In October 2013, the European Insurance and Occupational Pensions Authority (EIOPA) announced its intention to undertake stress testing of the European insurance industry in its 2014 work plan. They have now (20th January) updated their website to give more details on this insurance stress testing exercise.
The stress testing exercise is expected to include:
• Market risks under a combination of historical and hypothetical scenarios
• Insurance risks
• Impacts of low yields and low interest rates
EIOPA plans to consult with the industry in March 2014 and launch the Europe-wide stress test exercise by 30 April. National Supervisory Authorities (“NSA”) will collect and validate submissions by 20 June for onward submission to EIOPA. EIOPA currently expects to publish the results in November 2014.
As in previous stress testing exercises, each local NSA will be responsible for identifying and contacting individual insurers for inclusion in the exercise. It is not clear at this stage which insurers will be asked to participate in the stress testing exercise.
A link to the EIOPA 2014 insurance stress testing timetable and process is here.
If you have any questions or comments on this eAlert or any other aspect of Solvency II, please contact your usual Milliman consultant.
This e-Alert is intended solely for educational purposes and presents information of a general nature. It is not intended to guide or determine any specific individual situation and persons should consult qualified professionals before taking specific actions. Neither the authors, nor the authors’ employer, shall have any responsibility or liability to any person or entity with respect to damages alleged to have been caused directly or indirectly by the content of this e-Alert.
Milliman’s Neil Cantle and Martin Neil (of Agenda LTD) assess the implementation of Bayes’ Theorem as a solution in performing stress tests for capital allocation under Solvency II in this article published in InsuranceERM.com.
Here is an excerpt from the article:
The Bayesian recipe for successful stress testing involves blending expert insight, narrative history and data. But how does one go about cooking these ingredients? We know that traditional “frequentist” risk concepts get us only so far, so we need a different way of cooking our stress test model.
Compared with the frequentist approach, modern Bayesian approaches to stress testing are significantly more flexible and powerful in situations where there is little data but where we have some explanatory insight into the mechanisms underlying the risk process and could use this as the basis of the model.
You might think that this approach is inherently subjective. It is. But that doesn’t mean that it is “simply making up the numbers”, because it is a highly pragmatic realisation that all models are just that — models.
As such, there is no single “correct”, unassailable, objective model out there awaiting discovery. It should therefore be obvious that differing enterprises facing different risks, with diverse histories will develop different constituencies with dissimilar views on the risks they face. After all, if we all had the same model of risk there would be no marketplace.
Comparing the Bayesian approach with statistical approaches shows that purely statistical approaches are merely phenomenological, in that they do not offer any way to explain why events occur and so offer no means to influence these events in future. In essence, they can be summarised as offering nothing more than a “stuff happens” explanation of risk.