Dems Misread Their Baseline- And Their Base
As both sides seek to make sense of the stunning results of last week’s election, the temptation is to cling to the most sanguine interpretations while dismissing less flattering realities. Democrats are focused on a host of grievances — a rigged electoral system; voter suppression; FBI interference; unfair media treatment. Elated Republicans have convinced themselves that all is well, despite losing the popular vote for the sixth time in seven outings. Control of both ends of Pennsylvania Avenue combined with unprecedented dominance in state capitals around the country has a way of obscuring a remarkably close shave. As someone who suffered from an admittedly blinkered perspective throughout the course of the race, I humbly offer some clear-eyed insights about what this election revealed in hindsight.
Demographics are not destiny- at least not yet. For variety of reasons, from proportion to distribution, white working class voters still hold out sized sway. The electoral dominance of Barack Obama is not transferable. His formidable coalition does not simply convey to his party’s nominee, and must not be taken for granted. And in a race between two universally known and overwhelmingly disliked nominees, a restless electorate will err on the side of change. Hillary was viewed as a quasi-incumbent, and the undecideds broke accordingly.
One of the primary conceits of this campaign cycle was that Barack Obama had ushered in a new era in US politics. That his election and re-election was the harbinger of an emerging democratic majority, driven in a large part by a coalition of the ascendant — millennials, single women, African-Americans, Latinos, and others persons of color. While Republicans posted dominant showings in the ensuing midterms, that just served as proof that the electorate in non-presidential years didn’t reflect the country as a whole.
The parties had traded places — once a Congressional party locked out of the White House, Democrats suddenly enjoyed a hammerlock on the electoral college while Republicans had gerrymandered their way to a durable advantage in the legislative chambers. With once-ruby red Sunbelt states growing ever more diverse, Republicans were on their way to becoming a regional rump. Donald Trump merely served as an accelerant for this process.
Public polling confirmed as much- GOP bastions like Texas and Georgia appeared closer than the Rust Belt targets Trump needed to win. Even the Clinton campaign’s strategic decisions suggested that with 270 in hand, they were seeking to run up the score in places like Arizona. The early vote numbers only cemented this narrative. By the Friday before election day, Nevada — once a promising Trump target— was all but lost according to the number crunchers, with Hispanic voters turning out in droves. A similar phenomenon played out in Florida, albeit with less certainty as to the outcome. But the impression was clear: Trump had called down the thunder and awakened the sleeping giant of Latino voters, and now they would be his undoing.
This notion lasted until around 9 PM on election night when Donald Trump plowed through The Blue Wall like the Kool-Aid Man wearing a MAGA hat.
The exit polling data puts things into stark relief. In short, Clinton utterly failed to reconstitute the Obama coalition. Perhaps unsurprisingly in retrospect, her margins among non-white voters look as much like John Kerry’s or Al Gore’s as they do President Obama’s, even as these groups’ share of the electorate appear to have remained relatively stable. And just a small reversion toward the mean is problematic in a race this close, especially given the monolithic baseline.
For his part, Donald Trump did incredibly well among working class white voters, winning 67 percent. But Mitt Romney won 61 percent himself. How much could that uptick matter, especially when Trump fared little better among whites overall? As it turns out it does matter- big league. As the New York Times’ Nate Cohn wrote back in July, analysis of census data and voter lists suggests that Romney’s gains with this group were disproportionately located in the south. Thus his gaudy national numbers obscured significant regional differences. Not only did President Obama hold his own with working class voters in the industrial Midwest, he significantly improved on Kerry’s numbers. And as evidenced on Tuesday, there was still a great deal of meat left on the electoral bone.
When you sort the electoral battlegrounds by their relative shift toward Republicans from 2012 to 2016 the trend becomes clear. 10 states saw a swing of 5pts or more toward Donald Trump. Working class whites made up a plurality of the electorate in eight of these states. Trump flipped electoral votes in each of the six won by President Obama (71 in all). And in the two where he fell just shy — Minnesota and New Hampshire—more than half of the electorate was made up of the college educated whites.
