Captive Review, supported by Zurich, hosted an intriguing webinar exploring captive underwriting on Thursday, 26 January.
Dr. Paul Woehrmann, head of captive services in Europe, Middle East, Africa, Latin America and Asia Pacific for Zurich, was joined by Marcus Reichel, who is responsible for Merck KGaA’s captives in Luxembourg and Bermuda, and Pascal Prevost, Deputy Head of Group Risk Services at Nestle.
Christoph Betz, pricing specialist for captive services at Zurich, also provided a presentation during the webinar.
You can watch again below, while a number of questions that the panelists were unable to answer during the webinar have also been answered, in written form, below.
What revenue model for the captive is generally recommended? Profit, revenue neutral or cost centre? What are the impacts from all three?
Marcus Reichel (MR): We set our captives up as risk management tools, so we are targeting break-even results year-on-year.
Pascal Prévost (PP): Intercona Re operates as a profit centre and the targeted underwriting profit is 5% as per our risk management and ORSA policy approved by the board. Having said that, Intercona Re is the corner stone of Nestlé risk management strategy both for risk financing and loss prevention of insurable risks.
Which criteria do you apply to decide on the retention level of your captive?
MR: Various models, historical own data, benchmark data if available, own risk pattern, captive financials, all put into actuarial models.
PP: Same criteria, but additionally we will look at the capital requirements and profitability of the line of business, taking into account the minimum solvency ratio agreed with the BOD.
Can you specify which kind of models are used?
MR: Monte Carlo, Value at Risk.
PP: Additionally, the Expected Shortfall (TVaR) as required by the Swiss Regulator, but also expert judgement based on intuition and common sense.
Is your IT-tool self developed or bought externally?
MR: Riskeeper, which we bought externally.
PP: External systems for the collection of property insurance values, local policies, claims and premiums and capital modelling. A self-developed claims database for the reserving.
Do you use your captive to insure uninsurable risks, and if yes how do you assess frequency/severity and premium?
MR: No we do not at the moment, however, we would follow the same approach, review available data, define triggers and try to create sustainable statistics.
PP: We do not reinsure uninsurable risks as there is a limited list of risks we can underwrite as per the underwriting guidelines approved by the BOD. In any case, we rely on the underwriting expertise of our fronting partners (arms’ length principle).
In summary, what were the patterns found in the retail sector, in respect of captive structures? What drove this?
Paul Woehrmann: We have seen for property that the analysed captives have very few claims per year. But the claims that occurred often had a comparably large severity of more than €1m.
For liablity on the other hand, there has been a very high claims frequency, where the claims severity was comparably small.
These results were mirrored in the captive retention structure: for property the per occurence limit was in the range of several million Euros, with a per annum limit of usually two times the per occurence limit. For liability, the per occurence limit was usually a few hundred thousand Euros with no per annum limit or a very large per annum limit of X times the per occurrence limit.
We believe that the reason for the captive retention being so clearly in line with the claims structure is that our fronted captives take Risk Insights into account prior to setting the captive limits. For liability, the large frequency combined with low severity is most probably driven by the large portion of “slip and trip” claims in that segment.
Do you think the common claims patterns in the retail segment could make the creation of a mutual captive of retailers a viable option? Generally speaking, do you see mutualization of risks in the captive market?
Christoph Betz: I am not sure if this is a viable option since the risk is so similar. A mutual captive (or group captive) would be aggregating similar risks, especially if the retailers are geographically close, which could lead to high severity risks with a very high volatility, something that you actually do not want to have in a captive.
Instead, it would be more beneficial to cede different risks into a captive, to diversify the risk. For example, under Solvency II the ceding of both life and non-life risks leads – due to the diversification effect – to reduced capital requirements compared to two single captives with that same risk. The same benefit cannot be achieved by combining similar risks into one captive.