President Donald Trump says he has a mandate to build a border wall. “I was elected partially on this issue,” he declared at the White House on Thursday. Trump thinks that by claiming to have won the 2016 election based on his pledge to build a wall, he can bully Congress into accepting it. He’s bluffing. A close look at election results in 2016 and 2018 shows that Democrats who reject the wall—even those who represent states Trump won—are standing on firm ground.
Trump’s bluster about a mandate boils down to two claims. The first is that most Americans support a wall. “The people of our country, they want the wall,” he asserted at the White House in October. Three weeks ago, while visiting the Texas border, he suggested that Congress should fund the wall because “most people want it.” Feigning reasonableness, he allowed: “I guess there could be some pockets where they maybe disagree a little bit. But not very much.”
The second claim is more specific. It’s that when voters elected Trump, they did so because he had promised to build the wall. In November 2017, Trump told Democrats that he would beat them in a shutdown fight because the central issue would be the “border wall, which everybody wants. I got elected partially because of the border wall.” In June, he repeated, “That’s why I got elected.” Last month, in the middle of the latest shutdown, he vowed not to yield in his demand for a wall. “The people that voted for Donald Trump … those people are for it so much,” he declared. “And let me tell you: People that didn’t vote for Donald Trump are for it also.”
Trump is totally wrong. Let’s start with 2016. We don’t have to guess what voters thought of the wall that year because they were asked about it the national exit poll. The question in the poll was, “How do you feel about building a wall along the entire U.S. border with Mexico?” Fifty-four percent of voters opposed it. Only 41 percent favored it.
Trump got 46 percent of the vote—less than Hillary Clinton’s 48 percent, but enough to beat her in the Electoral College. He won Michigan by 11,000 votes, Wisconsin by 23,000, and Pennsylvania by 44,000. That’s fewer than 80,000 votes out of more than 130 million cast. He needed every ballot he got. And many of those ballots came from people who disagreed with him about the border. Fewer than 35 percent of voters supported a wall and voted for Trump. Another 9 percent opposed a wall and voted for him. That 9 percent, along with voters who took no position on a wall, put him over the top.
You could argue that this wasn’t a fair test because the question specified a wall along the “entire” border. You could also object that the question was asked in some states but not in others. So let’s dig into the numbers. The wall question was asked in eight states: Arizona, California, Colorado, Illinois, New Mexico, New York, Nevada, and Texas. Trump won two of those states: Arizona and Texas. But in none of the eight states did a majority of voters, or even a plurality, endorse the wall. And in none of them did wall supporters give Trump enough ballots to reach even 40 percent of the vote.
California, Illinois, and New York were wipeouts, so let’s set them aside. Of the remaining five states, Trump won two, and Clinton won three. In each of these five states, roughly 7 to 8 percent of respondents said they opposed a wall but voted for Trump. Subtract those voters, plus the few who didn’t answer the wall question, and Trump would have been swept. In Nevada, Trump got 46 percent of the vote, but wall supporters accounted for only 36 percent. In Arizona, he got 48 percent, but wall supporters accounted for only 38 percent. In Texas, he got 52 percent, but wall supporters accounted for only 39 percent.
So Trump’s election wasn’t a mandate for a wall. Wall supporters were outnumbered by 14 percentage points. They didn’t give him enough votes to reach 40 percent in any of the tested states—including Texas, where he won by 9 percentage points.
But let’s move on to the second objection, about the wording of the question. The 2016 exit poll specified a wall along the “entire” border. Maybe that was too extreme. Is there a mandate for a simpler project? What happens if we just ask voters about a wall?
The network exit poll didn’t ask that question in 2016, and it didn’t ask any question about a wall in the 2018 midterms. But a separate survey of the midterms, conducted by NORC at the University of Chicago for the Associated Press, did. This massive project, designed to capture early voters as well as Election Day voters, sampled roughly 120,000 people. It asked them, “Do you favor or oppose building a wall along the U.S.-Mexico border?”
There was no presidential election in 2018, so turnout was somewhat lower, especially in states that didn’t have competitive Senate or gubernatorial races. Other factors also affected turnout: Progressive voters went to the polls to oppose Trump, while anti-immigration voters came out in response to his scare talk about migrant caravans and crime. The gap between wall supporters and opponents wasn’t as big as it had been on the 2016 exit poll question. But again, opponents outnumbered supporters, 53 percent to 47 percent.
The 2018 poll, unlike the 2016 poll, asked voters in every state about the wall. That means we can zero in on the three states Trump narrowly won in 2016: Michigan, Pennsylvania, and Wisconsin. In each of these states, most voters in 2018 opposed the wall. So voters who supported the wall—and who showed up at the polls specifically because Trump emphasized that issue in the weeks before the election—weren’t enough to deliver those states to Trump or the GOP.
In deep red states, most midterm voters supported the wall. But even then, they weren’t sufficient to elect Republicans. Only one Republican Senate candidate, Marsha Blackburn of Tennessee, got enough votes from wall supporters to win her race. Every other Republican Senate candidate either lost or depended on wall opponents to deliver the winning margin. That includes Mitt Romney, who won in Utah by 32 percentage points, and Deb Fischer, who won in Nebraska by 19 percentage points.
Look at national polls taken in the past month, and you’ll see the same pattern. “Do you support or oppose building a wall along the border with Mexico?” Oppose, 55 percent to 41 percent. “Do you support or oppose building a wall along the U.S. border with Mexico?” Oppose, 54 percent to 42 percent. “Would you favor or oppose building a wall or a fence all along the border between Mexico and the United States, from Texas to California?” Oppose, 52 percent to 45 percent. “There is a proposal to substantially expand the wall along the U.S. border with Mexico. In general, do you favor or oppose this proposal?” Oppose, 58 percent to 40 percent.
Most people who voted for Trump in 2016 supported a wall. But nobody else, certainly not congressional Democrats, should care. If your job is to represent people in Michigan, Pennsylvania, Wisconsin, or most other states, your constituents oppose a wall. There’s your mandate.