Mrs S. A. Dreksler, F.I.A.: I want to make it clear that I am speaking on behalf of the working party tonight – this was not all my own work, it was a collaborative effort.

The members of the working party have been fantastic. It is their enthusiasm for the subject, and for sharing their experiences and thoughts, and their eagerness to help others climb the learning curve in this challenging topic, which has enabled this paper to be written.

I am delighted that Kendra Felisky is opening the discussion today. Kendra chaired the working party from its establishment in 2009 until I took over in 2011. She laid down the key principles behind the working party and behind the paper. We inherited these principles and they have remained pretty much unchanged throughout. One of the key principles was not to provide guidance but to share ideas and examples to promote discussion.

An unintended consequence of this is that it has made the paper fun to write – perhaps more so than it would have been had we been trying to provide guidance. It has allowed individuals to incorporate their personal perspectives, and although we have challenged these as a group – and we have had some interesting and entertaining debates – we have not attempted to form a common view. This is an important point. I do not think that you should expect any member of the working party to be entirely responsible for the whole paper or to defend any particular detailed aspect.

We tried to make the paper as comprehensive as possible, including discussion of wider considerations such as communication. Inevitably, though, we had to stop somewhere. I know that there are additional areas on which we could have spent time, such as expert judgement and consistency with capital models, and areas where we could have gone into much more detail. Please forgive us if we did not cover the problems with, for example, the practical application of the principle of correspondence in as much detail as you would have liked. You are welcome to set up a follow-up working party to deal with those points. I am sure we would all love to hear from such a group.

Finally, before I hand over to Kendra, I should like to thank those of you who have attended and contributed to the many events that we have spoken at in the last 4 years. A key objective of the working party, and of the paper, was to encourage discussion because it is through such discussion, across the insurance community, that we will find pragmatic and efficient ways to implement Solvency II that will work for insurers and regulators.

Solvency II set out to be principles based and in some respects it has remained true to this. We, as reserving actuaries, have been left to work out how to implement the details of the requirements for Solvency II technical provisions and the balance sheet. It is a key strength of our profession that we come together to do this, but we could not have that wider debate if you were not willing to turn up and contribute. So, thank you.

Ms K. M. Felisky (opening the discussion): I commend the authors for producing such a comprehensive and extensive paper on technical provisions. I know that it was many years in the making and I am pleased that it has finally come to fruition. I should also like to congratulate the authors on winning the Brian Hey prize.

I believe that this paper will become a useful resource for any actuary involved with technical provisions. It concisely summarises the various requirements for technical provisions within Solvency II, and provides further explanation and advice on the key aspects of the requirements that may be unfamiliar or new to actuaries not yet fully conversant with technical provisions.

To start the discussion off, I will go over some areas that I think are worthy of further debate.

Segmentation is an area of concern as it would be an unusual insurer who currently reserves at the Solvency II lines of business. Although the base projections are not often at the level of segmentation required, so long as an appropriate allocation methodology is used to apportion the reserves, that should be adequate. The difficulty is in ensuring that the allocation is sufficiently detailed and not too complex to follow or explain. This paper describes various allocation methodologies that can be used.

Another area new to reserving is the risk margin, which has already been the subject of workshops and seminars. The paper includes a detailed worked example of how to calculate the risk margin. It also has a good example of an approximation that can be used. I like that the authors have recognised the difficulty in communicating what a risk margin is, in particular, to management. They have also recommended an alternative term called the market value margin, which might catch on. But I think we will stick with the risk margin for now.

Speaking of changing names for things, let us talk about Events not in Data (ENIDs), formerly known as binary events. When this working party first started, one of the biggest topics for discussion was binary events. These ENIDs reserves reflect an additional amount that needs to be included in the best estimate of claim liabilities to ensure that this estimate is a true best estimate of all possible outcomes.

It is still contentious as to whether an additional amount needs to be included at all. In this paper the authors have done a good job of bringing together the various pieces of guidance on this issue to allow readers to form their own view. The authors propose an alternative definition for ENIDs reserves of “the balancing amount required to bring the best estimate before ENID up to an amount allowing for all possible future outcomes”, which emphasises that the ENID loadings will vary not only by the nature of the risk, but also by the insurer’s reserving philosophy. The authors also provide warnings for anyone trying to benchmark such loadings.

