The third candidate in this election is the economy

But how the economy might influence the results is less than straightforward. On the one hand, it has been growing steadily, generating millions of new jobs and pushing wages higher. On the other, prices are sharply higher than when President Biden took office and housing is less affordable, and those factors are weighing on Americans’ moods.

Harris hasn’t put as much emphasis on the jobs and growth figures as Biden did, instead focusing her message on what she calls an “opportunity economy." Trump has said he would push through an array of tax cuts and tariffs, while painting a darker picture of the economy. Over the weekend he warned that a Harris win would lead to “a 1929-style economic depression."

Americans have given the economy low marks during the Biden administration, driven by frustration over prices. Here, too, however, a sharp partisan divide comes into play. Consumer surveys from the University of Michigan show Republicans rate the economy as worse than even when the pandemic hit in the last year of Trump’s presidency. Democrats rate it as better than they ever did during Trump’s time in office.

Despite such downbeat assessments, what people have been doing with their wallets tells a different story. Last week, the Commerce Department reported that, adjusted for inflation, consumer spending was up 3% in the third quarter from a year earlier. During the first three years of the Trump administration, before the pandemic, spending grew at a 2.6% annual rate.

The interplay between robust growth and high prices might be a major reason why polls show the election as a tossup. This is also on view in a long-running presidential prediction model that Yale University economist Ray Fair first developed in the 1970s. Combing through data going back to the early 1900s, he has found that three economic variables did a good job of predicting the presidential vote.

The first is the per capita growth rate of real, or inflation-adjusted, gross domestic product in the three quarters before the election. The more growth the economy has delivered over the election year, the better it is for the incumbent party’s candidate.

GDP per capita grew at a 2% rate in the first three quarters of this year, the best pace going into an election since former President George W. Bush won his second term in 2004. GDP per capita grew at an even better 2.8% rate during Biden’s entire presidency through the third quarter—the best since former President Lyndon Johnson—but Fair’s work suggests voters focus most on the economy’s recent performance.

GDP per capita’s power in predicting the vote is a reflection of how, when it is growing strongly, other things such as job growth and wages are also usually doing well.

But Fair’s work shows that voters appear to have a longer memory for inflation than they do for economic growth. The second variable in his model is changes in a measure of inflation called the GDP price index. Fair has found that it is the price changes over the entire presidential term that matter. The more prices have risen, the worse it is for the incumbent party. The GDP price index expanded at a 4.5% annual rate in the first 15 quarters of Biden’s presidency, the fastest pace since former President Ronald Reagan’s first term.

Voters’ long memory on inflation has been on display this year. Anger over the sharp price increases that occurred earlier in Biden’s presidency has persisted even though inflation has cooled. And it has persisted even though economic analyses of Labor Department data show that most workers’ wages have risen more than prices did.

The final variable in Fair’s model is what he calls “good news quarters"—the number of quarters during the presidential term that GDP per capita growth exceeded 3.2%. There have been four of those during the Biden administration. There were three for Trump ahead of the 2020 election.

Based on the three economic variables, plus some noneconomic measures such as how long the president’s party has been in power, Fair’s model predicts that Harris will garner 49.5%, and Trump 50.5%, of the two-party vote share. In other words, just like what polling averages show—a virtually even race.

That doesn’t mean the actual results will look anything like a tossup. Fair’s model doesn’t include any of the noneconomic issues that can motivate voters, or how effective campaigns have been in their efforts.

Indeed, the model generated larger-than-usual errors in the past two presidential elections, though in the opposite direction of the polls. It predicted that Trump would capture most of 2016’s two-party vote, instead of Democratic nominee Hillary Clinton (who nevertheless lost the Electoral College vote). And it showed Trump winning the vote share in 2020, whereas polls predicted that Biden would win more handily than he actually did.

Plus, the snapshot of the economy people take with them into the voting booths can always change in ways that no model can pick up. Friday’s employment report showed the U.S. gained a paltry 12,000 jobs last month. That figure was heavily distorted by job losses because of Hurricanes Helene and Milton, as well as the Boeing strike. But people who read just the headlines might feel a bit more dour.

Or they might fill up their car Tuesday and come to a different conclusion. A gallon of regular gasoline averaged $3.10 a gallon on Monday, according to AAA, compared with $3.43 a year ago and a June 2022 peak of $5.02. Gasoline heavily influences people’s views of inflation: It is a frequent purchase, and the price is displayed on towering signs.

Whether it is Harris who wins or Trump, it will be possible to point to the economy and truthfully say, “That is why." But it won’t be the only reason, either.

Write to Justin Lahart at Justin.Lahart@wsj.com