Climate risk: how do banks measure it?

In the first blog on climate risk, we discussed that climate change is an important factor for banks. First of all, physical risks such as flooding may lead to damage to premises and products, which in turn leads to losses. Also, transitions risks arise because policies, rules and regulations will be adjusted, aiming to limit climate change. This is also recognized by regulators that have issued requirements (e.g. SFDR) that force banks to report on ESG measures.

In this second blog, we assess how banks measure climate risk. For this assessment, we will make use of a climate risk matrix that divides climate risk impact into 8 categories. We will look at examples of various banks, to make clear what the current market practice is for measuring climate risk.

Measuring climate risk

As described in our first blog on climate risk, climate risk is often divided in two types of risks:

1. Physical risks: natural phenomena such as drought, extremely high or low temperatures, perspiration, storms and flooding. Such events can result in damages to premises and products, thus resulting in losses.
2. Transition risks: risks borne by financial institutions in relation to changing policies, rules and regulations which aim to limit climate change.

These two types of risks can be combined with the conventional risk types ‘credit’, ‘market’, ‘operational’ and ‘other’, leading to the following overview:

Figure 1 An overview of examples of how climate change can be a driver of conventional risk types (source: DNB)

In this overview, it becomes clear how climate risk affects conventional risk types. Via the channels ‘physical’ and ‘transition’ climate risk, the 4 conventional risk types are affected. The question is of course, whether and how banks measure this impact. Without claim to completeness, we will look at examples of the Dutch banks ING, De Volksbank, Rabobank, ABN AMRO and also at HSBC.

ING

Finding results of ING’s climate risk assessments is easy. There is a clear “Climate risk report 2020” that elaborates on how ING deals with climate risk. There are three main approaches:

1. Scenario analysis: four long-term scenarios are identified, see Figure 2. It is interesting to see that ING focused on the extreme scenarios “Wait and See” and “Fast Forward”, because maybe the “Inefficiency scenario” is a more challenging one. This scenario planning exercise helps ING to better understand the drivers of the energy transition, such as the transition risk in energy-intensive sectors.
2. Stress testing of ING’s portfolio: using stress testing, transition and physical risks have been assessed. Whereas transition risk has been assessed on a global scale, for physical risk only the Dutch mortgage portfolio has been taken into account (e.g. flooding).
3. Transition risk heatmap: ING investigated the effect of transition risk across sectors in the form of a heatmap, based on the “Fast Forward” scenario. The aim is to anticipate how sectors are responding to steering to a low-carbon economy under a specific climate scenario as well as the identification of potential hotspots, see figure 3 for the main results.

Figure 2 Four scenarios identified for ING's scenario analysis.

The focus for all activities related to physical risk is clearly “property”. Both the residential and commercial property portfolios are assessed. It should be clear that when mapping this scenario analysis to the matrix presented in figure 1, it falls within the Transition Credit Risk category.

Although some activities such as the transition risk heatmap are broader, also within this category ING focuses mainly on property. For example, ING foresees policy moving towards favoring, or even excluding, certain energy labels. In this case, customer preference could shift towards the ‘better’ energy labels (or at least energy labels in line with the policy). Increased demand for these labels could in turn push up the property price. While at the same time there is a risk that properties with lower energy labels will decrease in price and value. However, there is not yet any concrete approach to assess this risk. Also here, the focus is on credit risk, implying that this assessment can be categorized as Physical Credit Risk.

Figure 3 Outcome of ING’s transition risk heatmapping exercise.

De Volksbank

De Volksbank has performed a climate risk scenario analysis and stress test to gain more insight in the impact of sustainability opportunities and risks, in particular climate risk. In 2019, a high-over qualitative climate scenario analysis was performed, but in 2020 a more structural and quantitative data-driven approach was developed, consisting of:

1. Impact analysis: an impact analysis to assess the direct impact of physical risks.
2. Scenario analysis:

A: One scenario based on the macroeconomic impact after a climate event materializes.

B: A second scenario focusing on the urge to accelerate the energy transition.

Figure 4 Effects of floods and pile rot in the CAS (Climate Adaptation Services) scenarios used by De Volksbank.

In the impact assessment, the effects of floods and droughts are analyzed for the residential mortgage portfolio, as it is the most prominent asset on the balance sheet. Impact is calculated on provisions, expected losses and Risk Exposure Amount (REA). This exercise can be categorized as Physical Credit Risk.

