With 44 days until the U.S. presidential election, it may seem to many as if the death of Supreme Court Justice Ruth Bader Ginsburg has dramatically changed the political calculus—shifting the national conversation away from President Trump’s performance in office and refocusing voters on the battle over who, ultimately, gets to select Justice Ginsburg’s replacement. But a wealth of new polling data suggests that the fundamentals of this election are not likely to change—and particularly telling are a trio of Gallup survey questions that, taken together, have a strong track record for predicting past races in which sitting presidents vie for a second term.

The three key metrics—gauging presidential approval, U.S. satisfaction, and economic confidence—which were in the field from Aug. 31 to Sept. 13, suggest that President Trump has a very difficult road ahead to win reelection, no matter what happens with respect to the vacancy on the Supreme Court. “With those three measures,” says Jeffrey Jones, a senior editor and researcher at Gallup, “you can get a good sense of the lay of the land. Together, they usually offer a good way to figure out if an incumbent is going to win or lose.”

In the latest Gallup survey, 42% of Americans say they approve of the President’s performance in office and just 14% of Americans say they’re satisfied with the way things are going in the country. The firm assesses economic confidence by way of an index that combines two related measures—how Americans rate current economic conditions today and whether or not they think things are getting better or worse. In both cases, in the latest poll (ending Sept. 13), the majority of respondents’ answers are negative: 31% rate current conditions as “excellent” or “good” versus 34% who say they’re “poor”; and substantially more people say the economy is “getting worse” (56%) than “getting better” (40%).

Overall, Gallup’s Economic Confidence Index, which calculates responses on a range of negative 100 to positive 100, is at minus 10. While that’s up from last month’s reading of minus 16—and sharply higher than its April low, when the sudden fury of the coronavirus led to widespread business lockdowns and an historic rise in unemployment—it nonetheless represents a major challenge for the Trump campaign.

“It’s understandable—but also really shocking—how much [economic sentiment] has changed this year,” marvels Jones. “We didn’t see it move that much in all of Obama’s presidency. But this year we’ve seen it go from the highest level in 20 years (a rating of plus 40) to rock bottom (minus 33) and now to minus 10.”

The current rating on U.S. economic confidence is “not as bad as in 1992, when George H.W. Bush lost,” says Jones, but it’s notably worse than the outlooks faced by other incumbents who won a return trip to the White House: Bill Clinton in 1996, who enjoyed a plus-23 confidence score at this point in his reelection campaign; George W. Bush in 2004 (plus three), and Barack Obama in 2012 (minus one).

Satisfaction overall with the direction of the U.S., according to Gallup, is notably more pessimistic than it is with respect to the economy alone. The 14% share of Americans who say they are generally satisfied with the country’s direction is mostly unchanged from the 13% who said the same a month ago, which itself was the dreariest level of satisfaction measured since the financial crisis in 2008.

That, too, is well below the lowest point at which any presidential incumbent was given a second term—33% in 2012, when President Obama sought reelection and won. The current figure is even below the level in 1992, when George H.W. Bush went on to suffer defeat that November.

As for the most familiar measure of the trio, presidential approval, the ratings for President Trump remain sharply below the level that typically ensures reelection. In June of 2012, 46% of voters queried by Gallup approved of then–President Obama’s performance in office, a level that rose to 52% just before election day. In June 2004, President George W. Bush was at 49% approval, which slid to 48% in the final survey before the election. Those, so far, have been the lowest approval ratings for victorious incumbents in the modern era.

“I would say the numbers definitely don’t look good for the President being reelected,” says Jones. “He’s still a little bit above [Jimmy] Carter and [George H.W.] Bush on approval, but in a two-candidate race, with a 42% approval rating, where is he going to get those eight percentage points to get reelected?” Some of that will come from turnout, Jones says. “We know Republicans tend to vote at higher rates than Democrats, though maybe less so this year because Democrats seem very committed about voting. So that’s a lot of votes to make up.”

In previous election years, Gallup has found that a given president’s approval ratings this close to November has generally matched the share of the vote he ultimately receives.

Jones, who has been carefully watching polls and studying elections for the past 20 years, explains that the 2020 race is “just fundamentally an incumbent election.” An incumbent election and an open election are different choices for voters. “An incumbent election is a referendum on the president, and it’s almost irrelevant who is running against him,” Jones says. “The simple question is, ‘Are you for him or against him?’ An open election is, ‘Which of these two do I like better?’”

That’s why the President’s approval numbers are so revealing today. “If a person has been in office for four years and you haven’t come around to supporting him, it doesn’t make a lot of sense that you’ll vote for that person,” he says. “This election is about, ‘Does this this guy deserve four more years or not?’”

Trump’s approval rating in the latest Gallup survey—in which 42% now have a positive assessment of the President’s performance versus 56% who disapprove, a 14-point spread—is slightly more negative than the polling averages compiled by both RealClearPolitics and FiveThirtyEight. In the former composite, as of this writing, 44.8% approve versus 53.7% who disapprove, for a gap of 8.9 percentage points; in the latter, the findings are similar: 43.2% approval vs. 52.7% disapproval, a spread of 9.5 points.

Jones says one reason for that discrepancy is that, in the Gallup surveys above, the firm polls “all adults” rather than using a model that tries to assess likely voters. And, in fact, in FiveThirtyEight’s current calculation of polls using that same criteria, Trump’s average approval rating (41.6%) is similar to that in Gallup’s survey, which has a margin of error of plus or minus 4 percentage points.

Even if President Trump’s approval rating is actually closer to the high end of that margin of error, however, it would still put him under water come November—unless, say, something dramatic (he outduels Vice President Biden in the coming debates, the threat of COVID-19 suddenly wanes, or perhaps the economy bounces back sharply) happens over the next six weeks. Though Mr. Trump did win in 2016 with only 46.1% of the popular vote (compared with Hillary Clinton’s 48.2%), the circumstances then were quite different than they are today. “2016 was a choice election and many people didn’t really like either choice,” Jones says. “Because of that, you had higher votes for third-party candidates, which easily could have tipped the scales in the three states [Michigan, Wisconsin, and Pennsylvania] where the election basically came down to fewer than 100,000 votes. In addition, you had lower turnout among Democrats and people voting, but not voting for president in some cases. All of that can certainly make a difference when an election is as close as it was then. I’m not sure that those things are going to hold this year.”

“Polls are not laser precision instruments,” he cautions. “They can give you a pretty good sense of where things stand in a given moment, but can they always pinpoint it to one or two points? Can they call an election in which one candidate is going to win by one percent or less? That’s quite a challenge. If they are right, a lot of that is luck.”

But the gap right now is wide, at least by the limited gauge of history. Says Jones: “It just seems hard to imagine, being in the low 40s with no third-party candidate, that the President has a good shot at this point.”

This story was originally featured on Fortune.com