Introduction to Climate Change Scenario Analysis
Climate change is considered to be one of the most pressing challenges of our time, with far-ranging and interconnected environmental, societal and economic ramifications. In response to climate change, central banks around the globe are taking steps to analyze the scale and impact it plays on their mandates. Most recently, the European Central Bank (ECB) conducted an economy-wide climate stress test, which examines the resilience of companies and banks to a range of scenarios. Over time, this could lead to greater scrutiny of the climate risks faced by individual banks. The Australian regulator (APRA) will also be conducting stress tests for banks this year and the UK regulator (PRA) will be running the second iteration of its climate biennial exploratory scenario (CBES).
Climate change stress testing is an important and complex endeavour that will highlight climate risks not yet uncovered. While we do not have all the answers, this blog aims to provide an introduction to the concepts of climate change scenario analysis and steps on how we have approached it.
What are the risks to portfolios from climate change and how can they be quantified?
These risks can broadly be broken down into two categories:
Physical Risks: These are the risks of a material impact to future cash flows due to changes in the frequency and/or severity of extreme events in response to climate change. Examples of such events include tropical and extra-tropical storms, floods, wildfires and droughts.
Transition Risks: These are the risks of a material impact to future cash flows due to policy changes aimed at addressing climate change. Examples of these include carbon prices, carbon taxes, stranded assets – i.e. limits on the amount of oil & gas which can be extracted.
These two sets of risks are complex and their interactions are often non-linear, as the speed of the transition to a low-carbon economy, and the time at which it occurs, drives the evolution of global temperatures and determines the relative prominence of the two risk groups. For example, a fast and early transition to a low-carbon world increases transition risks, but reduces the extent of physical risks, and vice versa.
In 2019 the PRA asked UK insurers to quantify the impact of climate change on their liabilities and assets. The intention of this exercise was to prompt insurers and financial market participants to engage in the construction and formulation of climate stress tests. On the liability front, they were often based on existing natural catastrophe modelling frameworks. The results were not published but the exercise identified data and knowledge gaps and enabled the Bank of England to fine tune their approach for the next round of stress tests starting in June 2021 which will include banks as well as insurance companies.
We have identified a three-step framework in response to the PRA’s requirements which can be viewed as a template for the broader stress testing of fixed income portfolios:
1) Identify sector and geographical exposures: The PRA provided a framework to identify which sectors will be most exposed to climate change (material sectors). In particular it focuses on sectors with heavy reliance on fossil fuels (metals & mining, oil & gas, non-electric vehicles (EV) autos, transportation) and those with heavy reliance on natural resources (e.g. water, agriculture).
2) Differentiate between winners and losers within material sectors: For example, in the automotive sector, company names that have already aligned their business models to produce EVs are set to benefit at the expense of those who are behind the curve. In the same way, energy companies which are already focused on lowering their carbon footprint and investing in renewable energy solutions are likely to benefit at the expense of those who have remained focused on their core business models. This part of the process is harder to quantify and relies heavily on the analysis of our career research analysts.
3) Quantify the potential market value impact under different climate scenarios: Once the risk is thoroughly understood, in its 2019 exercise, the PRA required an assessment of the potential market value impact under the following three scenarios.
- Scenario 1: Policy makers and companies do nothing & temperatures rise > 3.5 degrees Celsius by 2100. No transition risks but physical risks are significant.
- Scenario 2: Policy makers take some action & manage to contain temperatures to 2 degrees Celsius by 2100, but with a sudden, disorderly transition. Companies are exposed to both transition risks and physical risks.
- Scenario 3: Policy makers take decisive action & manage to limit global warming to 1.5 degrees Celsius above pre-industrial times in an orderly fashion. Companies would be exposed to significant transition risks but limited physical risks (PRA assumed none in this scenario). This last scenario is particularly relevant this year ahead of the landmark UN Climate Change Conference (COP26) due to be held in November 2021. Numerous countries are revising their pledges for reduction to be more ambitious and achieving these could require some sort of global agreement on carbon pricing.
