Non-technical summary

Research question

Climate risks are a priority issue for many central banks and supervisory authorities due to the potentially detrimental effects of climate risks on banks’ stability. However, measuring climate risks is fraught with challenges, because nature and magnitude of the shock transmission are plagued with uncertainty. Moreover, data limitations are widespread, further complicating measurement. This paper proposes a novel stress testing framework to quantify climate-related transition risks for the German banking sector. Our stress test allows the identification of vulnerabilities stemming from transition risks for both individual banks and the banking system as a whole.

Contribution

This paper contributes to the literature by proposing a novel climate risk stress test framework, which allows to quantify climate transition risks for a large sample of German banks. We capture model uncertainty in a unique way by applying two standalone models (a micro and a macro model) for translating stress scenarios into credit risk parameters. Reporting ranges of estimates provides a quantitative indication of uncertainty in relation to the green transition. Another key strength of our framework is the granular modelling of default probabilities (PDs). While traditional stress testing models typically produce country- and/or portfolio-level estimates of scenario-dependent PDs, we can flexibly estimate them at firm- or sector-level. Through our targeted, multi-layered analytical framework, we thus address the heterogeneity of climate risks across economic sectors and firms.

Results

For a scenario that envisages an orderly transition to net zero emissions by 2050, our results show an average increase of up to 40 % in PDs for non-financial corporations after three years. In an alternative scenario assuming an abrupt increase in the carbon price to Euro 200, PDs rise to a similar extent. Our results indicate a great deal of heterogeneity between firms from different economic sectors: PDs increase more strongly in the agriculture, utilities and transport sectors, for example, and the impact on credit risk also differs between firms in the same economic sector. Whilst significantly stronger increases in PDs emerge for emissions-intensive firms, credit risk is not as high for firms with low emissions. Furthermore, we uncover heterogeneity in credit risk effects across banks as a result of bank-specific differences in the composition of loan portfolios.