Democrats and Republicans have never been further apart in their expectations for the economy than they are right now.

The Trump presidency appears to have sparked an unprecedented political polarization in consumer confidence, which is often viewed as a leading indicator for the economy’s performance. While Democrats and Republicans have similar beliefs about how the economy is performing at the moment, they are now hugely divided over how it will perform in the future, new research shows.

“When we ask them about the economy … most Democrats expect a recession over the next year, and most Republicans expect economic growth,” says Richard Curtin, who directs the University of Michigan’s well-known survey. “That has been the big news: about how Democrats and Republicans shifted almost totally and instantly with the election of Trump.”

Perhaps surprising: That is a recent phenomenon. The research shows that Republicans were a little more optimistic than Democrats during the presidencies of Ronald Reagan and George W. Bush, although Curtin says the difference was not statistically significant. The gap was slightly larger under Barack Obama, with Democrats being more optimistic than their Republican counterparts. (The University of Michigan has collected this data only periodically, so it excludes the presidencies of George H.W. Bush and Bill Clinton.)

But with the election of Donald Trump, the partisan gap in consumer confidence split wide open, as the chart below shows.

Perhaps because Trump’s electoral win was unexpected by some, the data also show a sharp swing in the opinions of Democrats and Republicans before and after the election. As the chart below shows, Democrats quickly went from having more positive expectations for the economy than Republicans did to having a much dimmer view.

On average, the levels of confidence that Americans as a whole express in the economy are relatively high. After the presidential election, Michigan's measure of consumer confidence jumped to levels not seen since 2004.

Economists attribute this in part to expectations that the Trump administration would introduce tax cuts, infrastructure spending and other measures to buoy the economy — but they say the biggest factor was probably the continued strength of the economy, which recently qualified as the third-longest expansion on record.

Consumers have had plenty of reason to be happy: Unemployment is low, gas prices have come down, and the stock market is up, helping to buoy confidence.

But those post-election highs have ebbed in recent months. As former FBI director James B. Comey testified before the Senate on June 8, the University of Michigan saw its June 7-11 reading of consumer confidence retreat to 94.5, the lowest level since the election.

Few of the respondents that the University of Michigan surveyed mentioned Comey, but Curtin says the respondents probably were influenced by political events. "What they took from that incident was that Republicans thought that the probability of passing economic policies that Trump favors was much lower, and Democrats increased their dislike for those same policies," Curtin said.

Six months on from Inauguration Day, the division hasn’t narrowed, Curtin says. But it remains to be seen whether the advent of the partisan gap in how Americans view the economy is a fundamental change, or whether the gap will shrink as time goes on.

Curtin says he believes the gap might have arisen because Americans are now looking more to the government to help fix economic problems, like rising inequality and lower economic growth. But more polarized media coverage of the economy might also help generate these divisions.

It’s also still unclear what the impact on the real economy will be. It seems logical that Democrats who expect the economy to falter will behave differently than Republicans who expect it to boom. But Curtin and other economists debate how much consumer confidence spills over into real economic behavior.

Curtin says the impact on the economy is muffled somewhat because independents still remain the largest political grouping in the United States. Meanwhile, the optimism of Republicans helps to cancel out the pessimism of Democrats overall, he said.

Jed Kolko, an economist at job site Indeed who closely follows regional economic trends, says that states that voted for Trump have been growing faster than states that voted for Clinton in the first four months of the Trump administration, but those red states were already seeing faster job growth in the last months of the Obama administration.

“In fact, job growth has been slower in the first few months under Trump than in the last few months under Obama, in both red and blue states. But we'll continue to watch whether greater confidence might lead to accelerated growth in red states in the future,” Kolko said.

The research is not without critics. Ken Goldstein, an economist with the Conference Board, said he respected Curtin deeply and “would not be quick to challenge his assertion. But I think for most consumers it’s about my job, my paycheck, my chance for a raise, my chance for a promotion.”

“What happens in the labor market is far more important than what happens in the political arena,” Goldstein said.

It's a common observation among economists, who often argue that basic economic factors have much more influence over consumer behavior than events in Washington.

In a news conference in June, Fed chair Janet Yellen remarked that while businesses and household sentiment remained strong, these groups appeared to have a wait-and-see attitude when it came to political promises.

“I would say that based on my observation of actual spending behavior and my discussions with our wide range of contacts that I haven't seen very much evidence that thus far expectations of policy changes have driven substantial changes in either consumer spending or investment spending,” she said.

Josh Feinman, chief economist at Deutsche Asset Management, agreed, adding that a lot of trends in the economy transcend the political process.

“Politicians almost invariably give themselves more credit” for the economy than they deserve, Feinman said. “And, frankly, the other side almost always assesses more blame to government and policymakers than is really the case.”