The first scenario is related to the macroeconomic impact after a climate event materializes. For example a drop in GDP is foreseen. This can be categorized as Physical Market Risk, although the approach will be very similar to ordinary stress tests, where also a decrease in GDP will be taken into account.

The second scenario relates to transitional risk, for example additional taxes for CO2 emissions or lower ECB interest rates to stimulate sustainable investments. From the description it is not really clear if this is limited to credit or market risk, or a more general transition risk stress test.

Rabobank

Similarly to ING, Rabobank has published a 2020 report on their approach to limit climate change. Furthermore, being an active member of the UN Environment Finance Initiative (UNEP FI) working group, Rabobank was actively involved in the development of a scenario analysis methodology for assessing the risks and opportunities associated with the transition to a low-carbon economy. The methodology allows for the risk assessment of transition-related exposures in corporate loan portfolios. A blend between sector-level (top down) and borrower-level (bottom up) modelling techniques is used to convert pre-defined climate scenarios into conditional Expected Loss (EL) estimates. Within the framework Probability of Default (PD) and Loss Given Default (LGD) estimates are to be calculated conditional on a climate scenario. Transition risk is then measured as the difference between a base estimate of EL (excluding transition risk) and the conditional EL. More details on the methodology can be found in the report Extending Our Horizons. It should be clear that when mapping this type of scenario analysis to the matrix presented in figure 1, it falls within the Transition Credit Risk category.

ABN AMRO

An example of Physical Credit Risk is given by ABN AMRO’s study where the economic impact of future flooding in the Netherlands is assessed. The study considers the probability of a 200cm and 50cm flood within the period between 2020-2050 and evaluates how these events impact macroeconomic indicators, such as GDP, unemployment and the house price index. The study is part of a wider range of climate related stress tests with the goal to estimate the impact of physical risks from climate change to the mortgage portfolio.

HSBC

In their 2020 TCFD^[1] update HSBC provides insights from a pilot scenario analysis of corporate and retail customers. The goal of this pilot was twofold. Firstly, the pilot develops foundations for climate financial risk stress testing capabilities. Secondly, the pilot makes a preliminary identification of the most material drivers. In their methodology HSBC makes use of the three types of future scenarios as defined by the Network of Central Banks and Supervisors for Greening the Financial System (NGFS). The NGFS distinguishes three representative scenarios;

• Orderly: Early, ambitious action to a net zero CO2 emissions economy;
• Disorderly: Action that is late, disruptive , sudden and / or unanticipated;
• Hot house world: Limited action leads to a hot house world.

Under all three scenarios both physical and transition risks are considered simultaneously. Furthermore, the methodology of HSBC considers a 30 year horizon and makes simplifying assumptions of a static balance sheet and stable GDP growth. Due to data availability the test is only conducted on a sample of the full portfolio, which they admit is not representative for the entire portfolio of assets. Within the matrix depicted in figure 1, the pilot scenario analysis of HSBC can be categorized as Physical and Transition Credit Risk.

HSBC also tested the resilience of its infrastructure, including global data centers, office buildings, trading floors and service centers. In an undisclosed six-step modelling process all major climate-induced threats at specific locations and buildings was covered. This is a clear example of Physical Operational Risk.

Conclusion

From the above it can be concluded that banks are already actively setting up the foundations for quantitative climate-related risk management. However, at the same time it becomes apparent that most activities are of exploratory nature and that quantitative climate risk modelling is still in its infancy. The focus is still very much on analyzing physical credit risks. It would be good to enlarge the scope of climate risk stress testing to all 8 categories mentioned in the DNB overview. Also, we suggest banks to not only describe the methodology, but also disclose more about the actual inputs and results of the climate risk assessment. This will improve the transparency and comparability between banks and lead to better climate risk assessments.

Of course, it is essential to not only measure climate risk, but also take concrete actions. In the next blog on this topic we will zoom in on the actions and policies that banks have in place, and discuss if the link between measurement and actions is sufficiently tight, or whether there is room for improvement.

For more information on this topic contact Diederik Hemmes (Senior Consultant), Johannes Lont (Manager) or Vincent Schothuis (Manager)

If you want to join our RiskQuest team, please check our current job openings here

[1] Task Force on Climate-related Financial Disclosures