Banks and insurance companies are counting the costs from frequent physical climate events, such as floods, wildfires, droughts, heat waves, and storms. Comparing the period between 1980 and 1999 to the years between 2000 and 2019, there has been a marked 83% increase in recorded climate disasters. In 2022 and 2023 alone, economic damages reached $451 billion, representing a 19% increase compared to the annual average from the preceding eight years. Climate change is having real-world financial, operational, and regulatory consequences.

Why does climate change matter?

Currently only 25% of climate-related catastrophe losses in the EU are insured. This significant protection gap leaves banks more exposed to financial losses from climate events.2 Banks in regions like Southern Europe have higher exposures to physical risk; over 60% of their corporate loans are vulnerable to climate hazards such as floods and wildfires.

Yet, despite mounting losses and investments in data and analytics, assessing and managing physical climate risk remains a complex challenge for businesses, banks, and insurers. This is due to climate data being nonlinear, data-intensive, geographically specific, and deeply intertwined with infrastructure and finance.

Why physical climate risk is challenging to manage?

Unlike market or credit risk, physical climate risk is highly location-specific, requiring an understanding of exactly where all the assets are located—not just the headquarters of a company, but its operational sites, factories, and shops, as well as the locations of its key suppliers.

Managing physical risk also requires understanding not just of the general climate trends of the region, but specific vulnerabilities of the site itself: Is a factory near a floodplain? Is a warehouse in a wildfire corridor? How heat-resilient is a city’s power grid?

Finally, the impact of climate change is not distributed continuously. Rather, it manifests as discrete events which grow in frequency or severity, making prediction difficult and risk hard to quantify.

What about the data?

Over the last decade, physical climate risk data has evolved significantly, moving from broad, scientific climate projections to high-resolution, asset-level analytics.

Initial global climate projection models provided very wide spatial resolutions covering areas of 100 to 250 square kilometers. These models were too broad for banks and insurance companies to use, since they require a much tighter area to pinpoint the risk of a specific asset. Advancement and accessibility in satellite imaging has led to data being available at a higher level of detail, and now commercial datasets can provide hazard information at a resolution of roughly 10 to 30 square meters.

Furthermore, physical climate data also comes in unstructured data formats, such as CSRD (Corporate Sustainability Reporting Directive) reports, which are increasingly used as a foundation for physical climate risk analysis. As companies must disclose climate-related risks and opportunities (including physical risks like floods, heat, storms, etc.), scenario analysis under multiple warming pathways (e.g., 1.5°C, 2°C, 4°C) and the impact on business model and strategy. These reports are another input that banks and insurers can use to assess exposure to physical climate risks.

Geolocation challenges

There are many things to consider when ensuring the geolocation of an asset has been correctly assessed.

Many commercial datasets lack precise geolocation—i.e., the latitude and longitude coordinates for assets, properties, or infrastructure. The geolocation is typically based on the postal code in the address, which can be of mixed utility, depending on the country being assessed. This is because postal codes (or, in the U.S., ZIP codes) are primarily designed to optimize mail delivery, rather than exact geographic accuracy. UK and Singapore provide very granular postal codes, while France and Germany are very broad. Therefore, analysis on UK or Singapore assets provides a more accurate geolocation footprint than one for France, where a single postcode can cover thousands of addresses—resulting in all the risk being assessed at that specific point.

It is also important to consider the overall geo footprint of an asset. The Boeing factory in Everett, Washington, for example, covers more than 100 acres but only has one address and one ZIP code.

Additionally, while other financial services institutions, such as insurance companies, have long experience with geolocation data, banks traditionally do not. They often lack the proper data structures, identifiers, and analytical capabilities to use such data effectively.

Changing regulatory landscape

Data challenges remain, but things have advanced to the point where regulators are taking the first step to integrating environmental, social, and governance (ESG) risks into their supervisory stress tests.

The European Supervisory Authorities (ESAs) published joint guidelines on ESG stress testing in July 2025. The ESAs include the European Securities and Markets Authority (ESMA), the European Banking Authority (EBA), and the European Insurance and Occupational Pensions Authority (EIOPA). The initial focus is on climate and environmental risks, including physical and transition risks. The intention is to harmonize methodologies and practices among banking and insurance supervisors. The ESA guidelines are scheduled to be finalized in Jan 2026.

Challenges in combining datasets

Analyzing physical climate risk continues to be challenging, despite the availability of increasingly detailed data. This is due to the volume and fragmented nature of the data, as well as the need to combine multiple data sources that vary in format, from structured to unstructured data.

Advancement in computing power and the rise of tools like Quantexa's Decision Intelligence Platform has meant that merging large datasets at scale has become less onerous—even taking into consideration issues like partial and incomplete addresses, variations in spelling—making it easier to consolidate entities and assign relationships among them. And, ultimately, making it easier for financial institutions to identify and manage climate risk.

Contact us to find out more.