This study by Guy Carpenter revalidates the findings of the Under the Lens: Investigating Cyber Vendor Model Divergence report, based on the latest model versions while expanding the scope to include tail losses. Revenue continues to be the key driver of model divergence at the mean and was found to be an even more significant factor at the tail. Industry sector and country of domicile, on the other hand, were less important at the tail, where affected entities are highly correlated and individual characteristics are less distinct. Impact of coverage on model divergence varies and is liable to fluctuate based on vendors’ reaction to recent events and the current threat landscape.

Cyber catastrophe modeling is constantly in motion, due to ever-evolving threats and a continued drive to advance the precision and capability of the models. This study examined the observed divergence across 3 leading cyber catastrophe models, CyberCube Portfolio Manager, Guidewire Cyence and Moody’s RMS, using predictive analytics. Since the original study’s publication in June 2023, all 3 vendors have released annual updates to their model methodologies and parameters (Model versions are v5 for CyberCube, M6 for Guidewire Cyence, and v7 for Moody’s RMS.)

As they add more granularity and sophistication—and as new vulnerabilities and attack vectors appear—the models offer differing reactions to each input. Macro views of cyber aggregation potential may also evolve, potentially leading to greater…