Are Pennsylvania and North Carolina Really Pure Toss Up States?
Plus analysis of the latest polling data and the battle for Congress
Pennsylvania and North Carolina will be two of the three most important states in determining who wins the presidency (along with Georgia).
The polling margin in both states is razor thin — FiveThirtyEight’s averages have the contest even in North Carolina and have Kamala Harris ahead by two points in Pennsylvania.
Yet, history makes a case that the two states are actually mirrors: Pennsylvania is tantalizing for Republicans and can be won if things go just right, but usually falls a few points out of their grasp or more. The exact opposite is true in North Carolina: Democrats perpetually compete in the state and have high hopes, but narrowly lose far more often than not.
The chart above shows the share of the vote won by each party in gubernatorial, presidential, and U.S. Senate races in Pennsylvania since 2006. Only twice have Republicans topped 48.84% of the vote in 15 contests. And outside of the Republican wave election of 2010, they haven’t topped that number at all. By contrast, seven times in 15 contests, Democrats have topped 54% and only three times have they done worse than 48.84%.
The picture is unambiguous: Democrats have a much higher ceiling than do Republicans in the state. And while Republicans have increased their share of the vote in each of the last four presidential contests, and Donald Trump has done better than most Republicans, it still appears as if everything has to break just right for Republicans to eek out a narrow win.
When combined with the current polling, this augurs well for Harris (and Democratic Senator Bob Casey). The contest is likely to be close, because Trump is on the ballot and he draws out voters who don’t show up for other Republicans. Yet, there is a real question as to whether he can top 49% of the vote. Even if his share of the vote increased by the same margin as it did between 2016 and 2020, that would still only get him to 49.10 of the vote — which would require more of the vote to go to third party candidates than it did in 2020 for him to win. That’s possible given that there was no Green Party candidate on the ballot in 2020, whereas Jill Stein is on the ballot in 2024.
Overall, it’s clear that Pennsylvania is a state where Republicans can get to the high 40s, and in good years like 2010 and 2016, they can win. But the state is more a durable light blue than a total toss up. Additionally, if a candidate wins Pennsylvania by a large margin, it’s most likely to be a Democrat.
Some national reporters and commentators have characterized the state as a tough one for Harris based on internal polling and reporting. But the state’s history really doesn’t support this characterization. Instead, the public polling average seems about right.
North Carolina, by contrast, presents almost the opposite picture.
In 14 Senate, gubernatorial, and presidential elections since 2008 (there were none of those races on the ballot in 2006), Democrats have topped 49% four times — and three of them were in 2008, as Barack Obama became the first and only Democrat to carry North Carolina since Jimmy Carter in 1976.
Republicans, by contrast, have only received less than 48% of the vote in three of these 14 elections. Their floor is just higher — as is their ceiling. They’ve topped 50% five times in these 14 contests. Thus, while the state is typically only decided by a small amount, and Trump’s margin dropped between 2016 and 2020, it’s a state that has proved frustratingly elusive for Democrats.
And these trends in both states go back even further. I’ve only excluded the data because the liberal-to-moderate Republicans and conservative Democrats who ran in Pennsylvania through 2004 scramble the usefulness of the historical data. Nonetheless, Democrats have only won 4 of 18 U.S. Senate races in North Carolina since 1972, in addition to their poor showing in presidential contests (though they’ve done much better in gubernatorial races).
None of this means Harris can’t win North Carolina, or that Trump cannot win Pennsylvania. But if either happened it would signal one of two things: that the state might be shifting politically, or that everything had broken just right for that candidate.
It’s why I’d characterize Pennsylvania as leaning narrowly toward Harris, and North Carolina as leaning narrowly toward Trump — though the latter bears watching given the implosion by disastrous Republican gubernatorial candidate Mark Robinson. While Robinson isn’t likely to take Trump down with him, Trump’s effusive praise for the disgraced candidate could hurt at the margins, and Robinson’s spiraling campaign may give Democrats the benefit of a superior GOTV operation in the state…
One thing that is not receiving enough attention is the gap between polls with full size samples of “sub-populations” — young voters, Black voters, Latino voters, AAPI voters — and the data for those groups in larger polls.
A poll this week found Harris leading Trump by 38 points among Asian American voters — higher than Joe Biden’s 27% margin with this group in 2020. Similarly, the new Harvard Youth poll finds Harris beating Trump by 32%, better than Biden’s 24 point win with 18-29 year olds in 2020. Additionally, a Pew poll earlier in September found Harris leading with Latinos by 18%, a drop of only 7% from Biden’s 61-36 margin with this group in 2020 — far less slippage than some other polls show.
By contrast, in a New York Times/Sienna poll earlier in September that showed Trump ahead nationally, Harris only led among young voters by 8% (51-43), and only led with Latino voters by 14% (55-41) (the poll didn’t break out Asian American voters, but Harris only led the “other” racial category by 21 points, 54-33).
And this gap isn’t uncommon: as the veteran journalist Ron Brownstein has pointed out for months — going back to when Biden was the Democratic candidate — Harris has actually improved on Biden’s numbers with white voters and older voters. Trump is remaining competitive (or even ahead) in most polls because he’s making gains with non-white voters, and in some polls, with young voters.
But the gap between the full sample polls that focus on these subgroups — and have bigger samples — and other polling suggests that most polling may be underestimating Harris’s share with these subgroups. And there is also the possibility that polls are once again underestimating Trump’s support with white voters. That would offset any polling error with young voters or minority voters.
Yet, it’s something to watch — because it’s quite possible that if Harris truly does better with white voters and older voters than Biden did in 2020, she ends up winning fairly comfortably.
Keep reading with a 7-day free trial
Subscribe to The World According to Brian to keep reading this post and get 7 days of free access to the full post archives.