New models to unlock cyber
The infancy of historical claims data is perhaps one of the biggest hurdles for cyber risk modelling, but the advent of probabilistic modelling is a step in the right direction and can help the industry better manage the risk globally.
This is according to Yörn Tatge, senior vice president and managing director at Verisk Analytics, and Bernhard Reinhardt, manager of consulting and client services at AIR Worldwide, speaking to Baden-Baden Today.
AIR has developed a new probabilistic cyber model that has been calibrated with public, commercial and insurance claims data, including information on more than 60,000 incidents and the cybersecurity profile of over 100,000 organisations globally.
“It’s one of the largest datasets available,” said Reinhardt. “The beautiful thing is it’s now really a probabilistic model. We provide full transparency of the vulnerabilities and the assumptions that go into the events, and they are all accessible in the back end.”
This new model is capable of accounting for security breach and cloud service provider downtime incidents for insurance portfolios worldwide.
“The idea is that you generate thousands of scenarios for potential outcomes for the next year. There may be mutations of the event that show how it can develop; you basically have a full range of parameters that can be changed and come up with the worst-case scenario,” explained Reinhardt.
It is also part of AIR’s latest release of its Analytics of Risk from Cyber (ARC) platform, which can help companies estimate their additional losses from the EU’s general data protection regulation, for example from non-compliance, where they can be fined up to 4 percent of their revenue per year.
Tatge suggested that the expectations for cyber have been huge, and not having the tools to properly quantify the risk can make it difficult to provide the right insurance.
“On the lower return, higher frequency period you have relatively good data, but you want to have a better handle on the low frequency, high severity events,” added Tatge.
“This is what you can do with the model; you can determine the event as it may happen.”
Get all the latest re/insurance industry news with our daily newsletter - sign up here.
More of today's news from Baden-Baden