With the Inauguration only a day away, Democrats and the media continue their political postmortem of the 2016 election, trying to understand how they missed what the American electorate was thinking. The latest attempt revolves around a Comey-Putin narrative to explain the Democrats’ unexpected loss.
But an analysis of the national exit polls and a post-election Winning the Issues survey, done by the Winston Group on Election Night before the results were known, should put to rest the question of whether James Comey or Vladimir Putin tipped the scales against Secretary Clinton. The answer is no, and that conclusion isn’t speculation. It’s based solely on data.
To determine what issues were most important in driving voter decision-making, especially with late deciders, we tested a range of major policy issues, asking voters to rank them on a one-to-nine scale in terms of how important each was to their decision-making. We also included issues that dominated the news throughout the campaign, especially into the fall, such as the e-mail scandal, Russian hacking, Benghazi, Comey, Trump’s most controversial statements, the Clinton Foundation, and others. Not surprisingly, both partisan bases showed more interest in these issues, but even with the bases, none of these “scandal” issues placed in the top five.
Most important, independents put these issues at the bottom of their rankings. And this is where the media made another major miscalculation that continues today. For voters in the political center, the media’s fascination with what amounted to the “clash du jour” was white noise — in the background but not central to how they would choose the next president. Voters ranked Comey’s actions 17th in importance, and the allegations of Trump’s ties to Russia came in dead last at 20th.
I will leave the “who did what when” of both Comey and Putin to the official investigators. That isn’t the question under discussion here. What our analysis looked at was what actually motivated voters — especially those who decided in the final weeks — to elect Trump over Clinton. The exit polls and our survey found a different explanation than the one the media are driving today. The economy and jobs issue drove the election. It was not outside political forces, foreign or domestic; nor was it Comey or Putin.
What really happened in the 2016 election is not very complicated. It begins with understanding that the big unknown in this election rested on the fact that a significant number of voters, about one in five, had a negative opinion of both candidates. A lot of commentary at the time noted the high negatives of Trump and Clinton. What mattered more, however, and what would determine the final outcome of the campaign, was the context in which those voters who had a negative opinion of both would decide.
That unknown context — when and how these voters would make their choice — added significant volatility to the race. As a result, the election models, which were driving media narratives, proved unreliable. The political experts and pundits took polls and predictive models at face value because they almost entirely missed the volatility that was churning below the polls’ surface numbers, especially in the critical Rust Belt states.
Unhappy voters who disliked both candidates broke Trump’s way by 17 points nationally.
On Election Day, much to the surprise of the political elite, Trump won the election, in large part because these unhappy voters who disliked both candidates broke his way by 17 points nationally. In the Rust Belt States of Michigan, Ohio, Pennsylvania, and Wisconsin, the trend was even more pronounced. Trump’s margin with these voters ranged from 21 points in Michigan to 37 points in Wisconsin, winning majorities in all four key states.
The one-dimensional view of the race also led Democrats and the media to overestimate the state of play that existed in September and to underestimate the potential impact of the volatility in October and November. While 73 percent of voters made up their minds before October, with Clinton leading by six percentage points at the time, the exit polls showed that those who decided in October broke for Trump 51 percent to 37 percent. In the last week, Trump’s lead slipped, but he still won late deciders 45–42.
In the Rust Belt states, voters who decided the last week broke even more significantly for Trump. The numbers show that this late shift was enough to change the outcome of Michigan, where he was tied, and Wisconsin, where he was down two points going into the last week. Despite Clinton’s lead in national polls up to the last weekend, the traditional Democratic coalition in Michigan, Ohio, Pennsylvania, and Wisconsin was in trouble as frustrated voters, who felt left behind economically, struggled with their choice.
It came down to which candidate they believed would bring real change.
So what drove undecided voters, especially those with negative views of both candidates, to choose Trump? Ultimately, for 39 percent of the electorate, a significant number, it came down to which candidate they believed would bring real change. Of those who decided in the last month, that number increased to 46 percent. This was particularly true in the Rust Belt states, where the percentage of last-month deciders who rated “change” as their top issue ranged from 43 percent in Michigan to 54 percent in Wisconsin. Overall, Trump won these change voters by an astonishing 82–14 margin nationally, and reaching 84 percent in the Rust Belt.
What was behind this desire for change? As it has been for the last four elections, jobs and the economy were the top issue in both the exit polls and in the Winning the Issues survey. In the post-election survey, voters broke down this way in rank order.
2. Foreign policy/terrorism
Independents who decided the election:
2. Get things done in D.C.
1. Make the wealthy pay their fair share
Nowhere was this more apparent than in the Rust Belt states with three key demographic groups that historically have been part of the Democrats’ coalition: union members, people earning less than $50,000, and people with less than a college education. Two examples: First, in Wisconsin, among union households, President Obama beat Romney by 33 percent. In this election, Clinton won over Trump by only 10 percent. Second, in Pennsylvania, among those making under less than $50,000 a year, President Obama beat Romney by 36 percent. In this election, Clinton beat Trump by only 12 percent.
In those states, exit polls showed that 52–56 percent of voters chose the economy as their top issue. One key voter group — those who said the economy was “not so good” — is another prime example of the shift we saw from 2012 to 2016. President Obama won this group by 13 percent over Romney. Trump carried these voters, who represented 41 percent of the vote, by a margin of 53–40. However, in the Rust Belt States, he was getting an even larger percentage of these voters, ranging from 58 percent in Pennsylvania to 64 percent in Ohio.
Overall, these results generated an outcome not seen since Reagan’s win in 1984: a Republican presidential candidate carrying Michigan, Ohio, Pennsylvania, and Wisconsin.
The bottom line is that people were unhappy with the direction of the country. They wanted someone to rock the boat. Trump promised a different future focused on the economy and jobs; Clinton represented the status quo.