The 2016 presidential election was supposed to be close — but Republicans should have had the edge.
An election forecast built by Vox and a team of political scientists projects that a generic Republican should win 50.9 percent of the two-party vote in 2016. But Donald Trump isn’t a generic Republican — and he is polling at of the two-party vote, according to the Huffington Post Pollster. The difference between those numbers — , as of today — is what we’re calling the Trump Tax: the electoral penalty Republicans appear to be paying for nominating Trump.
Vox’s model projects the campaign's outcome based on the most historically accurate political science models incorporating "fundamentals" like the state of the economy, the president's popularity, and so forth. There's a margin of error, naturally, with estimates for the Democratic vote share ranging from 43.61 percent to 53.44 percent, but the point estimate was a narrow GOP victory — 50.9 percent of two-party vote, to be precise.
During the primary, head-to-head matchups between mainstream Republicans and Hillary Clinton reflected the Republican edge, and perhaps even Clinton’s weaknesses. RealClearPolitics' last average in early March put Marco Rubio up 4 points over Clinton. John Kasich versus Clinton was a veritable bloodbath, with Kasich an astonishing 7.4 points up as of late April. Trump's unpopularity was evident even then: Those same polls showed Trump trailing the Democrat, often by double digits.
What we appear to be seeing is a remarkable example of a major political party blowing a totally winnable national election.
There is uncertainty here, of course. Elections are complex things, and we can’t rerun this race in laboratory conditions. It is possible the difference between the GOP’s expected performance and Trump’s performance reflects an economy stronger than the GDP data indicates, or that Clinton is a far more effective campaigner than conventional wisdom (or her poll numbers) suggests. But given the difference between where Trump is polling and where more mainstream Republicans were polling, we believe the likeliest explanation is that this is the cost Republicans are paying for nominating a very unusual candidate who is running a very unusual campaign.
For much more detail on the model, how we built it, what the uncertainties are, and what it teaches us, head here.
The Trump tax waxes and wanes with changes in Trump and Clinton's polling. In mid-July, as Trump enjoyed a polling bounce coming out of the Republican Convention, the tax shrunk. All the same, Trump never improved enough to match the fundamentals model. And after the Democratic Convention, Clinton's bump pushed Trump further behind than he'd been all summer, and weeks later, that bump shows no signs of dissipating.
Forecasts from FiveThirtyEight, the New York Times' Upshot, and Princeton's Sam Wang rely primarily on polling, not fundamentals. Accordingly, they give Clinton a very high probability of victory. PredictWise, which averages betting and prediction markets and thus reflects conventional wisdom at the given moment among people willing to place money down, is also much more bullish on Clinton than Vox's fundamentals model.
Vox's model, developed by Washington University in St. Louis's Jacob Montgomery and Texas A&M's Florian Hollenbech, is a weighted average of six academic forecasting models. Three of those models have Trump ahead, and three have Clinton, but the one with the best past track record, Abramowitz's, predicts a Trump win.
Sign up to receive Vox Sentences and email updates straight to your inbox