Climate risks and banking performance
Loading...
Date
2025-09-16
Authors
Καραβασίλη, Αναστασία
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
This dissertation examines how climate conditions relate to banks’ profitability and asset quality in a multi-country setting. It assembles a harmonised annual country–year panel that combines standard banking outcomes—return on assets (ROA), return on equity (ROE), and non-performing loans (NPLs)—with balance-sheet indicators (bank capital–to–assets, domestic credit to the private sector), core macroeconomic controls (GDP per capita, inflation, unemployment), and two complementary measures of climate pressure. The first climate measure captures country vulnerability/readiness from the ND-GAIN framework; the second is a transition-pressure proxy defined as the log carbon-intensity of income, ln(CO2/capita)−ln(GDP/capita). Series are aligned by ID country code and calendar year and are used consistently across specifications; the panel is unbalanced due to gaps in reporting, and a locked estimation sample is employed to ensure like-for-like comparisons.
Methodologically, the analysis relies on Local Projections (Jordà) to trace dynamic responses of banking outcomes to climate shocks over medium-term horizons. The projections include country fixed effects, lagged outcomes and climate terms, and lagged banking/macro controls to mitigate reverse causality; standard errors are clustered at the country level. Lag length is selected by BIC from a small set, and horizons are set to capture multi-year adjustment. A static two-way fixed-effects regression is reported as a complementary robustness check.
The study addresses three questions: (i) how profitability (ROA/ROE) and asset quality (NPLs) respond over time to changes in climate pressure; (ii) whether responses are persistent or revert in the medium run; and (iii) how sensitive the results are to alternative climate measures and control sets. Across specifications, the impulse-response analysis delivers economically meaningful dynamics. In line with a risk-pricing/transition channel, higher climate pressure is generally associated with weaker bank profitability and a deterioration of asset quality over the medium term, although magnitudes depend on the chosen climate proxy and the conditioning set. Robustness checks—alternative lag orders, year effects, and restricted control blocks—do not overturn the main patterns.
Overall, the dissertation contributes a transparent LP framework for bank–climate interactions and a harmonised cross-country dataset that can support further work on policy design and supervisory assessment under climate risk.
Description
Keywords
Climate risk, Banking performance, Local projections, Panel data, NDGAIN, Carbon intensity, Profitability, NPLs