Physical climate risk raises bank credit risk, alters lending
A new ECB working paper finds that physical climate risk from floods persistently increases bank credit risk and temporarily alters lending activity. The study combines high-resolution flood maps with granular loan data to analyze four major European flood events.
Floods reshape credit dynamics
The study documents pronounced but temporary dynamics in lending volumes.
Loans to affected firms initially rise by 3.5 to 5 percent in the flood quarter, driven by immediate liquidity demand.
This is followed by a contraction of similar magnitude in the subsequent quarter as acute needs subside.
Interest rates follow a comparable pattern, showing a modest upward shift post-flood and a partial reversal.
Crucially, default rates on pre-existing loans persistently increase by approximately 0.7 percentage points, nearly doubling the average annual baseline rate.
This indicates that physical climate shocks can materially impair loan quality even when aggregate credit supply remains stable.
Demand-side factors primarily drive these lending dynamics, with limited evidence of systematic credit tightening by banks.
Mapping risk with satellite data
The research combines high-resolution Copernicus flood geospatial maps with the European Central Bank's (ECB) granular AnaCredit loan-level data.
It exploits four major European floods between 2021 and 2024, including catastrophic events in Germany, Belgium, the Netherlands, Italy, and Spain.
A spatial regression discontinuity design compares firms located just inside versus just outside flood boundaries (within 300-500 meters).
This quasi-experimental setup allows for isolating the causal effect of flood exposure, controlling for similar local economic conditions and pre-event characteristics.
The authors note that this methodology addresses challenges in identifying causal effects by treating flood assignment within narrow bands as effectively random.
Local shocks, systemic lessons
This study provides crucial causal evidence on how physical climate shocks impact credit markets, filling a significant gap in policy-relevant research.
While confirming material impacts on credit quality and localized lending, the findings suggest these shocks are local, not systemic, for resilient banks.
Granular data integration offers a pragmatic path for targeted supervisory frameworks and climate risk assessment.