Climate Risk: The mother of all data challenges

The financial services industry is becoming keenly aware of the risks posed by climate change, both to institutions and to the overall stability of financial markets. Regulators, investors and other stakeholders are demanding that institutions demonstrate an understanding of both the physical and transition dimensions of climate risk and take measures to mitigate them. It’s an immense measurement and prediction challenge that Microsoft is tackling on its own journey to net zero carbon emissions by 2030. We are gathering learnings about the sustainability data being used in every industry - data that can be used to help banks, insurers, asset managers and other FIs develop their climate risk management strategy and roadmap.

Assessing the full scope of climate risk

Physical risk refers to the impact rising temperatures are having on the physical environment and its associated impact on a firm’s business. Physical risk is more than storms, fires and floods that inflict damage to property. It also includes things like water availability on agriculture (and therefore CPG), warming oceans on fisheries and the deteriorating natural environment on tourism. Real estate values are already being impacted in many regions because of insufficient access to water. The Intergovernmental Panel on Climate Change (IPCC) has developed the Hazard-Exposure-Vulnerability model to help firms navigate the complexities of measuring physical risk:[1]

• Hazard: Defining climate change scenarios -?Climate change attributes and relationships are complicated and may not be fully reflected in current risk models. Climate models were designed to capture the global response to an increase in greenhouse gases and do that very well. However, the impacts of climate change at a local/regional level are highly mediated by uncertain atmospheric circulation. Moreover, in a non-stationary climate, risks may include events not previously observed in each location. Accordingly, models that depend only on historical data will underestimate the level of risk.
• Exposure: Identifying where those scenarios impact a firm’s assets -?While climate change is a global problem, its impacts are felt at a local level. Therefore, the risks cannot be assessed without knowing the precise location, function and importance of individual assets. It is also important to understand the exposure of a certain asset to impacts upstream and downstream in its supply chain, which can point to correlations in the impact of climate risk when events occur in different parts of the world.
• Vulnerability: Assessing the impact of climate events on the firm -?Climate events are not experienced equally. Their impact is highly influenced by the surrounding context. For example, the impact of a flood varies widely depending on the surrounding landscape, quality of infrastructure, economic activity, time of year and myriad other factors. The coverage and quality of data to characterize these contextual elements varies widely.

While physical risk focuses on the impact of the changing physical environment, transition risk aims to assess the potential losses that might occur as regulation, carbon pricing or shifting demand reduce the productive life of a firm’s assets. Known as “stranded” assets, these could include everything from coal-fired generation and cement plants to residential furnaces and internal combustion vehicles that don’t align with the path to net zero. Measuring transition risk introduces additional complexity:

• Transition and physical risk occur different timescales and can be driven by unforeseen catalysts (e.g. climate events in other jurisdictions). The data and processes needed to assess the cascading impact are still emerging and therefore could lead to inaccurate assessments and actions
• There is an inherent negative correlation between physical and transition risk. Rapid decommissioning of carbon-intensive assets mitigates physical risk, but heightens the risk of stranded assets. The interplay between physical and transition risk, and associated ability to measure and manage their cumulative impact, remains unclear
• Uncertainty regarding timing of the world’s response to climate change may limit the reliability of impact forecasts

Translating these risk inputs into actionable insights introduces additional challenges:

• Standards for climate data and services are still evolving. Data is collected from different sources and intervals with various degrees of scientific validity. This hinders the development of suitable products and services to manage risk. It also increases the difficulty of standardizing risk disclosure.
• Modelling and data management processes are opaque. Many current risk models are proprietary making it difficult to confirm their scientific validity and suitability. This could lead to inaccurate risk assessments and ultimately, rapid exit or unplanned exposure.

Charting a way forward

New, advanced, or adapted risk tools, data models and methodologies will be necessary for financial institutions to identify and assess material climate risks, establish climate-related risk strategy and stress test resilience to economic shocks from climate change. Based on Microsoft’s experience with climate data and risk analytics specific attention should be focused on:

1.?????Expanded Data Sets: Both the public and private sectors must invest in the data required to support the Hazard, Exposure, Vulnerability analysis recommended by the IPCC.

a.?????FIs will need access to climate scenarios that have sufficient spatial and temporal scales to inform asset-level assessment of physical and transitional risks.

b.?????FIs will need to collect a richer set of attributes to characterize the risks to individual assets, on an ongoing basis, as climate hazards can impact assets at any point in their lifecycle

c.?????Given potential disclosure limitations, FIs will have to explore alternative data sources (e.g. Satellite data, Natural Language Processing of unstructured information like text) to sufficiently understand their exposure. Effective use of such data sources will require firms to develop expanded data science capabilities to assess materiality and separate the signal from the noise.

2.?????Standardized Methodology: A common taxonomy will be necessary for stakeholders to assess aggregate risks to the financial system. FIs are already coalescing around the Sustainability Accounting Standards Board ?(SASB) standards and Task Force on Climate-Related Financial Disclosures (TCFD) recommendations; the IPCC framework is additive to this approach and can augment physical risk data prescribed under the TCFD.

3.?????Modern technology: Gathering, managing, and analyzing an ever-expanding data set requires modern technology. Cloud-native applications offer the flexibility needed to support the industry’s ongoing data journey. Machine learning tools and artificial intelligence will be essential to translating that data into meaningful risk signals and predictions. Many organizations are well along the path of their modernization journey, but further acceleration will help address the data and analytics intensity of climate risk use cases. It’s equally important for regulators to understand these technology capabilities so they harness it in service of their mandates.

What Gets Financed Gets Done

The financial services industry has a critical role to play in addressing the climate crisis by reallocating capital to enable the transition to net zero. After all, what gets financed gets done. This journey begins with understanding the risks climate change poses to the financial system, and that starts with understanding and reporting on the risks faced by the companies being financed. Then, firms can use that data to assess the exposure in specific loan, investment and insurance portfolios. FIs that invest in developing advanced analytics to more accurately assess and price climate risks can use this data to find opportunities in the expanding sustainable finance market. While new data and analytics solutions will be required, the industry can build upon its existing technology infrastructure and leverage partners like Microsoft that have a deep understanding of climate risk and how different industries are using data to address it.

?

[1] Determinants of Risk: Exposure and Vulnerability, Omar-Dario Cardona (Colombia), Maarten K. van Aalst (Netherlands).