Some people devote themselves to model airplanes. Others prefer model trains. Alan Abramowitz builds election models — formulas that use data, economic and otherwise, to predict the outcomes of presidential political contests.
In August, the Emory University political scientist took President Barack Obama’s approval rating and stirred in the second-quarter gross domestic product. Over this, he sprinkled the “time for change” factor — eight years of White House rule by Democrats.
“Basically, you lose the advantage of incumbency that you had in the previous election,” Abramowitz said. “Put that all together and you came out with a modest Republican advantage.”
But instead of giving a late summer nod to Donald Trump, Abramowitz disavowed his own numbers and forecast a November victory by former secretary of state Hillary Clinton. His model, he decided, was predicated on a mainstream candidate who was running a competent campaign.
“It seemed to me that [Trump] didn’t satisfy either one of those conditions,” Abramowitz said. In his own defense, the academic noted that his model was intended to predict the national popular vote – not Electoral College balloting.
“As it turned out, [Trump] did under-perform – because Clinton is going to end up winning the popular vote by about one percentage point,” Abramowitz said.
Even so, the episode fits neatly into a 2016 presidential contest that turned on what we refused to see. Clinton supporters were blind to a candidate tied to a 30-year reputation for evasion and close dealing. Trump enthusiasts looked past the misogyny, the race-baiting, the strong-man bluster, and their candidate’s aversion to facts — because they had finally found the grenade they had longed to hurl across the Potomac and into the District of Columbia.
Journalists may have seen hints of Nuremberg in some faces at Trump rallies, but were sightless when it came to the fury of an aging white electorate that believes it is being hustled off the stage too roughly, and too soon.
We shouldn’t be that surprised. Blurred vision is a rather routine political condition. Two days after Trump’s victory, a sentence in the middle volume of William Manchester’s biography of Winston Churchill leaped out at me:
“The prescient and the cock-eyed always arrive in a promiscuous rush, and most men in power, sorting through it, believe what they want to believe….while taking out insurance, whenever possible, against the possibility that the truth may lie in their wastebaskets.”
But even if belief often divorces itself from the facts, accuracy remains a virtue – one that is becoming harder and harder to come by when dealing with numbers, Abramowitz said. Polls, for instance.
Many national surveys, measuring the popular vote, were on the money or within the margin of error. Certain statewide polls before the election missed the Trump surge. Abramowitz isn’t sure why.
“One possibility is that late deciders broke for Trump. There’s some indication of that in the exit polls, but not enough,” he said. Another theory relies on “the shy Trump voter” who was too embarrassed by his candidate’s antics to admit his support. A stronger argument may simply be the fact that Trump relied heavily on support from less educated white voters – a segment of the population that’s hard to track down on the phone, and wasn’t rated high enough in turnout models.
The closely divided nature of the American electorate is another problem. The margin of victory is becoming ever slimmer, and thus harder for polling to detect. Early Wednesday morning, Trump victories in three Rust Belt states sealed Clinton’s fate.
And yet. “You add together Trump’s total popular vote margin in Michigan, Wisconsin and Pennsylvania – it’s about half of what Clinton won by in DeKalb County. Just to put it in perspective,” Abramowitz said. “Too bad DeKalb doesn’t have its own electoral votes.” Even though its population is larger than that of Wyoming.
Which is another matter that pollsters need to take up. Five times in this nation’s 227-year electoral history, the losing presidential candidate has won the popular vote. Two of those times have been in the last 16 years – with Democrats as the losers.
“I’m starting to think it’s not a coincidence,” Abramowitz said. One reason: Democratic populations are becoming more concentrated in large states.
At the same time, Republican domination of smaller states has introduced a GOP-bias into the Electoral College that is somewhat new. Electoral College votes are allocated according to the number of population-based congressional districts in each state.
But states also get two more votes – one for each U.S. senator. Of the 20 least populous states, Republicans now control 12, compared to five for Democrats, Abramowitz said.
Circling back to our original topic, I asked the Emory academic if he regretted not giving more weight to his own predictive formula in August. “Not entirely,” Abramowitz said. “But I think I exaggerated the importance of Trump’s peculiarities. He was not as weak a candidate as I thought he was.”
But again, the failure to accept a fact in hand isn’t as uncommon as we sometimes make it out to be. Nor, the Trump campaign might argue, is it always a bad thing.
In a final burst of late October campaigning in Georgia, Donald Trump Jr. breezed through Atlanta, making an early morning appearance before a group of black Republicans. “We have to have people in positions that actually know what they’re doing,” the younger Trump namesake said, expertly glossing over the fact America had never elected a president without prior political or military experience.
Sometimes, ignoring the data is simply the smart thing to do.