Much of the political commentary since the US presidential election has focused on two groups of party switchers: those who voted for Barack Obama in 2012 and Donald Trump in 2016 and those who voted for Mitt Romney in 2012 and Hillary Clinton in 2016.
Trump voters who previously voted for Obama are the subject of intense fascination because they are viewed as providing critical insights into the racial and class dynamics that helped determine the outcome of the election. On the other side, many analysts see Romney voters who flipped to Clinton as an illustration of how the Democratic Party survives in significant part by appealing to more upscale voters.
Frustratingly, however, these perspectives play down the importance of a crucial group of disaffected voters: those who voted for Obama in 2012 but then failed to go to the polls in 2016. Because this group is disproportionately young and black, this erasure is racially tinged.
Our analysis shows that while 9 per cent of Obama 2012 voters went for Trump in 2016, 7 per cent — that’s more than 4 million missing voters — stayed home. Three per cent voted for a third-party candidate.
We would hardly urge Democratic strategists to abandon Obama-to-Trump voters. However, Obama-to-non-voters are a relatively liberal segment of the country who have largely been ignored. They are mostly young and non-white, and they represent an important part of the Democratic Party’s demographic future. Given the likelihood that many Obama-to-Trump voters will remain in Republican ranks, it is hard to imagine how Democrats can win elections if this group remains on the sidelines.
To explore the characteristics and attitudes of these voters, we used data from the Cooperative Congressional Election Study, a large survey with a sample of more than 64,000 adults. We grouped all 2012 voters into one of five categories, three of which we focus on in this essay: Obama-to-Trump, Obama-to-non-voter and Obama-to-Clinton. (We used validated voter turnout data rather than self-reported turnout, which tends to overstate actual voter participation and which one of us used in a preliminary analysis.)
So who were the Obama-to-non-voters? Fifty-one per cent were people of colour, compared with 16 per cent of Obama-to-Trump voters and 34 per cent of Obama-to-Clinton voters. Twenty-three per cent of Obama-to-nonvoters were under 30, compared with 11 per cent of Obama-to-Trump voters and 10 per cent of Obama-to-Clinton voters. More than 60 per cent of Obama-to-nonvoters make less than $50,000 a year, compared with 45 per cent of Obama-to-Clinton voters and 52 per cent of Obama-to-Trump voters.
Before the 2016 election, these voters were often identified as part of the “rising American electorate” by Democratic strategists who hoped that demographic shifts would be a boon to the party. But these shifts are meaningless if Democrats can’t get enough young people of colour to the polls.
What do Obama-to-nonvoters prefer, policy-wise? Are they more similar to Obama-to-Clinton voters or to Obama-to-Trump voters? The answers to these questions have important implications for the future of Democratic Party politics, so we analysed the preferences of all three groups of voters across a broad range of domestic policy areas, including support for a path to citizenship for unauthorised immigrants, abortion rights, EPA regulation of carbon emissions, cuts in domestic spending, an increase in the minimum wage, and an end to mandatory minimums in criminal sentencing, as well as opposition to the repeal of the Affordable Care Act (ACA).
Two clear patterns emerged. First, it will come as no surprise that Obama-to-Trump voters adopt the most conservative positions. In fact, Obama-to-Trump voters express the most conservative views of any Obama voters in each of the seven areas examined in this analysis.
Second, the preferences of Obama-to-nonvoters are almost always closer to the preferences of Obama-to-Clinton voters than they are to Obama-to-Trump voters. For example, nearly three-quarters of Obama-to-Trump voters supported repeal of the ACA, while less than half of Obama-to-nonvoters did. The extremely high degree of support for repeal of the ACA among Obama-to-Trump voters clearly played a role in the 2016 election and in the negative reaction to Obama among this group of voters more broadly. However, considering how strongly Obama-to-Clinton voters in particular favour the ACA, it is hard to imagine how Democrats could incorporate anti-ACA voters into future Democratic coalitions.
Obama-to-nonvoters are most similar to Obama-to-Clinton voters on the minimum wage, although the proposal draws strong majority support from Obama-to-Trump voters as well. Obama-to-Trump voters are most out of line with the Democratic coalition on issues relating to race and gender. They are less supportive of a path to citizenship, and a supermajority (64 per cent) of Obama-to-Trump voters support deporting unauthorised immigrants. At 72 per cent, Obama-to-nonvoters are also far more in favour of abortion rights than Obama-to-Trump voters are (55 per cent). Overall, Obama-to-nonvoters are quite close to the emerging Democratic consensus on issues of class, race, gender and the environment.
Democratic strategists should recognise that Obama-to-Trump voters do not represent the future of their party. Obama-to-Trump voters diverge from the Democratic Party on many core issues, and in any case they are not particularly loyal Democrats: Less than one third of Obama-to-Trump voters supported Democrats down-ballot in 2016, and only 37 per cent identify as Democrats.
In stark contrast, Obama-to-nonvoters share the progressive policy priorities of Democrats, and they strongly identify with the Democratic Party. Four out of every five Obama-to-nonvoters identify as Democrats, and 83 per cent reported they would have voted for a Democrat down-ballot. A similar share of Obama-to-nonvoters said that they would have voted for Clinton had they turned out to vote. In short, while reclaiming some Obama-to-Trump voters would be a big help to Democratic prospects, re-energising 2012 Obama voters who stayed home is a more plausible path for the party going forward.
Whether Democrats can mobilise these voters is an open question, however. Significantly, only 43 per cent of Obama-to-nonvoters reported being contacted by a candidate in 2016, compared with 66 per cent of Obama-to-Clinton voters. While analysts have focused on why many conservative voters switched to the Republican Party, a better question might be why a campaign that sought to energise young voters of colour failed to do so. That’s the question that will decide the future of US politics.
Getting these voters to the polls on Election Day is the most important task for progressives. And given their outlook on the important issues of the day, Obama-to-nonvoters are also likely to be easier to mobilise after two years of a Trump presidency — never mind four.
— New York Times News Service
Sean McElwee (@SeanMcElwee) is a co-founder of Data for Progress. Jesse H. Rhodes and Brian F. Schaffner are political scientists at the University of Massachusetts, Amherst. Bernard L. Fraga is a political scientist at Indiana University.
