The leakage of methane gas from the Elgin North Sea platform raises a number of issues that underwriters and risk managers should be considering. The situation was described as an “explosion waiting to happen”, but the same could be said of every platform globally. Here Mat Maddocks argues that an understanding of all possible scenarios is vital to the successful management of an energy portfolio.
Deepwater vs. Elgin – Before the Event Exposures
When modelling total losses the calculation of the loss severity, and (re)insurers exposure, is straightforward as it is based on the policy limits. However for partial losses the situation is much more complicated as there are many different scenarios that are possible, each giving rise to different levels of losses to different policies or policy sections. To have a full understanding of the underwriting risk being assumed, underwriters need to have analysed as many of these outcomes as possible and to ensure that they are prepared, from a capital and risk appetite point of view, for those eventualities.
During the Event Exposures
What is needed is a system in place that can call on the before-the-event scenarios, but make it conditional on the event that has begun. In this instance we could investigate all the scenarios we had previously generated for the Elgin platform, review them in light of what has actually happened (some may be capable of being discarded or others added that were not previously considered possible), and by considering the relative probabilities arrive at an expected loss or distribution of possible portfolio losses for the event. Over time we continue this distillation process of learning more about the event and being able to remove or add scenarios, giving an increasingly more accurate and reliable exposure figure. Ultimately underwriters have a constantly evolving, real-time, picture of their exposure that is based on the most up-to-date information available about the event.
A final feature of the Elgin event was the impact that human response had, or could have had, on the (re)insured loss. The reaction from the authorities and platform operators following the leak could have resulted in a number of very different outcomes. From a scenario modelling point of view this is where things get incredibly tricky as any experience, statistical or otherwise, can be fairly irrelevant when it comes to predicting future actions and attitudes regarding the correct response. What is required is a framework for modelling these scenarios that can capture the underwriter’s and risk manager’s views regarding these parameters so that the outcomes of these assumptions can be clearly understood. Although not currently relevant to the Elgin event it is easy to imagine the different outcomes if the authorities had taken a strong intervention approach to the incident or a let things run their natural coursestance.
Even after we have a comlete understanding of what has occurred and an understanding of the incurred liability it is still not a straightforward task to calculate the liability of a (re)insurance portfolio. The inter-relationships between coverage types, and the complexity of programmes purchased (insurance, reinsurance, and retrocession), means that this calculation is complex and requires a sophisticated system to accurately arrive at an auditable final figure. Reinsurance contracts may have over a 100 features which would determine whether or not a recovery can be made, and retrocession contracts could be triggered by market level characteristics which may be difficult to determine. In addition to this there may be movements over time in the level of awards and changes in who is responsible or liable as issues are settled in court.
In conclusion I think the main lesson from the Elgin North Sea event is that the ability to generate and process a large number of scenarios is key to the management of an energy account. It is important for underwriters to be able to generate an event set that contains a large number of possible scenarios prior to accepting the risk, which can then be used for exposure management, pricing, capital modelling, portfolio analysis, and, as is the case with the Elgin event, to understand exposures to actual events as they unfold. Just looking at realistic disaster scenarios based on total losses or policy limits is not sufficient for a line of business that is full of complexity and demands a real-time, accurate, and transparent approach to risk management.