Climate Risk Stress Tests Underestimate Potential Financial Sector Losses

Climate stress testing is a developing field and researchers say it should incorporate “severe but plausible” scenarios such as green swan and Minsky-type events to improve understanding of tail risks. In this column for the Centre for Economic Policy Research, the authors say current central bank modelling may lead to significant underestimation of losses, and they identify various avenues for improving stress testing.

Central banks increasingly rely on climate stress tests (CST) to assess the potential and most detrimental effects of climate change. Stress tests are forward looking, making them useful for assessing the unprecedented impacts of climate change. Established applications of CSTs commonly use scenarios developed by the Network for Greening the Financial System, including orderly and disorderly transition, and hothouse world scenarios. There is also a growing strand of literature testing the effects of natural disasters on financial institutions.

Less attention has been given to the impact of “plausible but severe” climate events. The occurrence or increased understanding of such events can cause sudden changes in the perception of climate risk by financial actors, leading to shocks to asset values. The authors – Henk Jan Reinders, Dirk Schoenmaker and Mathijs van Dijk – consider two such shocks.

Firstly, shocks to real asset values could be triggered by changing perception of the state of the global climate system, including the risk posed by green swan events. For instance, if climate tipping points are reached sooner than previously thought, or if scientific insights regarding the timelines of those tipping points suddenly change.

Green swan events are almost impossible to predict with accuracy but are certain to occur and result in unprecedented and high-impact physical and economic damage if critical climate tipping points are passed.

Secondly, shocks to the prices of financial instruments could emanate from the financial sector if financial agents abruptly incorporate the latest climate science after continuously failing to do so. This could be caused by strongly improved climate risk data or increased awareness following a large natural disaster. The authors label this a “Minsky-type” shock as it originates from the realisation that persistent regulatory blind spots exposed the economy to systemic risk and potential collapse.

The authors discuss other drawbacks of existing central bank testing, including limited scope, over-reliance on macro-financial modelling, and low granularity. This is particularly significant as climate-related shocks often have specific sectoral and regional impacts, increasing the need for disaggregated, micro-based modelling approaches.

To address these shortcomings, the authors say CSTs should be developed to improve their use of microeconomic and disaster risk approaches. Microeconomic approaches assess the impacts of shocks on specific economic sectors and regions, while disaster risk tests link disaster risk models to financial sector outcomes.

Additionally, the authors say stress tests should consider relevant endogenous feedback loops which can amplify initial shocks between climate, economic and financial systems. The most important of these are intra-financial, macro-financial and climate-economic feedback.