s that the Democratic candidate, whoever it is, needs a convincing strategy for winning at least some of the battleground states that Trump carried last time. Failing to focus on this goal relentlessly would be inviting a repeat of 2016.
At least mathematically, the elements of a successful battleground-state strategy are clear. The Democratic candidate needs to excite voters in the Democratic base, particularly minorities and highly educated whites, while also trying to appeal to as many people as possible in Trump’s core demographic, which consists of whites who don’t have a four-year college degree. Contrary to some analyses, both of these things are necessary: it isn’t an either-or choice. The Democrats need a dual strategy.
A new study by Ruy Teixeira and John Halpin, two Democratic-leaning political scientists who have been tracking demographic and voting trends for decades, shows why. According to the study, roughly twenty-nine per cent of eligible voters next year will be nonwhite, about thirty per cent will be whites with college degrees, and slightly less than forty-two per cent will be whites without college degrees. On this basis alone, it seems likely that the Democrats could put together an electoral victory simply by turning out their base. But that is misleading, because none of the three major groups vote monolithically.
In 2016, about a third of Hispanics and Asians voted for Trump, according to Teixeira and Halpin’s figures, and so did more than four in ten college-educated whites. Conversely, even as Trump racked up a huge margin among white non-college-educated voters—thirty-two percentage points—almost a third of the people in this category voted for Hillary Clinton.
Regional differences also complicate things. In much of the Midwest, which has long been a key electoral region, non-college-educated whites still constitute a majority of the voters, or close to it. Teixeira and Halpin project that in 2020 this group will make up roughly fifty-six per cent of the eligible electorate in Wisconsin, fifty-two per cent in Michigan, roughly forty-nine per cent in Pennsylvania, and fifty-two per cent in Minnesota, which Trump lost narrowly in 2016 and is targeting again.
Because candidates can’t rely on monolithic voting patterns, they can’t rely on monolithic electoral strategies either. Successful Presidential candidates, even as they target their core supporters, somehow manage to limit their losses among groups that aren’t inherently favorable to them. That is what Barack Obama did in 2012, when he held Mitt Romney’s victory margin among white non-college-educated voters to twenty-two per cent, while racking up big victory margins among minorities and highly educated whites. This two-step garnered him three hundred and thirty-two votes in the Electoral College.
Given Trump’s popularity among working-class whites, and the emphasis that he and his campaign are placing on their vote, it would be very difficult for any Democrat in 2020 to match what Obama did in 2012. But this doesn’t mean that the Democrats should give up on this demographic. Even just preventing Trump from expanding his 2016 margin among non-college-educated whites could be sufficient to deny him a victory in key battleground states, and in the election over all, Teixeira and Halpin argue.
This argument hinges on two things. Trump’s margin of victory in 2016 was extremely slim; and the number of white non-college-educated voters, even as it remains large in absolute terms, is set to fall by about two percentage points in 2020 relative to other groups, such as college-educated whites and Hispanics. If everything else, including voting patterns and turnout rates, were to stay the same, that demographic change could prove decisive. As Teixeira and Halpin put it: “The Democratic candidate would take back Michigan, Pennsylvania, and Wisconsin to carry the Electoral College by 279 votes to 259 votes.”
None of this means that the Democrats should limit efforts to mobilize minorities, college graduates, and other Democratic-leaning groups. To the contrary, it is absolutely imperative that they continue, for example, launching enrollment drives in black neighborhoods in Milwaukee and taking steps to cement their 2018 gains in affluent districts north and west of Philadelphia. That is what it means to follow a dual strategy of attacking Trump’s weaknesses and trying to neutralize his strengths.
And paying attention to working-class white voters doesn’t necessarily mean tempering progressive policy proposals like raising taxes on the rich, tackling political corruption, providing universal day care, and guaranteeing health care to everyone. Sanders supports all of these things, and the Times poll showed him leading Trump by two points in Michigan. An anti-establishment, spread-the-wealth stance that emphasizes bread-and-butter issues can elicit broad popular support. Trump realized that in 2016. Speaking of his tax plan in Gettysburg, Pennsylvania, two weeks before the election, hesaid,“The largest tax reductions are for the middle class, who have been forgotten.” (Of course, the Trump-G.O.P. tax bill of 2017 gave crumbs to the middle class, and very large handouts to rich people and big corporations.)
The fundamental point is that the Democrats need to lay out a policy platform that appeals to a wide range of Americans, regardless of their race, location, and educational background, while also hammering home the message that Trump is divisive, fraudulent, self-dealing, and dangerously erratic. Among white non-college-educated women, if not their male counterparts, there is already some evidence of a willing audience for this narrative. Admittedly, it doesn’t come from the new Times poll, which found that “the president’s lead among white, working-class voters nearly matches his decisive advantage from 2016.” But other recent surveys have shown different results.
For example, a poll of seventeen swing states carried out last month by Democracy Corps, a nonprofit Democratic research organization, found that while Trump had retained his huge lead among non-college-educated men, Biden and Warren were both running even with him among working-class women. (In 2016, Trump carried this group by twenty-seven points.) In amemoanalyzing the poll’s findings, Stan Greenberg, the veteran pollster who is one of the founders of Democracy Corps, pointed out that working-class women represent nearly a quarter of the electorate, and erasing Trump’s big lead in this demographic would be a “historic” victory for Democrats.
Even if the Party’s 2020 candidate falls short of drawing even with working-class women, significantly reducing Trump’s advantage among these voters would go a long way toward assuring his defeat. Above anything else, that has to be the goal.