Once the technical provisions have been calculated, it is a requirement of Solvency II that they must then be validated, which is different to a standard peer review of reserves. This paper contains a good discussion of the issues around what is required of such a validation. Particular areas to note are in regards to governance (such as who is independent enough to do it), back testing (with some worked examples in the paper), and the benefits from formalising the validation process. All this takes time (for both doing the validation and the associated documentation) and it is up to the actuaries involved to make the case to management that this will benefit the business.

I did notice that the paper deliberately stayed away from addressing the stochastic approach to calculating technical provisions. Although a pragmatic approach, because this is an area that is continually being discussed with no move towards a consensus, I think that the working party could have helped move this debate on by discussing some of the key issues around the stochastic modelling of technical provisions in this paper. I would be interested in hearing any views from the floor as to whether this is needed or not.

All in all, I welcome this paper to the actuarial library and thank the authors for preparing it.

Mr C. D. O’Brien, F.I.A.: I have a number of comments. First, in section 3.7, the authors refer to reserving committees and say that reserving committees should be looking at both sides of the balance sheet – the assets as well as the liabilities. That seems to me to be correct because it is this in which the insurer is really interested.

For example, if the discount rate changes when interest rates change and the technical provisions alter as a result, then what has happened to the assets? The assets would also, one expects, have changed as a result of interest rate changes. So, yes, the two sides of the balance sheet are important.

In section 3.8, the authors refer to the actuarial function. There was a meeting here at the beginning of November 2013 talking about the role of the actuarial function in general insurers. There were some interesting issues about the fact that parts of the balance sheet are currently calculated by finance department staff, other parts by actuaries working in an actuarial area, and others by non-actuaries doing a job that could be done by actuaries. It would be interesting to see how that works out.

In section 3.8.8, the authors refer to life annuities. They say that some insurers may increasingly use stochastic models. It looks to me to be quite challenging. It will be interesting to see how that might develop.

The section on expenses, section 5.1, says that the new Solvency II rules are a step change in terms of complexity. I think that is correct. I would also suggest that it was correcting an omission. If an insurer takes on a liability under a policy to pay claims, then it is also expected to incur expenses in the course of doing so. Those expenses may be incurred before the time of claim, and that is the liability. The insurer needs money to meet those expenses.

Section 6.3.5 refers to ENID, which I found most intriguing – the Events Not in Data. I did like the definition of ENID to which the authors referred. It is slightly different from binary events in the sense of something that may or may not happen. Instead it is something that could happen but did not within the period of the data that have been used to work out the assumptions in the calculation of the provisions. I think it is a concept, which might also be of interest to life actuaries, to pursue how it could be applicable. But it is clearly very hard to calculate the probability of these ENIDs, and one can see that it might become an area where the company decides to use it as a prudence margin. If that happens, then it will be wasting an opportunity to better understand the business. I certainly think that it is an area worth pursuing in terms of not just binary events but extreme events that have not happened over the past few years.

Section 6.7.1 says that companies’ reserving policies will differ. Reserving policies may differ when the rules say that the insurer uses its discretion to calculate a prudent level of reserves. Insurers may decide to be very prudent or less prudent. But I am not sure that is relevant for Solvency II. Solvency II says this is how you should calculate the technical provisions. You are not meant to be very prudent. I am not sure that differing reserving policies should survive post-Solvency II. I would be interested to hear the authors’ views on that aspect.

Section 8.5 refers to the calculation of the risk margin and mentions that there is also a risk margin in IFRS in the new rules that were proposed in June, which the authors have not specifically discussed. I feel that the risk margin that is intended to be applied in the accounts is nonsense. At least the risk margin in Solvency II has the advantage that there is a way to calculate it that is set out, and the authors have conveniently given an example as to how that may be done, whereas for the risk margin in IFRS it seems to be difficult to work out what you are meant to do. It is good that the authors have given the example of applying Solvency II rules, which will hopefully help general insurance actuaries in practice.

Lastly, I applaud the way in which the authors of this paper have brought together so many different strands relevant to the debate.

The Chairman (Mr D. F. B. Newton, F.I.A.): I note that technical provisions have been assessed in some companies for a while now, so there should be some people here who have gone through the pain of calculating technical provisions and who may be prepared to share their experiences.

