In the final days of the campaign, Donald Trump is experiencing a significant downturn, shifting what previously appeared to be a clear victory for the former president into a likely win for Kamala Harris. This perspective is shared by Thomas Miller, a data scientist at Northwestern University, whose proprietary model has accurately forecasted past elections.
This decline in Trump’s prospects represents a dramatic change from the election dynamics less than 13 days prior. In the early weeks of October, Trump made a notable recovery, moving from a significant deficit to a commanding lead. With less than two weeks remaining before Election Day, Trump seemed on track for a substantial victory.
Harris responded to Trump’s surge by redirecting her campaign focus from critiquing Trump’s policies to highlighting his “unstable” personality and “obsession with revenge.” However, this approach initially failed to resonate with voters, as Trump gained momentum in the electoral vote count, as predicted by Miller’s framework. In contrast, Trump successfully tapped into underlying dissatisfaction with the Biden administration, particularly on economic issues.
As Miller explains, while macroeconomic indicators like GDP and job growth looked positive and were frequently touted by Democrats, many people felt financial strain in their personal lives. Concerns included rising grocery costs, high mortgage rates preventing home purchases, unaffordable car loans, and the necessity of working multiple jobs without savings. Additionally, rising unease about America’s involvement in foreign conflicts, specifically the Biden administration’s policies regarding military aid to Israel and Ukraine, troubled a significant portion of the electorate. Trump, meanwhile, adopted a quasi-isolationist stance that seemed to appeal to voters.
According to Miller, the time available for Harris was too limited to bridge the gap by adopting new policy positions, changing her campaign rhetoric, or enhancing her ground game. He believed only a significant unforeseen event could alter the election’s trajectory.
Such an event occurred, not through the Harris campaign, but via a misstep from Trump’s team. The pivotal moment was the rally held on October 27 at Madison Square Garden in New York City, an event intended to be the pinnacle of Trump’s campaign efforts.
The rally featured around three dozen speakers but lacked top politicians and respected statesmen. Instead, it showcased controversial Trump loyalists and prominent figures from entertainment and right-wing media, such as Rudy Giuliani, Stephen Miller, Robert F. Kennedy Jr., and Elon Musk. Also present were media personalities like Tucker Carlson and entertainment figures such as Hulk Hogan, whose performance was a crowd favorite.
Controversial statements at the rally created a stir at a critical juncture for Trump. Comedian Tony Hinchcliffe’s derogatory comments about Puerto Rico and Latinos, along with businessman Grant Cardone’s inflammatory remarks about Harris, were particularly damaging. Trump himself contributed to the negative tone by referring to his opponent as “a low IQ individual.”
Miller’s analysis suggests that the rally, which Trump called a “love fest,” instead projected hostility, particularly towards women. This sentiment was echoed by Nikki Haley, Trump’s main female rival, who expressed concern about its impact on female voters, particularly those who are college-educated suburban women.
Following this incident, the Democrats capitalized on the fallout from the rally, promoting the missteps heavily in the media. Meanwhile, continued missteps by Trump’s supporters, including a vulgar video shared by a super PAC and derogatory remarks about Liz Cheney by Trump himself, compounded the damage.
Miller’s model, which relies on prediction or betting markets like PredictIt rather than polling, indicated a dramatic shift towards Harris that would have seemed improbable weeks earlier. On October 26, Trump led with 367 electoral votes to Harris’s 171, a significant margin. However, following the rally, Trump’s numbers began to decline sharply, and by October 29, Harris had gained 58 electoral votes. By the end of that week, Harris had overtaken Trump with further electoral gains.
Miller’s model, while acknowledging potential Republican bias in the prediction markets, now suggests that Harris could have a larger lead than projected, despite the close race reflected in the data. This shift aligns with the political trend of success favoring candidates who appeal to the center, a position Harris has embraced while Trump has moved towards the political extremes.
As the campaign draws to a close, Miller observes that while it initially seemed unlikely for Harris to stage such a rebound, the rapid turn of events has altered the election landscape. The remaining question is whether Trump can overcome the trend he set in motion before Election Day.
