How Harris Wins

If she does.

---

(Photo by Andy Manis/Getty Images)

As we enter the final few weeks of the campaign, some analysts have begun sensing a “vibe shift” in the state of the presidential contest. Whereas Kamala Harris was thought to be a slight favorite for much of the past month, the race appears to have tightened in the last week or two. From mid-August to late September, her lead in national polling hovered around 2.5 to three points. In the last few weeks, however, that has narrowed, and this week it fell to under two points.

The same thing has played out in key swing states. A month ago, Harris held slim but enduring leads in Michigan, Nevada, Pennsylvania, and Wisconsin—which collectively could put her over 270 electoral votes. But those leads have evaporated, and she is now neck-and-neck with Donald Trump in all four states. (North Carolina has been—and remains—in a dead heat as well, while Trump has held a slight edge in Arizona and Georgia for some time.)

These shifting dynamics have led some people to more seriously grapple with arguments for how Trump might win. One such examination of this came over the past weekend from Nate Silver, whose election forecast now shows the odds of either a Harris or Trump win at exactly 50/50.¹ My colleague, Ruy Teixeira, has also previously detailed what a Trump win might look like.

Still, there’s also a reasonable argument that these recent developments don’t actually signal much of a change in the grand scheme of things. Since Harris entered the race in late July, no major election forecast has given either candidate decisive odds of winning. This is likely because polling has been tight all along, both nationally and in the battleground states.

In other words: this remains a true toss-up election, which by definition means that a Harris win remains a very real possibility. So today, I want to take a look at how this could happen.

Gains with white voters

One of the less-realized ironies of the Trump years is that despite the racial unrest the country experienced, Americans have actually become increasingly less polarized along racial lines. Take white voters, traditionally a Republican-leaning group. Averages of post-election demographic data indicate that while Trump carried white voters in 2016 by 17 points, his margin with them actually declined from Romney’s in 2012, albeit by just one point.² Four years later, Trump’s advantage shrank by another four points to 13. Meanwhile, the reverse has occurred among non-white voters, who tend to overwhelmingly back Democrats: the party’s support among black (-9), Hispanic (-16), and Asian (-5) voters all narrowed to varying degrees between 2012 and 2020.

Democrats’ gains with white voters during that time were heavily based on stronger support from the college-educated contingent, with whom the party’s margin improved by 17 points. But it’s not just college-educated whites boosting Democrats. This cycle, as Harris appears to be gaining even more ground with them, she may also be doing better with non-college white voters than Biden. According to the latest polling crosstab averages from Cook Political Report, Harris is winning white college grads by 15 points, six points better than Biden did in 2020. Meanwhile, she is also only losing non-college whites by 26 points while Biden lost them by 30.

If these gains materialize on Election Day, this could be the ball game. Not only do white voters continue to make up a massive share of the electorate (72 percent in 2020), but non-college whites, specifically, are overrepresented in the “Blue Wall” states. Harris may be able to offset eroding support from non-white Americans by making gains—even small ones—with whites.

One way to test this is by using Cook’s new 2024 forecasting tool, which gauges various election outcomes depending on how different segments of the electorate vote and what their turnout looks like. If we adjust Cook’s 2020 baselines for voters based on race and education to match their 2024 polling averages—and, for the sake of simplicity, keep the turnout rates for each group the same—Harris would narrowly win the Electoral College, 276–262, carrying all three “Blue Wall” states and Nevada.³

Strong support from Asian Americans

Although Democrats have been losing ground with non-white voters in recent years—something we’ve documented thoroughly at TLP—this trend is less clear with one group: Asian voters. To be sure, there is some evidence that Democrats have struggled recently with them. In 2020, many majority-Asian neighborhoods across the country shifted toward Trump (though those swings may have been largely limited to working-class Asian areas, specifically). Then, in the 2022 midterms, Asians swung harder against the party than any other racial or ethnic sub-group.

However, on the whole, the party’s Asian support in presidential elections has pretty much held up since hitting a highwater mark in the Obama era, at least nationally. In 2012, Barack Obama won roughly 68 percent of these voters. Four years later, Clinton won 64 percent, and in 2020, Biden won 66 percent. While Asian Americans still make up a relatively small share of the vote—and, thus, are not broken out as a separate demographic group in most national surveys—they are a fast-growing population, including in swing states like Georgia, North Carolina, and Pennsylvania.

A recent national survey of Asian Americans from APIA Data found that this year, 66 percent favored Harris, matching Biden’s final margin with them in 2020, while only 28 percent supported Trump (four points lower than his 2020 share). This included a majority of independents (60 percent) and even one in five Asian Republicans. Indian Americans, specifically, who constitute the second-largest immigrant group in the U.S. and who may be energized about the prospect of the first half-Indian American president, supported Harris at a higher rate (69 percent) than any other sub-group. Similarly, a battleground state poll of Asian voters from the Democratic pollster GBAO found Harris’s support hitting 68 percent compared to just 20 percent for Trump.

One possible spot of concern for Harris, according to a recent Pew survey, may be that Asians overall are less motivated than other groups to vote this year. However, the GBAO poll found that, at least in the presidential battlegrounds, this may not be true—and that pro-Harris Asians in those states appear more enthusiastic than pro-Trump ones. Suffice it to say, given this group’s overwhelmingly Democratic lean, strong turnout from them this year could improve her chances in key states.

