IE 11 is not supported. For an optimal experience visit our site on another browser.

Demographic shifts since 2016 could be enough to defeat Trump. But it's complicated.

Analysis: Trump's core demographic is declining. Our new interactive tool lets readers test how changes to the U.S. electorate might affect the election.
Image: Donald Trump and Joe Biden on a background of repeating United States map made of red and blue squares.
A new interactive collaboration by NBC News and The Cook Political Report finds that if 2016's rates of turnout and support were applied to 2020's new demographic realities, Trump would narrowly lose Florida, Michigan, Pennsylvania and Wisconsin.Chelsea Stahl / NBC News

Four years ago, President Donald Trump lost the popular vote by 2.9 million votes and still won the White House thanks to a near-perfect geographic vote distribution that allowed him to capture big Electoral College prizes by razor-thin margins.

The key? Trump's unprecedented 37-point margin among white voters without four-year college degrees, who are especially influential in the upper Midwest.

But as the U.S. becomes more diverse and college-educated, Trump's core demographic is steadily declining. In 2020, noncollege whites are on track to make up about 43 percent of the nation's adult citizens, down from 46 percent in 2016.

Meanwhile, whites with four-year degrees, who are trending blue and increasingly behave like a different ethnic group from noncollege whites, will make up 25 percent of adult citizens, up from 24 percent in 2016. And Black Americans, Latinos, Asians and other nonwhites, historically Democrats' most reliable supporters, will make up 32 percent, up from 30 percent four years ago.

A new interactive collaboration by NBC News and The Cook Political Report finds that if 2016's rates of turnout and support were applied to 2020's new demographic realities, Trump would narrowly lose Florida, Michigan, Pennsylvania and Wisconsin — more than enough to swing the presidency to Joe Biden. And, Trump would lose the popular vote by about four points, roughly double his 2016 deficit.

To let readers test their own assumptions about how these kinds of demographic shifts might affect the election, we’ve created an interactive tool that accompanies this article. To estimate the impact of changes in population, turnout and support on the Electoral College, you can use the tool to "swing the vote" and create your own November scenarios.

Right now, in the final stretch, Trump is doing everything he can to fire up his base, and he does have room to grow: In Michigan, Pennsylvania and Wisconsin, where Trump's combined margin of victory was just 77,444 votes, 4.9 million eligible noncollege whites didn't cast ballots in 2016. By contrast, only 1.6 million eligible nonwhites and 1 million eligible college-educated whites didn't vote.

But Trump might need to boost noncollege white turnout by about 5 points — from 55 percent to 60 percent nationally — just to offset the impact of their dwindling share of the electorate and get back to the same 306 electoral votes he won in 2016.

At the moment, Trump's bigger problem is that Biden is winning more noncollege whites than Hillary Clinton did in 2016. The latest NBC News/Wall Street Journal poll shows Biden losing them by 23 points, whereas exit polls showed Clinton losing them by 37 points. That would be more than enough to offset modest gains Trump has made since 2016 among Hispanics and other nonwhites.

Of course, race and education aren't the only prism through which to examine the changing American electorate. Breaking down the electorate by age, for example, reveals something different: Voters 65 and older, who narrowly supported Trump in 2016, have become a slightly larger slice of eligible voters since 2016 as more Baby Boomers have aged into that category.

However, age isn't as straightforward: To put it gently, plenty of the oldest 2016 voters have since exited the electorate. At the same time, many 18-22 year olds, who overwhelmingly dislike the president, have entered. And today, polls consistently show Biden doing several points better with seniors than Clinton, while Trump's support is concentrated among voters between the ages of 50 and 64.

Using the information above, you can start trying out the tool to see the impact of these changes. But to start, here are six possible scenarios of our own:

Six scenarios to swing the 2020 election

1. Biden makes a Sun Belt breakthrough

For years, Democrats have dreamed of breaking Republicans' grip on increasingly diverse and metropolitan Sun Belt states. In this scenario, Biden turns out enough nonwhites and wins over enough college-educated suburban whites to wrest Arizona, Florida, Georgia, North Carolina and Texas from Trump's column (in addition to Michigan, Pennsylvania and Wisconsin) — for a landslide victory.

2. Trump bounces back with seniors

Today, Biden is performing much better with voters 65 and older than Clinton did four years ago. But if Trump could find a way to get his margins with seniors back to 2016 levels, he could hang on to retiree-heavy battlegrounds like Arizona and Florida. In this scenario, Biden wins the popular vote by a full 3 percentage points but flips only Michigan and Pennsylvania, and Trump prevails by two electoral votes.

3. Biden squeaks by with white college graduates

In the 2018 midterms, Democrats' route to the House majority ran through high-income suburbs populated by white college graduates. If Biden were to replicate those gains, he could prevail, even if Trump matches or slightly exceeds his own 2016 numbers among other groups. In this scenario, Biden narrowly wins by flipping Arizona, Michigan, Pennsylvania, Wisconsin and Nebraska's 2nd District.

4. Biden coalesces the youngest and oldest

Today, Biden enjoys something of a "sandwich" coalition in polls: He's performing best with the youngest and oldest age groups. In this scenario, Biden roughly matches Clinton's share of the vote among voters under 65, but overtakes Trump with much stronger support from seniors. He wins by flipping the senior-heavy states of Arizona, Florida, Michigan, North Carolina, Pennsylvania and Wisconsin.

5. Trump prevails with inroads among nonwhites

One polling bright spot for Trump this year has been stronger support among nonwhite voters — particularly Latinos — than he received in 2016. If he were to combine those gains with 2016-level support among whites, he could actually improve on his showing. In this scenario, Trump holds all 30 states he won in 2016 and even flips Nevada red, narrowly winning the popular vote, too.

6. Biden rebuilds Obama's Midwest "blue wall"

In 2016, Democrats' "blue wall" of supposed Great Lakes strongholds crumbled. But Biden could resurrect the coalition he and former President Barack Obama built if he wins a higher share of white working-class votes than Clinton and restores Black turnout to Obama-era levels. In this scenario, Biden wins all 26 states Obama carried in 2012, including Iowa and Ohio, and picks up Arizona and North Carolina.

Notes: For purposes of this interactive, votes for independent and third-party candidates are counted as nonvotes in turnout figures. In 48 states, the winner of the popular vote gets all the state’s electoral votes. Maine and Nebraska award two electoral votes to the statewide winner and one electoral vote to the winner in each congressional district.

See the tool at