Wei Wenlong: The Climate Risk Business Director of Zhongchengxin Green Finance Technology (Beijing) Co., Ltd., a certified FRM holder, with dual Master's degrees in Financial Mathematics from the University of Minnesota and Environmental Science and Engineering from Tsinghua University. He specializes in the field of quantitative analysis of financial risks and has published over ten articles in core academic journals.
Abstract
The lack of historical data accumulation for managing and addressing climate risks, coupled with their long-term nature, makes it difficult for traditional risk measurement tools to be effective. As a result, stress testing has become an effective method for evaluating the impact of climate change and economic transformation. This article compares two methodological approaches to climate physical risk stress testing that have emerged in practice, and integrates physical risk into existing climate stress testing methodologies. The article analyzes the current domestic and international applications and development trends of the climate physical risk stress testing method.
Key Word: Physical Risk; Stress Testing; Commercial Banks
Chinese Library Classification Number: F832.2 Document ID: A
Article Number: 1009-1246(2024)01-0037-05
1. Background
Conducting climate risk stress testing is a powerful tool for financial institutions to actively respond to climate change, proactively manage climate risks, fulfill social responsibilities, and assist in transformation and upgrading. Climate risks are divided into transitional risks and physical risks. According to the definition given by the Central Banks and Supervisors Network for Greening the Financial System (NGFS), physical risks refer to risks related to climate patterns gradually changing due to climate change, specifically referring to chronic risks (such as gradually increasing temperatures) and acute risks associated with increased frequency and/or severity of weather events (such as tropical cyclones, storms, floods, and droughts).
In 2021, 23 major banks in China completed the first phase of climate risk stress testing, which focused on carbon transition risk stress testing with carbon emission trading prices as the main factor, targeting the three high-carbon industries of thermal power, steel, and cement. The test examined the impact of rising carbon emission costs on the repayment ability of enterprises with annual emissions of over 26,000 tons of carbon dioxide equivalent in the three industries, as well as the impact on the credit asset quality and capital adequacy level of participating banks.
Climate physical risk stress testing has strong spatial differences and involves many risk factors. Currently, it is in the exploration stage and has not yet formed a unified testing plan. Starting from the common processing ideas of climate physical risk stress testing in practice, this article summarizes two mainstream methodological approaches currently applied and selects one of them as a case study for practical application in a certain commercial bank.
2. Two methodological approaches for physical climate risk stress testing
One method is a combination of top-down and bottom-up approaches. Starting from the spatial characteristics of climate risk factors, this method predicts the spatial and intensity distribution of climate risks under different stress scenarios in the future, based on the three elements of climate physical risks: hazard, exposure, and vulnerability. The method evaluates the impact of these risks on the devaluation, cash flow, and other losses of financial institutions' assets in different regions. The advantage of this method is that it is easy to understand, but the disadvantage is that it has strict requirements and generally requires the use of geographic information systems (GIS), which have high requirements for data quantity and data granularity. MSCI proposed a physical risk quantification management plan that classifies six extreme physical risks (heatwaves, coastal floods, river floods, cold waves, typhoons, and wildfires) based on hazard, exposure, and vulnerability. A certain bank analyzed seven physical risk factors, including water scarcity, heatwaves, extreme cold, drought, floods, ecological environment damage, and sea-level rise, in a semi-quantitative form, combining with the spatial distribution characteristics of its assets in its 2022 environmental information disclosure report.
The other method is a top-down approach. Based on traditional macro stress testing methodologies, this method quantifies the impact of climate change on macroeconomic and asset risk variables under different stress scenarios by introducing core climate variables (such as temperature and precipitation) and constructing a relationship between core climate variables, macroeconomic variables, and risk variables of different industries of financial institutions (such as banks) using econometric models. The advantage of this method is that it uses existing risk quantification management systems, and the methodology is relatively mature and easy to implement. The disadvantage is that the explanatory power of core climate variables for macroeconomic variables is insufficient, and further research is needed to explore intermediate variables or build intermediate models to transmit stress. For example, a certain bank used ARIMA forecasting, Monte Carlo simulation, and Wilson models, selected macro variables such as GDP, PPI, CPI, M2, as well as climate variables such as precipitation and temperature, and constructed a manufacturing default probability PD model using data processing methods such as differential, year-on-year, and chain ratios to test changes in credit asset quality of the manufacturing industry under mild, moderate, and severe stress scenarios.
