Svetlana Borovkova: Climate stress testing – where to start?

Svetlana Borovkova: Climate stress testing – where to start?

Climate Change Risk Management Energy Transition
Svetlana Borovkova (foto archief Probability)

By Dr. Svetlana Borovkova, Head of Quant Modelling at Probability & Partners

What are the challenges in performing climate-related stress tests for various asset classes? And what are the emerging approaches and solutions?

In my previous column, I attempted to forecast the risks that will occupy financial risk professionals in 2024. One of those predictions focused on climate risk. And indeed, this issue is firmly on the agenda of financial institutions.

Regulation concerning climate risk is gaining momentum. By the end of 2024, banks are expected to incorporate climate risk drivers into their stress tests. However, challenges persist, such as a lack of approaches or datasets to facilitate the integration of climate risk drivers into credit and various other risk assessments. So, what are the challenges in performing climate-related stress tests for various asset classes, and what are the emerging approaches and solutions?

Physical versus transition risk

Firstly, it is important to distinguish between two main aspects of climate risk: physical risk and transitional risk.

Physical risk entails the risk to physical assets, such as houses (for mortgage portfolios) or assets of companies (for corporate credit or equity portfolios), resulting from extreme weather events such as floods or rising sea levels.

Transitional risk, on the other hand, encompasses the various financial risks arising from transitioning to a greener operational model, such as phasing out fossil fuels or improving energy efficiency of buildings. Such transitions can be extremely costly, potentially undermining the financial position of borrowers or significantly impacting the financial performance of affected companies and industries.

For corporate credit and equity portfolios, the main climate risk drivers are related to transitional risk, with less emphasis on physical risk, except for specific industries or regions. This is in contrast to mortgages, where both transitional and physical risks play equally important roles.


The lack of methodologies and data in the climate risk domain is further complicated by the high exposure granularity and relatively long-time horizons required for climate stress tests (typically set to 3 years for short-term 10 years for long-term).

Three main classes of approaches are emerging:

    • Top-down: Operates on the level of the macroeconomy and the ecosystem as a whole​.
    • Bottom-up: Operates on the level of individual assets. ​
    • Meso-approach: Operates on the level of sectors and geographical regions, where equity or credit portfolios are mapped into a sector/region matrix and climate risk assessment is performed separately for each cell of such a matrix.  

While the first, top-down approach is increasingly viewed as less promising, both bottom-up and meso-approaches can be useful. The bottom-up approach is clearly preferred, providing more insights into the exposure of individual loans or investments to various climate risk drivers. However, the meso-approach offers a viable alternative if granular data on individual exposures and sensitivities to climate risks are not available.


In any climate risk approach, scenario analysis plays a crucial role. Fortunately, we do not have to devise scenarios ourselves, as these are prescribed by the Network for Greening the Financial System (NGFS). The NGFS offers six climate scenarios, ranging from orderly to disorderly transition and the persistence of current policies. For each of the six scenarios, The NGFS provides forecasts of macroeconomic variables, emissions, carbon prices, and others, on both global and regional levels.


The main challenge in conducting stress tests for equity and corporate credit portfolios lies in translating NGFS scenarios into stressed default probabilities (for credit) and stressed equity value (for equity portfolios). This requires an understanding of how factors such as increased carbon price, carbon tax, and carbon reduction costs influence a company’s fundamentals, including leverage, profitability, revenues, and other variables. These, in turn, impact the company’s default probability and the Net Present Value (NPV) of future cashflows. This step involves a considerable amount of modelling and, at times, guesswork.

For instance, one approach is to quantify the increase in costs and the loss of revenues resulting from all expenses associated with carbon emissions and reduction. This can be achieved by using a cost pass-through rate typical for a particular industry and the price elasticity of demand. It is crucial to have reliable data on a company’s emissions – although it remains unreliable and often unavailable.

In the meso-approach, sector-wide averages for emissions and other essential variables can be used. This allows for the identification of sectors that are particularly sensitive to carbon transition risks. In a recent extensive exercise, we discovered that only a few industries, including mining, manufacturing, power generation, and real estate, are particularly sensitive to transition risks, while others are comparatively unaffected by them.

For instance, in the mining industry, the loss of value due to transition risk can vary from 20% to 60% for more severe and disorderly transition scenarios. In contract, for manufacturing, power generation and real estate industries, these losses range from 5 to 20%.   

Main takeaways

It is evident that climate risk and stress testing will gain more prominence as the regulatory burden in this area grows. While established methodologies are yet to emerge, progress is slowly being made. The main challenge lies in the realm of data: either its scarcity, difficulty in getting it, or its complete absence. Contrary to intuition, it is not physical but transitional risks (which are short-term in nature) that pose the greatest threat in financial terms. This is also related to the unpredictability of governments concerning their climate policies.

There exists huge variability in the impacts of transitional risk across industries and sectors, with some being highly vulnerable to climate transition while others remain practically unaffected.

Probability & Partners is a Risk Advisory Firm offering integrated risk management and quantitative modelling solutions to the financial sector and data-driven enterprises.