With Trump’s 14 point gain among non-college whites offset by a commensurate drop among their college grad counterparts, his overall margin ticked up a mere point above Mitt Romney’s 20pt advantage. But this socioeconomic arbitrage carried with it geographic implications that proved critical—and given the concentration of working class white voters in Rust Belt battlegrounds, the trade accrued overwhelmingly to Trump’s benefit. While Clinton gained 10 points on President Obama’s 2012 performance among college educated whites, Trump nonetheless carried the group, and those votes he lost were spread broadly among states of varying electoral value, many of them wasted in pockets like Orange County, CA and Fairfield County, CT.
Trump’s victory is likely to be hailed as a vindication over the “missing white voter” thesis popularized by Real Clear Politics’ Sean Trende (and contorted beyond recognition by countless others.) And Trende’s analysis of the numbers and the shifting dynamics over the past eight years has proved exceptionally prescient. But the MWV shorthand may be a misnomer for what we’ve just witnessed. It’s not necessarily that these voters were missing, it’s that they were hiding in plain sight. Working class whites remained in Barack Obama’s column to a degree that few appreciated until they were the difference between victory and defeat.
And yet still the election was decided on a knife edge. Had Clinton done as well as once expected among the affluent, suburban, GOP-leaning demographics she was targeting, she might have been able to withstand both the Obama demographic hangover and the blue collar jailbreak. But in the end enough Republicans came home to deliver Trump the inside straight he needed to pull off a shocking upset.
So what happened with the polls? They generally pegged the Clinton share of the vote about right—they just undershot the Trump vote (and therefore the overall margin) enough to make the difference. This phenomenon was particularly acute in the upper Midwest, with big misses in Wisconsin and Michigan. Were the crooked media cooking up rigged numbers? Probably not, given that even the rosiest internal models from the Trump campaign and the RNC still had him coming up just short. A shy Trump vote? Maybe, maybe not. But here’s a clue that might explain it. According to the exit data, voters who decided in the final week broke for Trump by a 5 point margin. A modest advantage, not nearly enough to swing the election. But in the states where it mattered, the picture was very different. In Michigan late deciders favored Trump by 11 points. In Pennsylvania Trump carried them by 17 points. And in Wisconsin, a state few saw as truly in play, the fence-sitters—fully 14 percent of voters—broke for Trump nearly 2 to 1.
You could chalk it up to a Comey effect. You might argue it was simply reluctant Republicans coming home. Perhaps we were so focused on Trump’s ceiling that we missed Hillary’s. But what’s clear is that in the states where it ended up mattering, Clinton performed like the embattled quasi-incumbent she was. She held firm with the 46 percent of the vote she brought into Election Day, but she simply couldn’t get over the top.
The exits lend further credence to this theory. When voters were asked which candidate quality mattered most, a strong plurality cited ability to “bring change.” These voters preferred Trump, 83 to 14. Likewise, 62 percent of the electorate said that the country was on the wrong track- these voters broke for Trump 69–25. That’s not to say they necessarily liked Trump or even thought him to be up to the task. 60 percent of voters reported an unfavorable opinion of Trump. An identical 60 percent didn’t even see him as qualified. Clinton’s fared significantly better in both metrics. Yet on both questions, a significant chunk of voters cast a pox on both their houses. And among those with an unfavorable opinion of both, Trump won by 20 points. As for the 14 percent who found each unqualified, Trump prevailed by a staggering 55 point margin—more than 4 to 1. That’s the ballgame.
The failure of Clinton to excite, turn out, and sustain Obama margins with the so-called Rising American Electorate hurt her badly. But even more fatefully, the Clinton campaign — and Democrats generally—appear to have misread the very make-up of their coalition, and in doing so ignored the apparent glue that was holding together the deceptively fragile Blue Wall. In the end, a little more than 100,000 votes separate genius and glory from folly and failure. Democrats are left to ponder what went wrong from the electoral wilderness while Republicans marvel at the scope and scale of their newfound political power. Peering prematurely around the electoral corner, both sides are balancing great opportunity with broad exposure, a dissonant echo of previous waves. But the 2016 campaign experience serves as a healthy reminder that in politics things are never quite as good as they seem, nor as bad as they appear.