3. Application of top-down approach for physical climate risk stress testing in a commercial bank: a case study of a certain commercial bank
01 Main objectives
The main objective is to explore the impact of physical risks on different industry credit customers of the bank, identify physical risk factors, analyze stress transmission paths, construct stress transmission models, and calculate and analyze the impact of climate risks on risk control indicators such as the credit asset quality and capital adequacy ratio of the bank under different scenarios. Considering that the top-down approach uses the existing risk quantification management system, is mature in methodology, and easy to implement, the top-down approach is chosen to execute the physical climate risk stress testing.
02 Influencing factors
Statistical data shows that heavy rain and drought are the main climate factors that have a significant impact on the bank's region. As climate change intensifies, temperatures continue to rise, and droughts occur, the temperature causes systematic interference to the atmospheric water cycle, resulting in an increase in extreme weather disasters such as frequency and intensity of heavy rain events. Since the data series for heavy rain and drought are not long enough, variables such as precipitation and average temperature are used as characteristic variables.
03 Stress transmission
The main physical risk factors of climate change are determined to be precipitation and average temperature. Stress transmission paths from physical risks to economic variables and then to changes in credit asset quality are constructed. On the one hand, extreme heavy rain can cause floods or waterlogging disasters, increase precipitation, and to some extent cause a reduction in agricultural production, damage to industrial production, directly causing regional economic losses, affecting the production and operation activities of market entities, and leading to a decrease in operating efficiency and debt repayment ability, ultimately causing a decline in the credit asset quality of banks. On the other hand, extreme drought causes a reduction in agricultural production and supply, and the pressure of rising agricultural product prices is transmitted to industries and other fields, causing a rise in production costs and impacting the debt repayment ability of market entities, ultimately leading to a decline in the credit asset quality of banks.
04 Determining the physical climate risk stress model
Based on the Credit Portfolio View (CPV) model, a physical climate risk stress testing model is constructed. Near non-performing loan rate data for a certain industry is selected, and variables such as local and provincial GDP, precipitation, average temperature, and national PPI, M2, Shibor, etc. are collected and transformed into pre-processing indicators such as year-on-year growth rate, cumulative value, difference, and Logit conversion. The autoregressive function (ACF), partial autocorrelation function (PACF), and white noise test (Ljung-Box) are used to eliminate sequence correlation and model the relationship between variables. A CPV panel model is constructed with relevant hypothesis testing. The length of the data used in the model is 8 years, with a quarterly frequency.
The main physical risk factors of climate change are determined to be precipitation and average temperature. Stress transmission paths from physical risks to economic variables and then to changes in credit asset quality are constructed. On the one hand, extreme heavy rain can cause floods or waterlogging disasters, increase precipitation, and to some extent cause a reduction in agricultural production, damage to industrial production, directly causing regional economic losses, affecting the production and operation activities of market entities, and leading to a decrease in operating efficiency and debt repayment ability, ultimately causing a decline in the credit asset quality of banks. On the other hand, extreme drought causes a reduction in agricultural production and supply, and the pressure of rising agricultural product prices is transmitted to industries and other fields, causing a rise in production costs and impacting the debt repayment ability of market entities, ultimately leading to a decline in the credit asset quality of banks.
05 Determining physical climate risk stress scenarios
The three temperature rise scenarios corresponding to the six NGFS scenarios are adopted, namely 1.5℃, 2℃, and 3℃ scenarios, which correspond to mild, moderate, and severe stress scenarios respectively. It is assumed that the bank's credit asset exposure, non-performing loan rate, risk control indicators, and macroeconomic factors for the participating industries remain unchanged at the 2021 level, and 100% provision for new non-performing loans is adopted to analyze the impact of temperature rise on the credit asset quality and risk control indicators such as the capital adequacy ratio of the bank.
06 Analysis and discussion of test results
Based on the model economics interpretation, it is found that under the background of climate change, assuming other factors remain unchanged, the higher the temperature, the greater the impact on the credit asset quality of the participating industries of the bank. The test found that after provision for impairment losses, the impact of the participating industries on the credit asset of the bank was limited under mild, moderate, and severe stress scenarios. In the test results, the non-performing loan rate increased, reaching 1.71% under the severe stress scenario; the provision coverage ratio decreased, reaching 117.14% under the severe stress scenario, still higher than 100%; the capital adequacy ratio decreased, reaching 15.10% under the severe stress scenario, which meets regulatory requirements and is higher than the minimum requirement of 5.1 percentage points; the Tier 1 capital adequacy ratio decreased, reaching 11.10% under the severe stress scenario, which meets regulatory requirements and is higher than the minimum requirement of 2.6 percentage points; the capital adequacy ratio decreased, reaching 11.10% under the severe stress scenario, which meets regulatory requirements and is higher than the minimum requirement of 3.6 percentage points.
