During September, RMS ran a series of cyber risk seminars in London and New York. These half-day events coincided with the release of RMS Cyber Solutions version 4.0 and featured both RMS and industry experts discussing cyber risk and the opportunities for the cyber insurance industry.
At both events, the day kicked off with Dr. Andrew Coburn, senior vice president for RMS, examining recent developments within the cyber risk landscape by outlining the approach RMS takes to tracking and categorizing the wide range of evolving threat actor groups. He also proposed some key future trends, such as the potential impact of a “gloves-off” nation-state cyberattack and its implications for the cyber insurance industry.
Former ethical hacker Eireann Leverett dug deep into the topic of contagion mapping and how hacking groups – both good and bad, are utilizing innovative techniques to map out the digital world. He also touched on the growing use of deepfakes in spear phishing attacks, whereby executive identities are faked to trick employees into fraudulently transferring funds out of the business.
To provide the industry’s perspective, we were delighted to be joined by two expert panels in London and New York discussing the cyber market and the role of models to support growth. Thanks to Jamie Pocock (Guy Carpenter), Laila Khudairi (Tokio Marine Kiln), Rory Egan (Munich Re), and Kirsten Mitchell-Wallace (Lloyd’s) for participating in London, and to Anthony Shapella (AIG), Jon Laux (Aon), and Kara Owens (Markel) in New York.
For the second half of the agenda, members of the RMS cyber team focused on the release of RMS Cyber Solutions version 4.0. This release features substantial enhancements to the RMS model and capabilities across several key areas including exposure data enrichment, expanded model data sources, and new stochastic modeling approaches to quantify cyber risk.
Dave Gatey, senior director – modeling for RMS, revealed how new modeling methods, such as agent-based modeling and multi-compartment models were being used in RMS Cyber Solutions v4. Chris Vos, lead modeler for RMS, took to the stage in New York, and myself in London, to give context as to how these improvements to the model and software will assist clients in understanding their cyber risk and therefore making better decisions for their business. In New York, the RMS cyber seminar was followed by a half-day terrorism seminar.
Introducing RMS Cyber Solutions Version 4.0
For many insurers, obtaining complete and accurate exposure data from cyber submissions remains a challenge. Often, these submissions are missing key information such as business revenue, profit, or business sector – all attributes that are critical to understanding the potential effect of cyber events.
To address this, RMS has released a company database consisting of 13 million companies across 30 countries, alongside a data enrichment engine that uses a custom similarity matching algorithm to allow users to enrich their exposure data. This will help ensure the inputs into the model are as accurate as possible, reducing model uncertainty, and minimizing an insurer’s data collection efforts.
Although historical data does not show you the whole picture when it comes to cyber risk, it is still critical to inform the lower return period scenarios. To enable this, RMS has invested substantially in automating our historical event data collection techniques by employing bespoke machine learning algorithms that extract event data from hundreds of thousands of unstructured data sources. These new data sets cover multiple event types including breach, malware, ransomware, and cloud outages and allows our v4 model to be run at a significantly increased level of granularity, supporting greater risk differentiation.
RMS has continued to research the causal processes that drive cyber risk, working closely with our partners across cybersecurity and academia, to map out and build simulations of these underlying processes. By stochastically modeling these individual components and applying game theory models to explore threat actor behavior, we can extract probabilities associated with both short- and long-tail cyber events.
Investing in Cyber-Physical Loss Models
Finally, RMS has maintained its substantial investment in cyber-physical loss models. These models take data from the EDM (the RMS property exposure data store) and other casualty classes to quantify the impact of clash-type cyber catastrophe events such as power blackouts. This allows insurers to explore the potential for silent cyber losses across their business, supporting regulatory reporting. Many insurers are exposed to this type of cyber risk, even if they don’t write affirmative cyber insurance policies.
These new insights and models continue to be delivered within an open modeling framework, allowing complete transparency into each of the modeling components. This transparency allows users to validate each component and create custom models to support their own view of risk.