One of the questions I have – and this could be for the working party itself – is this: you have put together a paper that is 120 pages long and that discusses how to calculate technical provisions. It is clearly not a straightforward process. How does one manage a process like that in practice? It is obviously not as straightforward as the old GAAP reserving. Is there an easy way of managing this? Those who are practitioners may have some views on this question.

Mr M. J. Wheatley, F.I.A.: I was a member of the working party. In response to that question, in many ways, it is not that hard. I am reminded of when Lloyd’s changed from an underwriting year basis to an earned basis. The first couple of years were chaos. How could somebody possibly calculate numbers on an earned basis? With many of the manipulations that are needed for the technical provisions, I think we will look back in 10 years’ time and wonder what all the fuss was about. I think that there are some new assumptions to be made regarding payment patterns. As an industry, I think we have a great “get out of jail free” clause at the moment that this would be introduced although yields were so low. Maybe there are some issues around some other assumptions. Once you have agreed on an approach, much of it is quite mechanical and straightforward. Certainly, within the Lloyd’s market, you are now seeing some people moving towards a fairly straightforward business-as-usual-type basis.

I think that question will be answered in a few years’ time once people have become more familiar with it. It will be interesting whether or not it is seen as a fuss about nothing.

Mr A. J. Newman, F.I.A.: It is a similar issue to the fact that under Solvency II the actuarial function is responsible for technical provisions. Why do people not just take responsibility for other people doing the work when they are not in the same area? That was alluded to earlier, I believe, by Mr O’Brien.

It is an issue that has long puzzled the profession. I remember in an earlier decade of this millennium when we had a working party on relying on the work of other professionals, trying to see what other professions in the United Kingdom do and what the actuarial profession does in territories overseas. They seem to have sensible ways, both the accountants in the United Kingdom and actuaries in America, for example, of going about relying on other people.

The actuaries in America use other people’s models – for example, catastrophe models – which you as an individual do not have the expertise to reproduce yourself. Nevertheless, you can still evaluate how good those models are. You can still evaluate other people’s inputs into your technical provisions.

Mr D. Paul, F.F.A.: I am responding to the comment about whether, in a few years’ time, it will seem like it was no fuss at all, and that there was not really a big change. There are still Solvency II technical provisions being calculated as an afterthought from the GAAP technical provisions. You have a company process that involves the finance team as well as actuaries, and then the actuaries, perhaps, do some differencing to get to the answer for Solvency II. We have not yet seen a migration into the full production process for the Solvency II technical provisions.

It seems to me that the Solvency II version of technical provisions is more complicated, more scientific. You would think that you would design a process that determined the technical provisions first of all, and then you would solve more simply for the GAAP provisions that you do not need to discount. That is, you did not really need the cash flow in the first place.

I feel that much of this is still being done in the wrong order. I am interested to hear if anyone else feels like that.

The Chairman: Mr Paul picked up on the added complexity of technical provisions under Solvency II. I think it would be interesting to explore how to explain these technical provisions to the various stakeholders, particularly how to make them fully understood by the board that is expected to be wholly responsible for them. How can we ensure that they have adequate understanding?

Indeed, what is adequate understanding? I would be very interested to hear views regarding what level of understanding boards need and how that understanding is conveyed.

Mr C. N. Critchlow, F.I.A.: I was not expecting to speak this evening. I hasten to add that I am not a general insurance expert. I hope that I will not say anything untoward.

My own experience of working with life companies is that the knowledge on boards is incredibly diverse. I should imagine that it is no different within the non-life sector.

For those firms where experience and understanding is not as great, it is a matter of taking them through a journey of understanding the various components of the balance sheet, both what they look like currently and how they might look under the new regime.

Of course, there are two possible outcomes. You may end up with a level of solvency before and after, which is very little different, in which case you are posed the question: “Well, why have we paid you such vast sums of money to give us the same answer as before?”.

But then you need to explain why and explain the additional level of understanding that Solvency II provides in terms of the information and the level of risks inherent within the business that perhaps you did not have before.

The alternative outcome is that you end up with a very different answer to that under Solvency I. This potentially leads to a much more interesting discussion with the board because then you can identify exactly where the level of prudence within the original basis may have existed, and why removing those margins then leads you to a recognition of those profits and what are the true underlying risks within the business and how you may be able to adapt to those risks over future years.

In terms of discussions with the board, it is just a matter of promoting those discussions, and education.

