Ever since humanity identified the concept of a “future,” people have been trying to determine how to predict it: palms, winds, hexagrams, shamans, oracles … every culture has sought to identify the best way to foretell events.

In recent weeks, attention has focused on several models that seem to have worked in the past and that predict the outcome of the 2020 presidential election. The models’ designers claim that they can be more accurate than using more commonly accepted voter polls to assess whether President Trump will win re-election.

The Fair Model, designed by Yale University Professor Jay Fair, uses a combination of economic growth, the power of incumbency and the voter tendency favoring change to estimate the popular vote share of the incumbent party. It got a boost from The New York Times recently, which pointed out that the model accurately predicted Trump would win in 2016, and that it predicts him getting re-elected in 2020.


The Keys to the White House Model – developed by American University Professor Alan Lichtman and Russian scientist Vladimir Keilis-Borok – uses even more factors, including the strength of the incumbent party, party unity, apparent scandals, social unrest, and charisma of the candidates. It also looks at the success of policy changes by the incumbent and economic strength to predict who will win.

Both models currently forecast that Trump will easily win re-election. And several other models, including ones that focus on the Electoral College, also suggest a second Trump term.

At the same time, recent polls suggest that over half the electorate would vote against Trump, with his job approval rating in the mid-40 percent range.

The most recent Fox News poll showed several potential Democratic nominees with solid leads over Trump, with former Vice President Joe Biden beating the president 49 percent to 38 percent.

Who deserves our faith? The econometric and incumbency modeling shamans? Or the polling oracles? You need both to understand what’s going on politically as we head toward the 2020 elections.

I’d argue the models are correct to conclude that the state of the economy and the strength of incumbency have been key factors in affecting how people think (and therefore vote).

But while the models have worked in the past, they’ve not done as good a job in predicting the vote for Trump. Indeed, both the models mentioned suggested Trump would win the popular vote in 2016 – but he lost by 2 points.

Who deserves our faith? The econometric and incumbency modeling shamans? Or the polling oracles? You need both to understand what’s going on politically as we head toward the 2020 elections.

Still, the polls are right that Trump remains vastly less popular than a president ought to be under current conditions.

The American economy is on a tear right now, and we’re about to enter the longest economic expansion in history. That is an ideal environment for an incumbent.

If voters were faced with the choice of an unnamed Republican incumbent versus any Democratic challenger under current conditions, voters would clearly opt for the incumbent by a wide popular vote and Electoral College margins.

But an unnamed Republican incumbent will not be on the ballot next year. Trump will be.

Trump’s persona takes attention away from the successes of Trump’s administration. In fact, Trump is such a powerful presence that people wind up considering not should we re-elect Current Policies (I’d bet we would), but should we sign up for another four years of The Trump Show?

The polls suggest the public is looking for a different show.

So in the coming months, I’d look at both the models and the polls.

Should Trump submerge his personality to enable the public to focus on the success of his policies, one should expect that the models will turn out to be more accurate. But if the president’s personality remains in the center ring of the political circus, I’d focus primarily on betting with the polls.

In reading polls, it’s important to look underneath the numbers. Too often, folks focus on the difference between the main candidates, and each candidate’s actual number.

In 2016, many people focused on how Hillary Clinton was ahead of Trump in many key states in the Midwest, and therefore was likely to win. But in those states where she was leading, she was actually sitting in the mid-40s, while Trump was in the low 40s.


There were a large number of undecided voters. When folks just looked at the difference, they were expecting the undecideds to split evenly among Trump and Clinton. But they didn’t. They went disproportionately to Trump, enabling him to eke out a victory in Michigan, Pennsylvania, and Wisconsin, and thus move into the White House.

If the polls continue to show the Democrat with close to or over 50 percent of the vote, then I’d expect the Dems to be able to win. Should the Democrat hover in the mid-40s – with a large undecided vote – then it’s very possible Trump will be in a position to win over a large chunk of the undecided voters and get another four years in the White House.


Source link