Mr. Bader, the impact of climate change on the financial market is extensive; where do you start with Climate Stress Testing?

There are typically two types of climate risks that we look at: physical risks and transition risks. Physical risks pertain to the immediate effects of climate change on the loan and investment portfolios of banks, such as those resulting from rising sea levels or severe weather events. Transition risks, on the other hand, emerge from the shift towards a low-carbon economy. These include, for instance, substantial expenses associated with the energy-efficient renovation of buildings and plants. With Climate Stress Testing, we have created a tool that enables banks to incorporate the impacts of climate-related risks into their risk assessment calculations.

Risk management has long been a fundamental practice for banks. What are the new challenges in assessing climate risks?

Conventional risk management practices in banking typically rely on the examination of historical data. The objective is to recognize patterns that suggest a heightened risk. One example: To evaluate the likelihood of a borrower defaulting, historical data on previous events is analyzed. For companies, this may involve examining the development of return on equity up to insolvency. This enables the identification of factors that may have contributed to this default. By analyzing a substantial number of portfolios, it is possible to ascertain threshold values for default probabilities.

However, when it comes to climate-related risks, the historical data record is insufficient. We currently lack a clear understanding of how the evolution of the CO2 price correlates with a company's profitability. So, we can only make assumptions about increasing costs. The actual response of companies to CO2-prices remains unclear. Currently, there is simply no evidence for this.

What strategies can be employed to address this lack of empirical data?

We rely on future scenarios rather than on data from the past. The Network for Greening the Financial System (NGFS) along with supervisory authorities provide models that present a range of credible future scenarios. These range from moderate physical and transition risks if global warming is limited, to high transition risks if timely action is not taken. An unregulated increase in temperature, resulting in significant physical hazards, is another potential scenario. We develop financial mathematical scenarios based on these models, enabling banks to assess default probabilities.

What are the use cases for climate stress tests in the banking business?

There are currently two main use cases: Firstly, in traditional risk management to ensure the bank's risk-bearing capacity. The supervisory authorities establish the regulatory framework for this matter, which has recently expanded to incorporate ESG risks within stress testing protocols.

Secondly, Climate Stress Testing is used to incorporate climate aspects into the assessment of credit risks within the portfolio.

A third use case is more future-oriented. The model could be utilized not only for assessing climate-related credit risks at the portfolio level but also in the lead-up to lending, ensuring that climate risks are originally integrated into the scoring system.

In contemporary discussions surrounding data-driven decision-making, artificial intelligence is often at the forefront of consideration. What role does the use of AI solutions play here?

The use of artificial intelligence presents significant opportunities for enhancing risk modeling. As already mentioned, currently, the primary obstacle in climate stress testing is the insufficient availability of substantial data necessary for effectively training AI models. The situation will evolve once the obligation for sustainability reporting is broadened, as this will greatly enhance the database within the ESG landscape.

The second difficulty is the strict regulations: While the supervisory authorities have recently granted explicit approval for the use of AI solutions in risk management, they simultaneously impose stringent requirements regarding the explainability and verifiability of the employed models. In practice, these can therefore only be used to a limited extent.

In addition to regulatory obligations, in what ways does managing these risks influence the resilience of companies during crises?

The calculations conducted by the NGFS model indicate that, irrespective of the climate scenario that unfolds, we should anticipate a modified risk situation in the markets. This may manifest as declines in property values in areas that are significantly impacted by climate change, or as increased financial pressures on businesses resulting from high costs associated with emissions.

Companies ought to take proactive measures to anticipate potential shifts in the procurement markets at this time. They should enhance their resilience by assessing possible physical risks throughout the supply chain and incorporating these considerations into their sourcing costs.

Consequently, it is essential not only for banks but all businesses to enhance their management of these risks. This specifically entails evaluating the impacts of climate change proactively, even in the absence of empirical data.