Analytical approaches for the climate-related risk estimation of commercial banks’ credit activities: challenges, opportunities, and the way ahead

ABSTRACT

Banks typically attempt to quantify climate-related risks, whether physical or transition ones, by adopting a top-down or a bottom-up analytical approach for the risk estimation of their borrowers. The two analytical approaches for risk estimation are regarded as mutually exclusive, when, in reality, they can complement each other in a mutually beneficial way. We discuss the challenges and opportunities of both analytical approaches with a focus on their applicability for commercial banks’ loans, and highlight directions for future research.

KEYWORDS:

• Climate change
• climate-related risk
• analytical approaches for risk estimation
• commercial banks
• loan portfolio

Over the past few years, banks have become increasingly aware that climate risks can strand assets in their portfolios and affect financial stability (Caldecott 2018; Caldecott et al. 2021). Banks typically attempt to quantify environment and climate-related risks by adopting either a top-down or a bottom-up analytical approach (BIS 2021; FSB2020). The top-down approach to risk estimation defines risk first at the broadest level, such as, for instance, sectors or industries, and subsequently maps the aggregate risk measure to individual exposures, such as individual credits (BIS 2021). Conversely, the bottom-up approach first estimates risk at the lowest level of aggregation, such as, for instance, an individual borrower, and extrapolates them up to a broader level of interest, such as the bank portfolio, or the whole banking sector of a country (BIS 2021).

Exposure to climate change risks increases corporate default risk, impacts the creditworthiness of corporate loans and threatens banks’ stability (Capasso, Gianfrate, and Spinelli 2020). How can commercial banks address climate-related risk that may result in financial risk for their loan portfolios by focusing on both top-down and bottom-up approaches for risk estimation? We first discuss the key features of the analytical approaches and subsequently introduce recent contributions with an emphasis on applications for commercial banks’ credit portfolios. Furthermore, we discuss the challenges and opportunities as well as future possibilities for the development of analytical approaches by commercial banks.

The FSB (2015) has classified climate change-related risk into three categories: physical, transition and liability risks. The first category covers climate and weather-related impacts on the value of financial assets; the second the risks that could emerge during the transition to a lower-carbon economy, and the third the consequences that could arise if a party who has suffered damage from the effects of climate change requests compensation from those it considers responsible. The data sources used typically differ depending on whether the modelling emphasis is on physical or transition risks (FSB 2020) as the usually chosen main focus in the existing analytical approaches for climate risk consideration that are typically translated into standard financial risk and valuation metrics, such as credit rating and Value-at-Risk (Bingler and Senni 2020).

When the goal is to quantify the financial impact of physical risks, top-down applications use climate data and can combine them with macroeconomic variables and sectoral data (Ortec Finance in NGFS 2020). In the case of transition risks, the top-down approaches can incorporate sectoral and macroeconomic data and combine them with information from the European Union's NACE database (Weyzig et al. 2014; Battiston et al. 2017; Monasterolo, Zheng, and Battiston 2018; Nieto 2019). By building on the latter data, Esposito, Mastromatteo, and Molocchi (2019, 2021) develop a sectoral environment risk-weighted assets tool that accounts for the pollution damage of Italian banks’ borrowers. Faiella and Lavecchia (2022) propose loan carbon intensity as an indicator. For every borrowed euro, the greenhouse gas emissions are calculated, originating from the fact that they are defined by the authors on the basis of NACE as more critical in terms of emissions and loan exposure sectors. PCAF (2020) has proposed a methodology with a similar goal to measure banks’ financed emissions for various asset classes. Teubler and Kühlert (2020) have applied this methodology in practice with the data of a concrete bank.

While top-down approaches have the ability to generalise findings and to prioritise external validity, they lack local relevance and contextual understanding (Loveridge et al. 2020). A possible approach to resolving these issues is to use a bottom-up approach, focus on a finer level of analysis, such as an asset or a company, and gather data on the exposure to physical and/or transition risks. In the former case, high-resolution geo-spatial data, based on an asset's physical location or whether it belongs to a climate-vulnerable region, can be used (Industrial and Commercial Bank of China, Tsinghua University, Four Twenty Seven and Acclimatise in NGFS 2020). In the latter case, a credit portfolio's asset level exposure measures tend to be based on its current impact on the climate, but also uses additionally implemented corporate sustainability information (PWC and Banca Intesa Sanpaolo in NGFS 2020).