These scenarios are being refined for the new iteration of the exercise, emphasizing the speed of the low-carbon transition and a shorter horizon to 2050. Please note that we are highlighting the above scenarios as examples rather than prescribing their broader use.
Focusing on the near-term risks, we have designed a proprietary climate stress test to address Scenario 3 which includes detailed country and sector inputs from our in-house research analysts. It aims to capture the immediate knee-jerk market impact from potential climate policy changes, taking into consideration a corporate holdings’ sector.
Following the three-step framework previously laid out, we first identify which sectors and countries we consider to be at most risk to spread widening based on their exposure to carbon intensive activities. Within sectors we focus on autos, energy, basic materials, utilities and transportation. For countries, we focus on Emerging Market (EM) countries and classify each EM country into one of two categories based on factors such as their past carbon emission levels, whether they are a commodity exporter or importer, and the impact of a hypothetical carbon tax as a percentage of GDP. We still apply a shock to the remaining EM countries that do not have direct exposure to commodities; a tail event would likely trigger a risk off sentiment. Due to the risk off sentiment, there could be a potential small rally for developed countries but we expect the impact to be minimal, hence we do not model this in more detail for this scenario.
In the next step, we differentiate between companies within sectors based on their emissions and readiness for carbon transition (See carbon transition framework for more details). To address this in a systematic fashion, we implement and use the Carbon Emission Quartile data from MSCI to differentiate between the winners and losers within each sector. This data includes companies’ actual carbon emission levels as well as their readiness for de-carbonization. Quartile 1 represents companies best prepared and quartile 4, the worst. During the stress test scenario, we apply a substantially lower base shock to each company in quartile 1 and a higher base shock to a company in the lowest quartile. The extent to which the magnitude of the shock differs by quartile depends on the sector. This is because certain sectors, such as energy, may not have fully operational solutions to reduce carbon emissions.
The last step is to determine the magnitude of the shocks applied. For sectors, we based our analysis on prior moves these sectors have seen during historical stress periods. For the purpose of this model, we simplified the credit spread widening to 100% of their current level on average for sectors that would be hit hardest by the climate policy change in automobile, energy and materials. The credit spread widening assumption for an average company in the utility and transportation sectors will be 75% and 50% respectively.
For countries we used a proprietary model to estimate the median cost of a hypothetical carbon tax which came to around 5% of GDP. This is similar to the EM fiscal deficit widening experienced during the Global Financial Crisis and most recent Covid-19 pandemic. We designed the country shock to be in line with the magnitude of impact in these periods. For the purpose of this model, we have used 100% sovereign credit spread widening for the countries with higher carbon emission level and 50% for the countries with a moderate carbon emission level.
The result from this particular climate scenario highlights a few interesting points:
- The markets most impacted would be the Emerging Market Debt and High Yield Debt markets, which could experience a loss of around 4% - 5% in absolute performance. This could be mitigated with an active management approach which considers climate risks in its decision making.
- Products that adopt a sustainable investment approach suffers a relatively smaller loss than the equivalent peers with traditional investment approach (-0.5% loss vs. -2% loss). The relative outperformance mainly comes from security selection within impacted sectors. The strategies with a sustainable investment approach typically hold higher allocations in companies in Quartile 1 & 2 (MSCI Carbon Emission Quartile Data) in comparison to a traditional investment approach.
Climate scenario analysis is an arduous task and most central banks and regulators are still working out how to include it into their policy frameworks and mandates. A good starting point is to identify risks within investment portfolios and, as we have laid out in this blog, explore potential impacts under fixed assumptions as a first step. A more in-depth analysis can then focus on precisely quantifying those impacts and their sensitivity to different climate and policy scenarios, using advanced tools that integrate physical and transition risks. Climate change risk analysis is a fast-evolving endeavour, and the rapidly increasing availability of granular data and advanced tools will only help tackle the problem more effectively.