We know a majority of Americans plan to vote by mail or in person before Election Day, and at least 75 million have already cast ballots. This is an enormous uptick in votes cast early from 2016, and as a result, there’s a real possibility that we won’t know who won the presidency on election night.
Take Pennsylvania, arguably the most important swing state this election as it’s currently the likeliest “tipping-point state” in our forecast, or the state that could determine the winner of the Electoral College. About 2 million mail-in ballots have already been returned so far, almost one-third of the total number of votes cast in the state four years ago. But because the state can’t begin processing mail ballots until 7 a.m. on Election Day, it could be awhile before we get full results in Pennsylvania. State officials have said that it could take until Friday to finish counting most ballots. And between Tuesday and Friday, there could be a pretty big shift in terms of which party is favored in the vote tally, given that Biden supporters are far more likely to vote early or absentee than Trump supporters, who are far more likely to vote on Election Day.
How the media projects the winner of an election | FiveThirtyEight Politics Podcast
In other words, we need to brace ourselves for a “blue shift” in states like Pennsylvania. That is, states that primarily report Election Day results first could show Republicans with an initial lead on election night only to then shift toward Democrats as more mail ballots are counted. To be clear, this won’t be true everywhere. Some states, like Florida, are already counting early votes and will likely report a lot of its early in-person and mail votes ahead of many Election Day votes.
But it’s important to stress that this pattern — where the votes counted last shift the result in a Democratic direction — has existed in previous elections. Historically, this is in part because some Democratic-leaning voting blocs, such as young voters, are more likely to cast last-minute absentee votes as well as provisional ballots, which are counted later. But with the massive uptick in voting by mail and the strongly Democratic lean of mail votes — polls suggest Biden might win around three-fourths of Pennsylvania mail-in ballots — the blue shift could be far larger this year than in past years.
Suffice it to say, we should be ready for a sizable swing in the vote, and Pennsylvania is probably not the only battleground state where we will see such a shift. However, we’re not driving entirely blind: The results from Pennsylvania’s June 2 presidential primary give us some clues for how a blue shift might happen, and more importantly, what it might look like.
In that election, the more rural and Republican-leaning parts of the state reported a sizable majority of their results on election night, as their votes were disproportionately cast on Election Day. In the state’s more populous and Democratic-leaning areas, it took roughly two and a half weeks to report most of the vote, due to the huge influx of mail-in ballots. As a result, on election night, 3 a.m. Eastern, only 46 percent of the final Democratic presidential primary vote had been reported, compared to 71 percent of the GOP total, based on data from ABC News.
So what would this mean if something similar happened on election night in Pennsylvania? Well, if we use what our presidential forecast currently says about the race in Pennsylvania, we can game this out a little to understand what the vote count might look like on Nov. 3. We should be clear, though, that this analysis has its limits. We can’t know for sure how many Pennsylvanians will actually vote in November or the pace and trend of the count. Not to mention, counting the votes in two separate primaries is different from counting the votes in one presidential contest.
Right now, our forecast estimates that about 6.8 million votes will be cast in Pennsylvania, on average, and the average popular vote result gives Biden about a 5-point edge over Trump, 52 percent to 47 percent. If that panned out, that means Biden would win about 3.6 million votes to Trump’s 3.2 million (with a few additional votes for other candidates). And if the vote count followed what we saw in the primary, less than half of Biden’s votes would be in by 3 a.m. on election night, whereas around 70 percent of Trump’s would be reported.
That means we could be looking at a situation where Trump has about a 16-point lead, 58 percent to 42 percent, based on approximately 60 percent of the total expected vote. But over the course of the next few days — again, assuming the same pattern we observed in the primary — Biden would win two-thirds of the remaining votes, which would precipitate a 21-point shift in the overall margin from 3 a.m. on election night to the final result, as the chart below shows.
Now, it’s entirely possible the actual shift isn’t as large as the one outlined above. After the primary, local election officials in Pennsylvania now have some experience in handling a huge number of mail ballots. The state has also made investments to expedite the vote count. Plus, a sizable share of mail ballots will have already been turned in by Election Day, so administrators could have the majority of all mail votes when they begin processing them on Election Day morning. Pennsylvania officials also want to make sure that voters aren’t surprised by a sizable vote shift so plan to be as transparent as possible. Secretary of State Kathy Boockvar has said she will call on county election officials to regularly report the progress in counting mail ballots, instead of reporting them all at once, to avoid a huge swing in the vote that might cause confusion. And she’d also like the state to share how many ballots are left to count, where they are in the counting process and the party affiliation of the voters who cast those remaining votes.
Still, we’re talking about a lot of votes that have to be processed. Just 2.7 million people cast ballots in Pennsylvania’s presidential primary, about 40 percent of the total turnout we’re expecting in the general election. In fact, more people have requested mail ballots than voted in Pennsylvania’s presidential primary, period. That means even if officials are better prepared, they will still have far more mail ballots to process than in June. And while they can begin processing at 7 a.m. on Nov. 3, that still means verifying and processing each ballot, and only then counting each vote. This could take days, which could affect how quickly we know who won the election.
Will millions of dollars on TV ads make a difference in the race for the presidency?
And this brings us to one potential worry about a big blue shift, especially if the election is close: Trump could try to declare victory based on initial returns, as he’s claimed the result should be known on election night. Now, declaring victory doesn’t really amount to much, as it has no legal bearing. But doing so could help cast doubt on the results, especially if there is a subsequent shift toward Biden that’s accompanied by false claims of voting fraud, such as the ones the president has made numerous times already. Experts from organizations such as the Transition Integrity Project have even expressed concern that, in such a scenario, Trump could try to figure out ways to interfere with post-Election Day counting to stop a potential blue shift.
Even if Trump doesn’t question the result, Pennsylvania’s blue shift could be affected by post-election court rulings in the state. Take the fight over whether to count ballots that were postmarked by Election Day and received by Nov. 6. The U.S. Supreme Court has twice now upheld the state’s Supreme Court ruling that extended the deadline, but it’s not clear whether the court has entirely ruled out revisiting the issue. Justice Samuel Alito, joined by two of the court’s other conservative justices, wrote that while there was a “strong likelihood” that the state’s Supreme Court decision violated the U.S. Constitution, there simply wasn’t time to decide before the election. However, ballots arriving after 8 p.m. Eastern on Election Day will be separated out, so the issue can be revisited if necessary after the election. And if the Supreme Court does take up the case again after the election, deciding to not count these ballots could reduce the size of the vote shift in Biden’s favor because more Democratic votes would likely be affected, given the partisan split in mail voting. For his part, Trump said on Wednesday that he hoped the courts would stop states from counting ballots beyond Election Day.
It’s impossible to say just how much the vote may shift after election night in Pennsylvania, but the primary results in the Keystone State suggest that we should be ready for a sizable swing in the vote. And it’s probably not the only battleground state that will see a blue shift of some note. For instance, Wisconsin can’t begin processing mail ballots until Election Day, and Michigan’s count is expected to stretch on for a few days, too.
The better prepared Americans are for lengthy counts and major vote shifts because of the increase in mail-in ballots, the less likely they are to believe misinformation about what those swings mean.