Physical climate risk, credit risk and lending activity

We study how physical climate risk shapes bank lending activity and credit quality by combining high-resolution Copernicus flood geospatial maps with loan-level AnaCredit data. We exploit four major European floods (2021–2024) in a spatial regression discontinuity design comparing firms located just inside versus just outside flood boundaries (within 300–500 meters). We find that immediately after floods there is an increase by about 3.5 to 5% in lending, driven by liquidity demand, followed by a contraction of similar magnitude in the subsequent quarter. Interest rates follow a similar pattern, while default rates rise persistently by around 0.7 percentage points. Exploiting multiple lending relationships and firm–time fixed effects, we show that demand factors dominate: banks with greater exposure to affected firms do not systematically tighten credit supply. Nonetheless, relationship banks extend roughly 10 percentage points more credit to affected firms while imposing tighter collateral requirements, consistent with risk-sharing rather than unconditional support. Sectoral composition and pre-existing firm risk are the primary axes of heterogeneity in the immediate response. The findings shed light on how physical climate shocks propagate through credit markets and inform financial stability analysis.