Hillary Clinton and Donald Trump. (Getty Images, Reuters)

Donald Trump’s poll numbers have slumped over the past three weeks. His campaign is perpetually being re-configured. Meanwhile, a new survey of political experts by the site PollyVote forecasts that Hillary Clinton will win 347 electoral votes to Trump’s 191. She is, at least tentatively, planning her work as president.

To be sure, with just under three months to go, it’s too soon to be hanging the proverbial Oval Office drapes. Trump still has somewhere between a 15 to 30 percent chance of winning, according to various poll-based election models, the prediction markets and the forecasters at Good Judgment.

But if Clinton does win comfortably, one of the most important lessons will be: It wasn’t supposed to be that way.

Back in January 2014, I wrote a post titled: “The Democratic Party’s uphill battle to 270 electoral votes in 2016.” Ha-ha, right? Silly me, right?

Not at all. At that moment in 2014, given economic growth and President Obama’s popularity and the challenge of winning the White House for a third term, the odds slightly favored the Republicans.

Fast-forward two years and the same is true. In a new article in the Annals, Michael Tesler, Lynn Vavreck and I show that economic and political conditions as of early 2016 slightly favored the Republicans. Not much has changed in the few months since we drafted that piece. Currently, the various statistical forecasting models average out to a narrow Republican lead. At a minimum, the election should be a toss-up. Thus far, it’s not.

Speaking at a campaign rally in Charlotte, N.C., Republican presidential nominee Donald Trump energized voters during his speech where he called for equality for African Americans, gays and Hispanics. (The Washington Post)

Now here’s where it gets interesting. One explanation for why Clinton is favored to win in this environment is simple: Donald J. Trump. By design, statistical models of elections don’t build in the idiosyncratic features of individual candidates. Usually that’s not an issue, because neither has a massive advantage in political talent or campaign resources.

Trump, however, seems to have unique liabilities, worse even then Clinton’s liabilities. So Vox calls the gap between the polls and these statistical forecasts as the “Trump tax.” This is how far Trump is running behind a generic Republican candidate.

But Trump alone isn’t enough to explain the Democrats’ advantage. Even in late 2015 or early 2016, before any caucuses or primaries, both the prediction markets and political experts gave the Democrats better than a 50 percent chance of winning. At least part of Democrats’ advantage predates Trump’s nomination.

Why? There’s no one clear answer. Tesler, Vavreck and I speculate that prognosticators could have been accounting for the country’s changing demography — specifically, the fact that the growing nonwhite population could favor the Democratic Party.

Most people know that whites are shrinking as a proportion of the U.S. population. But in our article we highlight these other key facts:

  • The percentage of Asian Americans voting for Democratic presidential candidates has increased sharply.
  • An increasing number of Latinos have come to identify with the Democratic Party during Obama’s presidency. Latinos were 15 percentage points more Democratic in 2012 than they had been on average from 2002 to 2007.
  • In the two decades leading up to Obama’s presidency, black identification with the Democratic Party had actually declined. This trend reversed itself after 2008: African Americans were 10 points more Democratic than they had been before Obama’s election.

More generally, American politics has seen an alignment of racial and partisan attitudes. Here are some relevant graphs from our article:



White Democrats and Republicans have diverged on how warmly they rate whites versus blacks, whether they approve of interracial dating, how favorably they feel toward Muslims and whether they support immigration restrictions. This makes it easier for Democrats — and harder for Republicans — to craft platforms that appeal to minorities.

There was nothing inevitable about this. As the political scientists Zoltan Hajnal and Taeku Lee noted in their 2011 book, “Why Americans Don’t Join the Party,” many Latinos and Asian Americans at that point didn’t have strong ties to any political party. Moreover, history is replete with cases where demographic changes were supposed to — but then failed to — give one political party a chronic electoral advantage. For these reasons, I have always cautioned against interpreting Obama’s victories in 2008 and 2012 as fundamental realignments.

One reason to be cautious about proclaiming every election a realignment is that political parties rationally adapt to demographic changes. After the 2012 election, I, among many others and many Republicans, suggested that the GOP should take a position on immigration that would appeal to Latino voters. This, I argued, would ultimately increase the party’s odds of winning future presidential elections. With the nomination of Trump, however, Republicans have gone in the exact opposite direction.

In the short term, this means that the Democrats’ uphill path to 270 votes has now become a downhill path.

In the long term, a lot will depend on what stories politicians in both parties tell themselves about why they won or lost. These stories will guide the direction of the Republican Party in particular.

If Clinton does win 347 electoral votes, a key part of the story should be: It wasn’t supposed to be that easy.

The Washington Post spoke to veterans in the battleground state of North Carolina about the upcoming election and their choice for the next commander-in-chief. (Dalton Bennett/The Washington Post)