 Speech by Julian Adams at the Insurance Day Summit
I know that today is about the changing regulatory environment more generally, both internationally and here in the UK, and I’m sure that the panel session which I’m due to join immediately after the break won’t be short of material to cover. Right now, though, I want to spend my limited time with you covering one important driver of both domestic and international regulation, namely Solvency II.
First, I want to set out again the journey towards model approval from where we are now to the implementation of Solvency II; and second, I want to give you some specific feedback on some of the issues we have seen in the course of our work to date in order to help firms prepare for their interaction with us in these areas.
Model approval process
Firms will shortly start to move from the phase of our process called ‘pre-application’ to the submission phase. The submission phase shadows to the maximum extent possible what a final application will look like, recognising that, first, we do not yet have the legal power to accept an application under Solvency II or make a decision in relation to it, and, second, that the policy position in some areas is not completely settled and we are working with draft text.
As I say, firms will shortly begin to make their submissions to us – and indeed some will do so this month – in accordance with the timetable we have already agreed with them, and we will review that submission against the requirements we currently have. At the end of this phase, we will aim to form a view on the model and communicate this to the firm in question.
This could be that we believe that there is no further work to do, and the model meets the requirements as they stand; it may be that there is further work for the firm to do; or it may be at that point that we believe that the firm has made insufficient progress and stands no chance of approval before the implementation date.
At a later date, we will then accept formal applications from firms. I cannot tell you precisely when this will be, since this date is dependent on when we assume our formal legal powers under Solvency II, which is in turn dependent on the finalisation of the Omnibus II Directive. We believe that this date will be at some time in 2013. The nature of the application phase will vary from firm to firm, and will also depend on the extent to which the final policy position differs from that which we are working to now. Our approach is designed to ensure that we can do as much work up-front as possible, and minimise re-work where this can be avoided.
So that is a brief re-cap of what our timeline looks like in terms of model approval from now until the implementation date. I am aware that some firms are clearly anxious that the process is not moving as quickly as they would like. Obviously we will continue to keep our processes under continuous review but there is a very real sense in which our progress can only be as swift as that of the firms we supervise. If firms are not making sufficient progress, or are producing work which is below the standards we would expect, we won’t be able to proceed in the way that we have planned.
My message to you, therefore, is that our ability to make progress is every bit as dependent on you as it is on us, and that what some firms regard as undue delay in the model approval process may in fact be us deploying the appropriate level of challenge and scrutiny to ensure that the – rightly – high standards associated with model approval are maintained.
Feedback on internal model review
I’d like to turn now to some of the specific points we have seen during our model review work to date. Yesterday I wrote to all firms involved in our model approval process setting out this feedback, and today we have published this letter on our website, but I thought it would be useful to share some of our conclusions with you, particularly those points which are of special relevance to the London market.
I appreciate that feedback of this nature is always prone to sounding somewhat negative. But I make the comments in the spirit of wanting to point firms to areas where we have had the most questions to help them prepare for their interaction with us. I also fully recognise the huge efforts being put into the model approval process by the industry. Although the topics I’m about to cover are areas of common weakness, they are given in the context of a high level of engagement and excellent progress, which gives us a good degree of confidence in the work being done by firms.
The first point is in relation to methodology and assumptions, particularly in pitching these at an appropriate level so that they are adequately reflective of risk without being too complex. In many cases, firms have not considered up-front the extent to which certain issues may or may not be material to their overall position, and then applied an appropriate degree of rigour. For example, we have seen portfolios of bonds with specific, unusual features, modelled in exactly the same way as a standard portfolio, and we have seen blanket assumptions made across differing regions, lines of business and tail durations. Conversely, some of the work we have seen is so complex that it stands little chance of being understood or properly challenged, either by firms’ own boards or by us.
Similarly, we have seen issues with model scope. Some firms with very material catastrophe exposures have tried to claim that the catastrophe models which underpin all of the assumptions in this area are out of the scope of their internal model and we should therefore not look at them or how they are used. We do not regard this as being tenable, and will clearly need to look at the underlying catastrophe models where these are material. What is important in all cases is that firms can demonstrate how they have made choices about materiality, and how the judgements they have made have fed through to the level of complexity they have used in their model.
The second area where we think it would be useful to give feedback relates to aggregation and dependency assumptions. The amount of diversification credit a firm seeks to take is obviously critical to the overall output of its model, and in particular whether it presents a capital number which is adequately reflective of risk.
We expect firms to be able to explain the choices and assumptions which have been made in this area, and to have validated them accordingly. We also expect firms to carry out sensitivity testing on their diversification assumptions, so that they and we can understand better which assumptions are the most important drivers of the overall result. Our experience to-date is that firms have spent less time in this area than we would have expected and their work has been less convincing as a result.
We have made a number of separate observations in the area of validation and validation policies. The first and overarching point is that some of the validation policies we have seen have been so vague that we have not been able to draw any assurance from them. More specifically, we have seen a number of validation policies where judgements as to the materiality of certain elements have been made without necessarily being justified, or with little or no supporting analysis. We would expect decisions about materiality thresholds for validation to be more clearly articulated and justified. To clarify, the level of detail should reflect the materiality of the elements of the model; we are not asking for a detailed justification where it is clearly evident that the element of the model is immaterial.
The importance of a challenging governance process has also been highlighted in this area. During our review, we saw a validation policy which had been through the full governance process within the firm, and yet when we reviewed it a significant number of areas had been missed, including, for example, ad-hoc triggers for additional validation, the appropriateness of validation tools, the validation of partial internal model and standard formula integration techniques, and independence between the design and validation teams. The governance process needs to add value and challenge, and not simply be a box-ticking exercise through committees.
Turning away from validation, another important issue is the Use test. It is fundamental to the spirit of the Directive that an internal model is fully embedded within the business and that this can be demonstrated. For instance, we would expect the model to drive important business decisions, and to be central to the systems of risk management, governance and decision making. For those firms with early submission slots, we expect to see evidence of whether and how the firm has been embedding the use of the model and its plans to do so in time for implementation. Later, we will need to see clear statements as to how and where the model is used, for all firms.
I also want to mention very briefly model change policies. These sometimes seem to have been produced in abstract, and with materiality thresholds which are too high. When challenged, firms have sometimes struggled to give examples of how the materiality threshold could be triggered, and we would expect to see some form of back testing to ensure that change policies would be triggered under reasonable circumstances.
The final point is a specific one in relation to the London Market, and catastrophe-exposed business in particular. Catastrophe modelling has increased in sophistication in recent years, but it is still to some extent restricted to traditional perils in well-understood catastrophe zones. The increasing importance of emerging markets means that the nature of catastrophe exposure is changing, as is the ability of standard modelling techniques to understand it.
A recent example of this is the flooding in Thailand. We have seen a number of different approaches to the assessment of this type of risk, some of which have been sufficient and others not. The important point is that this type of unmodelled catastrophe risk, and the approach to assessing it, needs to be kept under review, particularly if business plans suggest an increasing focus on this sort of business. Equally, we would expect firms to keep under review the state of the art of catastrophe models, and the extent to which those are adapted over time to take account of wider risks or territories.
Conclusion
I hope that the foregoing has been a useful recap of some of what we have seen to date. I hope it also reassures you that we are trying to keep our focus on material areas of model development and use which can have a material impact on the levels of capital firms are required to hold. Of course, there is still much for us all to do if we are to keep on track to implement the Directive and we continue to keep our approach under review to ensure that it is operating effectively and efficiently. I am confident that by working together, we can achieve this.
|
