Banks lend less during heatwaves, Banque de France study finds

Heatwaves more likely to hit real estate, construction and services, while transport, leisure and mining hurt by chronic temperature rises as well

July 7, 2026|Written by
People play in a water fountain in front of the Eiffel Tower.

Photo by Chris Walts

Both long-term rises in temperature and acute heatwaves significantly cut loan growth for companies, although the impact varies across sectors, and policymakers should reflect this in climate risk frameworks, according to a new French study of bank lending.

The working paper was published by Banque de France (BdF) research director Oliver de Bandt and Skander Maraoui, an economist from Université Paris-Est Créteil. It examines the impact of chronic and acute heat on bank lending to French micro, small, and medium-sized enterprises (SMEs), using data from the BdF risk register.

“Both dimensions reduce the supply of bank credit but when we look at them in a distinct way we we can see that their effects on credit supply are different across loan maturities and sectors and even across time,” Maraoui said.

Increasing heat waves mean less bank lending

Acute heat shocks generate sharp but transitory declines in lending, persisting for up to five years, while chronic warming triggers a persistent decline in lending, lasting up to eight years, the paper says.

Almost half of the world’s population will be living with extreme heat by 2050 if global warming continues on its current path, according to a study from the University of Oxford. Research from the European Central Bank has found that extreme weather events could account for a loss of up to 5% of euro area economic output in the next five years, which would be on par with the great financial crisis. 

De Bandt said the study is unique in that it uses very granular data which shows different impacts in different French regions, such as ski resorts suffering from a lack of snow due to warming, which might limit access to credit in that area.

The study showed that sectors like transport, leisure, manufacturing and mining are affected by both types of temperature shocks, while acute heat waves are more likely to hit sectors like real estate, construction and services.

In some sectors, temperature risk leads banks to shorten maturities for loans rather than withdraw credit altogether, the study shows.

“Banks actually reallocate credit across maturity because they don’t stop lending to the firms but they prefer reducing long-term loans, reshuffling their portfolio exposure towards shorter maturity loads because of the uncertainty,” De Bandt said.

Banks not always considering long-term heat risks

A recent ECB⁠ report warns that while banks are improving their modelling of acute physical risks, they are lagging when it comes to modelling long-term chronic risks.

Maraoui said the BdF study shows that French lenders have already been taking acute and chronic temperature risk into account and adjusting their credit behaviour even before the ECB flagged the issue, making it a leader on this topic because many other central banks just focus on acute risks rather than chronic ones.

France has scored as a top leader in the Green Central Banking Scorecard for several years in a row, largely due to the extensive climate research from the BdF. 

 “The ECB went the right way to distinguish between average temperature rise and chronic,” he said, noting that the ECB issued a guide in 2020 on how it expects institutions to consider climate-related and environmental risks when formulating and implementing their business strategy and governance and risk management frameworks. 

De Bandt noted that standard bank credit assessment models do not usually capture the risk posed to rising temperature exposures, as credit ratings are based on historical data rather than forward-looking projections.

The paper calls on policymakers and supervisors to note the importance of climate-risk frameworks that are granular, sector-specific, and attentive to the maturity structure of lending.

“This will be useful for calibrating stress tests. You would need to take different scenarios and look at the differential effects,” De Bandt said.

This page was last updated July 6, 2026

Written by

Emma Thomasson is a British journalist, consultant and trainer based in Berlin. She is an expert in economics, politics, business and technology. She previously worked for Reuters as a correspondent and bureau chief in Germany, Switzerland, the Netherlands, South Africa and the UK.