According to the latest forecasting model from Decision Desk HQ and The Hill, the presidential election is the closest this century, and the stakes are incredibly high for next week's first debate between former President Trump and Vice President Harris.
The Decision Desk HQ model gives Harris a 55 percent chance of winning, but suggests the race will essentially be a jump ball contest, with seven states at stake for both sides.
This makes Tuesday's debate even more important, as it could completely change the course of the battle depending on whether either Harris or Trump performs well or suffers a crushing defeat.
“This debate is going to be just as important as the one between Biden and Trump,” Scott Tranter of Decision Desk said in an interview.
Tranter said the numbers could fluctuate because the debate showdown is highly anticipated, and early voting is set to begin in some states, so voters may end up filling out their ballots immediately after the debate.
While other models vary slightly (Nate Silver's model, for example, gives Trump a 60 percent chance of winning), the individual predictions don't vary that much.
At first glance, the roughly 10-point difference between Decision Desk HQ's and Silver's models may seem like a big difference in indicating one candidate is more likely to win than the other, and Silver's model has projected that Trump's chances of victory have increased in recent days.
But both models show how close the race is, with either candidate likely to win.
“Statistically speaking, Silver's model is not an 'outlier,' and it is understandable that the public would view his model as significantly different. In fact, Silver gives Trump a very slight edge, while we give Harris a very slight edge,” Tranter said in a memo explaining the model to Decision Desk HQ/The Hill.
“There are many mathematical reasons for this, [it] “This is a sign of what the situation may look like as we head into the final weeks of the week,” he added.
Tranter said Silver's model includes momentum and other factors that Decision Desk HQ's model does not include. Polling adjustments It was taken during or shortly after the Democratic Convention. Has some elasticity Such provisions are often made for candidates at party conventions, and this is no different in Harris' case.
“We both grill pretty good steaks,” Tranter said, “and he has some ingredients that we don't use.”
silver He pointed out in his update Starting Thursday, several post-Labor Day polls have Harris performing well, leading some to expect fewer adjustments to models of her uptick at the convention than polls have shown over the past few weeks.
Harris is leading Trump by about 4 points in The Hill/Decision Desk national average, but the lead is narrowing or they're tied in seven key battleground states.
The Democratic candidate's lead is largest in Wisconsin, at about 3.5 points, the only state where both candidates hold leads of more than 2 points, in Michigan they have about 1.5 points, and in Pennsylvania, Nevada and Georgia they have leads of less than 1 point.
In Arizona and North Carolina, the candidates received roughly the same number of votes.
The 21st century has been characterized by close elections decided by just a few states and tiny margins within those states. The largest candidate victory of the century was in 2008, when former President Barack Obama defeated the late Sen. John McCain (R-Ariz.) by about seven points.
There hasn't been a true landslide victory in an election since 1988.
But 2024 may be the closest yet. Tranter wrote in the memo that the race is the “closest” this century, possibly since John F. Kennedy's victory over Richard Nixon in 1960 or Harry Truman's victory over Thomas Dewey in 1948.
The 1960 election was the closest of the 20th century, with Kennedy winning the popular vote by just 0.2 percentage points. Conquered three states The victory was decided by just one point.
Although Truman won the popular vote in 1948 by a few points, the two states that helped him win were Ohio and California. By just a few thousand votes.
This could also be the case, as Tranter said the winner would likely only receive around 300 electoral votes, just over the required 270-vote threshold.
He said the 2020 election model was “pretty clear” before Election Day, predicting that now-President Biden had an 80% chance of winning. In 2016, FiveThirtyEight's model clearly predicted that Hillary Clinton had a favored odds ahead of Trump.
Looking at 2024, “all the good forecasters are right about 50% of the time, with a margin of error of a few percentage points,” Tranter said.
So even Silver's model, which gives Harris a roughly 40 percent chance of winning, isn't a bad probability compared to other models.
Party conventions often produce some, even if small and temporary, improvement in a candidate's approval ratings, but Harris's boost after the Democratic Convention two weeks ago appears to be minimal.
Harris' lead in the Decision Desk average of national polls is only a tenth of a percentage point larger than it was before the convention began on August 19. She is roughly equal to Trump in Arizona, Michigan, Pennsylvania and Wisconsin, but has grown by a percentage point or two in Georgia and Nevada.
Before the competition began, she held an average lead of one point in North Carolina, but the candidates are now tied.
Generally speaking, candidates this century have seen smaller increases in popularity after their conventions compared to previous elections in the late 20th century. Minimal voting improvements They will do so after the 2020 tournament, Silver said.
Tranter said an argument could be made that Harris' surge came after she entered the race in late July and early August, a few weeks that allowed her to make up Biden's large lead nationally and in key states, making the race competitive again.
“It's never been the case that we thought we were going to go to a convention with one candidate and then we had another candidate,” Tranter said.
Correspondingly, however, her lack of an increase in approval ratings does not necessarily mean that her approval ratings will decline over time after the convention.
Just because the race is close at the moment doesn't mean it will remain close over the next two months.
Tuesday's debate in the key state of Pennsylvania could be the next game-changer.
