PERILS AG, the provider of industry-wide European catastrophe exposure, industry loss data and indices, has launched a new website called Wind-Jeannie which provides forecasts of potential industry losses from European windstorms.
For PERILS this is a first step into pre-catastrophe event loss forecasting, rather than the established post-event loss calculation and reporting services that have been its staple since it launched.
PERILS has now been collecting data on European windstorm losses, as well as property insurance exposures in Europe, for over six years. By combining its expertise in understanding the magnitude of windstorm insurance losses, with the latest meteorological and weather data, PERILS has created a tool that can provide a loss forecast for an approaching storm as far in advance as 72 hours.
Wind-Jeannie has been designed by PERILS to assist insurance companies to prepare for large windstorm events. By providing information on the expected magnitude and geographical extent of a catastrophe loss in advance of it occurring, insurers can better understand their individual potential exposure enabling them to mobilise claims teams and prepare for the event.
The Wind-Jeannie website, which is available to users of PERILS services and those companies that report data on exposures to PERILS, provides insured property market loss forecasts for windstorm events across Europe for the next 72 hour period.
The estimated insured loss estimates are calculated using wind gust forecast data provided by the German Weather Service (DWD). The loss forecasts are updated every 12 hours, so insurers and users can gain a view of how the potential loss changes as a storm approaches.
Loss calculations are based on vulnerability data derived from storm events captured by PERILS and on market sums insured taken from PERILS’ Industry Exposure Database. The information is broken down by country and CRESTA zone, with detailed footprints of the forecasted gusts and the resulting losses available to be downloaded from the website for each individual forecast.
Luzi Hitz, CEO of PERILS, told Artemis that PERILS has been running a beta of the service for a year, comparing the output with storms in real-time and also back-testing the tool against historic storms. PERILS now feels that Wind-Jeannie’s forecasts are accurate enough to be truly useful to re/insurers and so is officially launching the tool today.
“We began testing Wind-Jeannie during the 2014/2015 European windstorm season and were very pleased with the information generated by the system. It takes less than 15 seconds to check the forecast and if conducted regularly enables users to establish a clearer picture of the potential impact of an approaching windstorm event and where the most significant insured losses might occur,” Hitz explained.
Hitz continued; “Post-event scenario loss calculations have been done for quite some time now, including by PERILS. But now that we change from a post-event to a pre-event view, and give a forecast of a windstorm approaching Europe, it really is something quite unique.”
Hitz told Artemis that with enhancements in new numerical weather and climate models have made this possible, with the technology improving tremendously and making the time right to put this data to use in a pre-event forecasting tool.
Having a view of the magnitude of a catastrophe event before it actually strikes is unique and could also have ramifications for the much discussed real-time risk management that the insurance, reinsurance and of course insurance-linked securities (ILS) market is moving gradually towards.
It raises the prospects of real-time trading of risk, with those insurance and reinsurance (or ILS) companies exposed to a storm being able to make an assessment about acquiring top-up coverage, in a ‘live-cat’ type scenario.
Hitz commented; “I can see going forwards that this will be used for short-term trading, maybe not in the immediate future, but once there are a number of such tools available for comparison.”
Hitz sees Wind-Jeannie as a tool that will help the re/insurance industry with reinsurance and retrocession claims handling, as well as with risk transfer in the future.
“When they know in advance the order of magnitude of a loss that could happen, it helps people to alert their internal claims department, mobilise them, even before a windstorm really strikes,” Hitz said.
Edi Held, Head of Products at PERILS, added; “Our aim in designing Wind-Jeannie was to provide a system similar to a standard weather forecast site, which would be user-friendly and generate data that was easy to understand and accessible via mobile devices. Instead of providing a weather forecast, it provides an industry loss forecast. Users can also set up email alerts for when a windstorm event is forecast to generate a market loss in excess of EUR 100m.
“Wind-Jeannie provides a new perspective on windstorm activity and I hope it will prove its worth during the upcoming windstorm seasons. Moving forward, we will look to improve Wind-Jeannie by continuing to collect loss information from actual events and comparing these to the forecasted losses provided via the website.”
While Wind-Jeannie is only available to companies working with PERILS or subscribing to its services, the website features a very interesting visualisation of future wind gusts over Europe. We encourage you to have a look.
“Wind-Jeannie” loss forecasting, example Storm Niklas (31 Mar to 3 April 2015): Wind-Jeannie uses gust forecasts from the German Weather Service (DWD) and converts the data into estimates of the expected insured property market loss. The loss forecasts are for the forthcoming 72 hours and are updated twice a day. Losses and gust values per CRESTA zones can be downloaded for each forecast as maps or in Excel format. The below charts show the loss forecast as would have been calculated by Wind-Jeannie on 31 March for the upcoming storm Niklas.
Subscribe for free and receive weekly Artemis email updates
Sign up for our regular free email newsletter and ensure you never miss any of the news from Artemis.