Sneaky (relative) strength with seniors

Older Americans have long been a sought-after constituency by both parties, as they constitute a large share of the electorate and historically turn out at high rates. Generally, seniors are a Republican-leaning group—no Democratic presidential nominee has won them since Al Gore in 2000. However, given the party’s advantage with younger voters, they don’t necessarily need to win older voters outright but rather just be competitive with them.

Here, there is some potentially encouraging news for Harris again. After seniors backed Trump over Clinton by eight points in 2016, Biden cut that advantage in half, losing them by only four. This cycle, Harris may be doing even better. According to Cook’s polling averages, she trails Trump with them by 1.4 points, which would mark a substantial improvement on Biden and potentially boost her in swing states whose populations are older than the national average like Michigan, Pennsylvania, and Wisconsin.

However, another demographic crosstab tracker from Democratic pollster Adam Carlson shows Harris losing ground with these voters from 2020. According to his averages for the first half of October, Trump was ahead with them by seven points, representing a 3.3-point rightward shift. And yet, other individual, high-quality polls show Harris actually holding a slight edge among seniors. The latest New York Times/Siena poll, for instance, has her ahead of Trump by two points.

Regardless of which one is correct, seniors’ outsized influence on American elections is clear. Harris may not ultimately win them outright, but if she even comes close to matching Biden’s margins, that probably bodes well for her.

Independents leaning toward Harris

Perhaps the biggest unknown of this election is how arguably the most important voting bloc—independents—will break. Since 1980, independent voters have backed the winning presidential candidate all but once (they narrowly favored John Kerry in 2004). Clearly, winning them is a crucial part of any candidate’s path to victory.

Both Carlson and Cook have shown Harris with a very small advantage among independents since early August. As of this week, she held a small lead over Trump with them of about one point. We should note that this constitutes a significant swing to the right from 2020, when Biden won these voters by 10 points. Still, Harris may not need a huge advantage with them to win. In 2012, Barack Obama only carried them by two points but went on to comfortably secure re-election. If Harris holds on, even narrowly, it could be enough.

A favorability advantage

Aside from demographic considerations, Harris is also viewed far more favorably than Trump. According to FiveThirtyEight’s poll trackers, roughly even shares of Americans have favorable and unfavorable impressions of her while Trump’s favorability has been mired in the low 40s all year and his net favorability currently sits at minus nine points.

Moreover, unfavorable attitudes toward Trump appear to be more intense than similar attitudes toward Harris. According to a recent Pew Research study, 38 percent of Americans say Trump would be a “terrible” president compared to 32 percent who say the same about Harris. Similarly, nearly half of Americans (48 percent) say Trump would “change the way things work in Washington…for the worse.” By comparison, just 41 percent say the same about Harris.

Of course, the close head-to-head polling between Harris and Trump suggests that a somewhat significant share of voters who do not hold favorable views of Trump are planning to vote for him anyway. But the disparity in favorability views isn’t nothing, either. Most campaign managers would surely prefer the public hold a positive image of their candidate than a negative one.

More reliable base voters

One of Trump’s biggest electoral vulnerabilities is the fact that as the two parties’ coalitions have shifted in recent years, Republicans have attracted a greater number of lower-propensity (or less reliable) voters while Democrats have brought in a greater share of higher-propensity voters. This means that Republicans may find it necessary to do more work to get their voters to the polls than Democrats do.

Political scientist Dan Hopkins examined election data from the last three general elections (2018, 2020, and 2022) and found that the voters who cast a ballot in all three were far likelier to back Democrats. This was true across racial lines. But as a voter’s engagement declined, the likelier they became to back Republicans. So essentially, Trump’s ability to turn his competitive polling into an actual win hinges on getting many of these infrequent voters to show up and vote.⁴

It’s certainly possible that he can accomplish this. However, the 2020 results show how even turning out less frequent voters can still leave him coming up short. Indeed, Trump attracted new voters that cycle and won the second-largest number of votes ever for a presidential nominee (behind only Biden)—but it still wasn’t enough to win a second term. That doesn’t mean he can’t succeed this time; he may, though, face stronger headwinds when it comes to hitting his vote goals than Harris does.

---

So this is probably Harris’s best case scenario. But ultimately, despite the ebbs and flows of this election cycle, there has been one, consistent story: it’s going to be a close race, one that likely comes down to fewer than 100,000 votes across three or four states. There are strong cases to make for either candidate winning. The best advice I can give is to ignore the noise in these final weeks and just wait for election week. We’ll have results to parse through before you know it.

1. In addition to Silver, three other forecasts have shown Trump overtake Harris recently by a slight margin, while one continues to show Harris as the slight favorite.

2. Data showing how various demographic groups voted comes from averages across four datasets: the Edison exit poll, AP VoteCast survey, Catalist “What Happened” report, and Pew validated voter analysis.

3. Cook’s 2024 averages do not include Asian voters, so I used the results of a recent APIA Data national survey of Asian Americans as the baseline for that input in Cook’s forecasting tool (more on that survey in the next section). Also, changes in turnout for each sub-group could obviously have a substantial impact here as well.

4. And so far, many Republican operatives don’t seem especially thrilled with the state of the Trump campaign’s ground game.