07 Optimization direction
This study explores physical climate risk stress testing starting from two macroeconomic variables, average temperature and precipitation, and the related methods have certain reference significance. However, there are also two shortcomings. First, static testing methods are used, assuming that the balances of relevant accounting items on the bank's balance sheet remain unchanged, which does not match the actual situation. Second, precipitation and average temperature variables do not directly affect macroeconomic variables, so it is necessary to study potential intermediate variables or intermediate models to transmit stress and improve the interpretability of the model.
4. Conclusion
This article analyzes and compares two methodological approaches for physical climate risk stress testing, and uses a certain commercial bank as an example to elaborate on the practical case of integrating the top-down approach for physical climate risk stress testing into existing climate stress testing methodologies. Based on case analysis and application practice, the following conclusions are drawn.
Firstly, the top-down approach integrates physical climate risk stress testing into existing climate stress testing methodologies but needs to be improved. On the one hand, whether climate risk variables are used as exogenous variables to affect macroeconomic variables or to affect the asset quality of financial institutions together with macroeconomic variables, the specific logic of intermediaries needs to be improved. Intermediate variables or intermediate models should be sought to transmit stress and improve the interpretability of the model. On the other hand, NGFS predicts the macroeconomic variables under the global warming scenario and their interactions with various economic sectors based on the Integrated Assessment Model (IAM), the Global Change Analysis Model (GCAM), the MESSAGEix-GLOBIOM model, the REMIND-MAgPIE model, etc. Therefore, the SVAR model can be used to better connect climate variables with macroeconomic variables.
Secondly, with a certain data foundation, the top-down approach can integrate physical climate risk stress testing into existing climate stress testing methodologies and be used for localized physical climate risk stress testing. On the one hand, since climate variables have spatial differences, the top-down approach can be applied based on the idea of one model per region (province/city/county). On the other hand, constructing a macroeconomic forecasting framework that conforms to the actual climate change background is also a potential solution, and it is necessary to actively explore macroeconomic stress testing for climate risk scenarios.
Thirdly, physical climate risk stress testing will still be mainly based on the combination of top-down and bottom-up approaches in the future. Physical climate risk stress testing needs to analyze and manage the three elements of climate physical risk hazards, exposures, and vulnerabilities, and both academia and industry are exploring their application. As this type of method retains the quantification principles of catastrophe models in the insurance and financial industries, it has a good scientific basis and explanatory power for different types of climate physical risks and disasters, and its model mechanisms have been tested in practice. Therefore, this type of method may become the mainstream approach for physical climate risk stress testing in the future.
Fourthly, financial institutions should pay attention to physical climate risk stress testing. As an important component of climate risk, physical climate risk stress testing is gradually being understood and mastered by domestic financial institutions. According to the results of the climate risk stress testing conducted by the European Central Bank in 2022, the ECB is considering incorporating the testing results into the internal capital adequacy assessment process (ICAAP) of some financial institutions, which may have specific quantitative effects on the second pillar of the Basel Accord in the future. At the same time, some European financial institutions have already included climate risk stress testing in their ICAAP, prudently managing regulatory capital and actively responding to climate risk.
References:
[1] Sun, T., & Miao, M. (2023). Reflections on Financial Management and Coping with Climate Risk. China Finance, 16, 12-15.
[2] Xie, L., Zhou, F., Chen, S., et al. (2022). Analysis of the Impact of Climate Change on Financial Stability Based on SVAR Model: A Case Study of China's Banking Industry. Fujian Finance, 12, 3-13.
[3] Monetary Policy Analysis Group of the People's Bank of China. (2022). China Monetary Policy Execution Report (Q4 2021) [R]. Beijing: People's Bank of China.
[4] Research Bureau of the People's Bank of China. (2023). Promoting Effective Connection between Green Finance and Transformational Finance [R]. Beijing: People's Bank of China.
[5] IPCC. Climate Change 2022: Impacts, Adaptation, and Vulnerability. Contribution of Working Group II to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change[R]. London: Cambridge University Press, 2022.
[6] MSCI. MSCI Real Assets Climate Analysis: A TRANSPARENT APPROACH TO CALCULATING CLIMATE RISK[R]. New York: MSCI, 2022.
[7] NGFS. NGFS Climate Scenarios Database Technical Documentation V3.1[R]. Paris: NGFS, 2022.
[8] NGFS. Physical Climate Risk Assessment: Practical Lessons for the Development of Climate Scenarios with Extreme Weather Events from Emerging Markets and Developing Economies[R]. Paris: NGFS, 2022.
[9] ECB Banking Supervision. 2022 climate risk stress test[R]. Frankfurt: European Central Bank, 2022.
