A central question of the 2020 presidential election is this: What will become of the Obama-to-Trump voters? These voters totaled approximately 6 million and were crucial to the outcome of the 2016 election. Without them, Donald Trump would have lost to Hillary Clinton in the Electoral College by a sizable margin. What should we expect of them this time around?
A study I conducted with my colleague Jiwon Lee, using the American National Election Study — the most precise accounting of actual voters — offers some clues.
While diverse in many ways, Obama-to-Trump voters are more likely to be white, working-class and to live in the Midwest. Because they voted for Barack Obama in 2012 before supporting Trump in 2016, we can assume that their votes were not informed solely (or even strongly) by racial preference.
In 2016, these voters had comparatively low levels of political interest and did not see a reason to strongly support the positions of either party's establishment leaders. As a consequence, they tended to make their voting decisions late in the game and were less likely to talk politics with friends and family than were party loyalists. We have seen no evidence that social media shaped their voting preferences in 2016.
On economic matters, Obama-to-Trump voters are centrists, except when it comes to free trade, which they view as a greater threat to jobs and wages than both Democratic and Republican party loyalists.
While some have argued that these voters' emergent support of Trump is attributable to surging white nativism and other types of racial identity, we favor an explanation that places identity responses in a secondary position. Those who voted for Obama in 2012 and were attracted by Trump's economic populism only joined his coalition if they could accept, or at least ignore, his disparaging remarks about Mexican immigrants, his prior claims about Obama's citizenship and his criticism of Black Lives Matter.
Indeed, this argument implies that if Trump had used less racialized rhetoric, without changing his positions on trade and immigration, he might have won more easily and not had to rely on the electoral irrationality of those who voted for Jill Stein, the Green Party candidate. Had the Stein voters supported Clinton instead, she would have carried Michigan, Pennsylvania and Wisconsin.
What could Clinton have done differently? Her campaign, either because it was hemmed in by Bernie Sanders or because it employed a poor strategy, undermined its chances by failing to craft an economic argument with sufficient populist appeal. It would not have been hard to argue that a debate on appropriate levels of immigration was just as important as a debate on a citizenship pathway for Dreamers, those brought to the country as children.
And it would not have been hard to embrace the idea that it was time to reassess the North American Free Trade Agreement, more than two decades after its passage. The battleground states Clinton lost were engines of prosperity until the 1970s. The passage of NAFTA in 1994 did not reverse the precipitous declines in labor demand for lower-skilled workers that began in the 1980s, and increased trade with East Asia since then hasn't done so, either. Global trade has generated many benefits, but too few of them for high school graduates in the Midwest, or elsewhere.
The greatest risk Democrats face in 2020 is nominating a candidate who does not appeal to the economic interests of voters in the middle of the ideological spectrum, who have revealed that they have populist economic impulses. At the same time, Democrats should not nominate a candidate who is so threatening that the Republicans who supported Gary Johnson in 2016 decide to finally support Trump in 2020. If Johnson's voters return to the Republican Party — perhaps because they approve of Trump's tax cuts and appreciate how recent economic growth has promoted their own economic interests — then the usual risk aversion that supports incumbents' advantage could be enough to re-elect Trump, even if many Obama-to-Trump voters are won back.
A Democratic proposal to undo the Trump tax cut for the highest-income households could be attractive to centrist voters, especially if it is pitched as a way to invest directly in infrastructure and K-12 public schooling. Voters would also likely rally to a targeted "fresh start" program that pays off the loans of students who were recruited by for-profit colleges that exaggerated their graduation and job-placement rates, especially if the program is rolled out aggressively to support veterans who have accumulated debt without acquiring bachelor's degrees. However, if the plan is pitched as a way to make college free, or to pay back the student loans of college graduates employed in an economy that has been restructured to their benefit, it will backfire and drive away centrists.
Should the Democratic Party avoid nominating Cory Booker, Julian Castro or Kamala Harris, based on the argument that they cannot win back Obama-to-Trump voters because they are not white enough? Such arguments are wrongheaded. Voters whose support is determined strongly by racial attitudes chose their parties long ago. The voters who are up for grabs are those who do not vote based on racial matters, and that is why they backed Obama in 2012 before shifting to Trump in 2016.
The message for both parties is clearer than ever: It's still the economy, but now it's more than unemployment rates and economic growth. The country needs a new plan to distribute the wage benefits of economic growth more broadly. Trump has a plan: So far, he has managed to talk tough on immigration and start a trade war without destabilizing the economy. Democrats need to spend more energy on figuring out how to do better. Vacuous statements about new green jobs and the economy of the future were not persuasive in 2016, and they won't be in 2020, either.
Stephen L. Morgan is a Bloomberg Distinguished Professor at Johns Hopkins University. He is a sociologist and a principal investigator of the General Social Survey.
© 2019 Bloomberg News