Research Press Release

Energy: New model predicts when US fuel demand may recover from COVID-19-related lows

Nature Energy

July 17, 2020

United States gasoline demand is growing slowly since plummeting in April, as a result of COVID-19-related lockdowns, but may not make a full recovery before October 2020, according to a modelling study published in Nature Energy.

The US is one of the largest energy consumers and oil producers in the world, and it has had 25% of the reported global COVID-19 cases; public health policies to reduce mobility and ‘flatten the curve’ have led to a decline in US oil consumption. Previous research suggests that even after the pandemic peaks, there may be resurgences in cases until 2024, which may mean extended lockdown measures to control the spread of infection. Therefore, a model that can reliably forecast oil demand, a critical leading indicator of the state of the US economy, is necessary to manage business activities and investment decisions.

Shiqi Ou, Xin He and colleagues designed a machine learning model to project the US medium-term gasoline demand and study the impact of government interventions, such as lockdowns and travel bans. The model incorporates different pandemic projections as well as people’s resulting travel activities and fuel usage.

Using current infection data, the authors find that US gasoline demand has been growing slowly after a quick rebound in May, but is unlikely to fully recover before October 2020. Under a pessimistic infection scenario, continual lockdown could reduce motor gasoline demand temporarily, but could help demand recover to a normal level faster due to its ability to bring down the infection rate. A second wave of infections in mid-June to August could significantly lower the gasoline demand once more, but it would not be worse than it was in April. Under an optimistic infection scenario, gasoline demand could recover close to the non-pandemic levels by late September 2020. These findings add to the evidence that lockdown measures influence infection rate, and this in turn impacts mobility and fuel demand.

The authors note that with appropriate modification, this model can project other transportation-related fuel demand (such as for freight trucks) and project fuel demand changes in other countries in the near future. This model will be updated regularly with new data as the pandemic evolves. The results can be found at and


Return to research highlights

PrivacyMark System