The existing top-down and bottom-up analytical approaches rely on different data, quantitative methodologies, data granularity and availability, as well as the time horizon and breath of scenarios, which makes the comparison of model estimates among commercial banks challenging. In addition, the currently widely used NACE-based EU green taxonomy may be failing to incentivise green behaviour (Caldecott 2019). Concerning the pollution data used as the input, CO₂ emissions are a common modelling choice that is widely supported by the scientific community. However, although, for instance, methane also belongs to the group of essential anthropogenic greenhouse gases and is surging in recent years (NOAA 2021), it remains overlooked by commercial banks. This is an important omission because IPCC (2021) suggests that methane has added about 0.5C to the global warming since 1850–1900. Hence, commercial banks should target a more comprehensive consideration of factors that could affect their credit portfolios.

Since the financial system is interconnected (Roukny, Battiston, and Stiglitz 2018; Battiston et al. 2016), not least because a key borrower may be a client of several commercial banks, the lack of the possibility of a transparent and possibly comprehensive horizontal comparison among banks of their analytical approaches, but also vertically by commercial banks’ regulators, can trigger incorrect risk perceptions and business decisions. Therefore, future work by stakeholders on tools that are more comprehensive and comparable is needed, without the implication that a one-size-fits-all strategy should be targeted. Due to its importance for the stability of the financial system, credit risk interconnectedness among commercial banks could also be considered by adapting it to measure the similarity of their climate change policies in their lending practices by building, for instance, on Abbassi et al. (2017).

A further current challenge for the practical utility of the analytical approaches is that they are often regarded as mutually exclusive and are rarely implemented together (Acclimatise and Vivid Economics and Oliver Wyman in NGFS 2020), when, in reality, they can complement each other (Nicholls and Zochowski 2020). Consequently, an opportunity is missed in the context of the bounded flexibility concept of Ruff (2013). This suggests that rather than a single solution, a range of acceptable valuations for any issue exist and consensus efforts should focus on finding this range (Nicholls and Zochowski 2020). With respect to the analytical approaches, the top-down approach provides a range of acceptable average valuations within which boundaries the estimates of the bottom-up approach exist and are achievable (Ibid.). Therefore, a knowledge gap that commercial banks should address concerning their credit portfolios by building on the bounded rationality concept is how to best integrate the top-down and bottom-up approaches and, hence, achieve optimal external validity and local perspectives (Loveridge et al. 2020).

Another shortcoming of most current top-down and bottom-up approaches that are applicable for commercial banks’ credit portfolios is that they often address either physical or transition risks. However, it is important to assess physical and transition risks simultaneously because of the systemic nature of climate change and possible feedback loops between these two risks (CISL 2022). A potential way for commercial banks to consider the latter point is to adapt for their lending business the principles and workflow for the integration of both risks suggested by CISL (2022). Furthermore, considering the well-established trend in recent years for more climate-related lawsuits (NGFS 2021; GRI 2022), banks should introduce as a usual practice the consideration of liability risk, at least for a subset of sectors and countries that are most likely to suffer from it.

The analytical approaches for risk estimation are useful tools for the climate risk management of commercial banks’ credit portfolios. However, the alignment between climate risk management and climate outcomes (for instance, melting glaciers, floods, and droughts) should not be erroneously or sometimes intentionally conflated to happen by default in all cases (Caldecott 2022). In reality, this alignment may be defined as optional (Eggen and Stengel 2019) or happen by pure chance (Caldecott 2022). While making climate outcomes targets and transition plans mandatory certainly is a way to increase their alignment with climate risk management (Ibid.), an additional productive way forward for commercial banks might be to adopt a stakeholder perspective (Lubell and Morrison 2021) and consistently consider the voices of the local players that are currently missing from analytical approaches applied by them.

Author contributions

Conceptualisation, Writing – original draft: I. M.

Validation: A. B.

Writing – review & editing: I. M.

Disclosure statement

No potential conflict of interest was reported by the author(s).