The model by Oxford Economics uses unemployment, disposable income and inflation to forecast election results to predict the elections outcome.
According to the model, Mr Trump will lose in a landslide, capturing just 35 per cent of the popular vote, according to a report by CNN.
The model has predicted the winner of the popular vote in 16 of the past 18 elections and is a complete reversal of what the model was predicting before the coronavirus outbreak hit the US.
Before the public health crisis, Oxford Economics predicted that Mr Trump would win about 55 per cent of the vote, CNN reported.
“It would take nothing short of an economic miracle for pocketbooks to favour Trump,” Oxford Economics wrote in the report, according to the broadcaster.
It added that the economy will be a “nearly insurmountable obstacle for Trump come November”.
The forecaster said that the downturn has emerged in light of the poor economic backdrop the country now faces, and the model assumes that this would still be the case by the time of the election.
“The economy would still be in a worse state than at the depth of the Great Depression,” Oxford Economics said.
The poor economic backdrop means our state-based & national election models both suggest President Trump will lose the popular vote. But while economic factors point to a clear Democrat win, voter turnout and the evolution of the pandemic could be crucial https://t.co/0OhzZgSOtj pic.twitter.com/H64pV4ymfL— Oxford Economics (@OxfordEconomics)May 20, 2020
Experts have also said that the outcome of the election could still very much depend on how the pandemic pans out in the next six months.
“If new infections really pick up, people will conclude Trump opened the country too soon,” Greg Valliere, chief US policy strategist at AGF Investments told CNN. “But if new infections drop, Trump will get some credit.”
Experts also note that the six months run up to polling day could give Mr Trump time to reframe his campaign against Joe Biden and pass the blame of coronavirus onto China.
Voter turnout could also swing the election one way or another, according to the report.
The model is said to have correctly predicted the popular vote in every election since 1948, other than 1968 and 1976.
However, researchers admit it has “inherent limitations”, excluding non-economic factors such as a candidate’s agenda or likeability.
“Traditional models work in normal times. But we’re not in normal times right now,” Mr Valliere said.