Mr J. Ryan, F.I.A.: One of the issues is that in trying to explain to boards, and indeed in managing the business, Solvency II reserves, in my view, is not where you start. You need to explain to the board that you have to have a sufficiency to meet the regulatory requirements. The risk metrics required under Solvency II are not the right ones to manage a business. Probabilities do not make any sense whatsoever. When you are putting in risk margin you need to take into account not so much probabilities of ruin but the extent of loss and more coherent risk measures.

When you are doing that – if you do that – some of the explanations about Solvency II reserves become very different. These issues are important and difficult, sometimes, to explain to boards – the ENIDs, and so on. The analysis becomes more complicated to do because you have probably more to come through. If you only have a probability of ruin, a lot of the events just go away. So I am all right for Solvency II but I am not all right if I am trying to run my company properly.

I appreciate that this is about Solvency II, and therefore the working party did not want to go into it, but I think it would be a shame if people using the paper – because it is an excellent paper and an excellent way to do things – did not realise that there are many useful things here, but there are other things that you need to do if you are going to manage your business properly as opposed to running Solvency II.

An interesting example of this is doing a small consultancy project for an overseas government that has a solvency issue. Interestingly, no money is required at all if I apply Solvency II correctly. That does not make any sense. But I would be able to sign a Solvency II reserve without any problem. That is something to think about.

Mr Critchlow: Just in response to that last point, I do think that Solvency II can be powerful in managing your business. This is because, even though the standard parameters that you may apply within the pillar one computation may not be particularly relevant for the risks in the business, surely this sort of discussion should be coming through in the pillar two regime, where you are looking at the own risk and solvency assessment? There you are trying to use the results of the pillar one computations to assess how your business may react in reality over the next 3, 4, or 5 years.

I think that Solvency II has to be made to work for your business.

The Chairman: Does anyone from the regulators have any views on this topic?

Mr J. Kirk, F.I.A.: I was on the working party too and I should first like to thank Mrs Dreksler for her effective oversight of the group.

The whole design of the working party was to look at the practical side as opposed to the theoretical side, and one reason the paper took so long was that we changed tack a couple of times. For items we first thought might be problematic, such as binary events and risk margin, we found that there are practical ways of dealing with them. When doing these calculations, you find the practical challenges. Items like interaction with your finance team, and knowing how your reserves make their way into accounts, are where the challenges lie.

Something to remember is that GAAP accounts are different to solvency accounts. Solvency II technical provisions are to assess the solvency of the firm. They are not necessarily reported in your accounts although there is a link – you have to understand that link. Mr Newton was asking for views from a regulator and we already use technical provisions on a Solvency II basis at Lloyd’s. I think the biggest issue I have seen is the interaction between the actuarial teams and the finance teams.

If you are being honest, do you come up with a set of reserves and hand them over to the finance team, and understand what is done to them? Do you understand the adjustments? You are going to have to reverse those adjustments out to determine the reserves on a Solvency II basis, and then put the equivalent adjustments back in. That is what I think people are finding difficult – this interaction with others and the bigger process.

The second point to think about is validation. It may be that some people skim over validation of technical provisions. When asked what you do to validate your technical provisions, is it something that you just ignore? Is it an elephant in the room for reserving actuaries?

That is a controversial question, but I would like others’ views, please.

The Chairman: In the discussion so far we have concentrated on the UK perspective but one of the things that interests me is what happens across Europe. Does anybody have a wider, pan-European, perspective on technical provisions for Solvency II?

Mr J. Courchene: At least on the Continent, where I have been practising for the last 6–8 years, it is rare that topics such as binary events or ENID come up. When they do come up, they often come up with remarks such as, “We have heard about this because we attended a GIRO and it is interesting”. But they do not end up finding their way into the actuarial reports.

If there were a remark to be made, it would be that the markets on the Continent may be a step or two behind in terms of thinking or addressing these types of topics, which is a compliment to the working party (of which I am a member).

The Chairman: Does that mean there is a material distortion in the level playing field across Europe?

Mr Courchene: I guess it would depend on the materiality of the ENID adjustment. It is difficult to conclude that it should be categorised as a distortion.

There are certainly other ways in which markets on the Continent insert a level of prudence, which is essentially what an ENID adjustment is doing. For example, in the German markets, they have an accounting principle that is named appropriately the “prudence principle” which necessarily brings their case reserves up to a level that is not expected to be paid but it is prudent to hold with the expectation that it might be paid under not a worst case scenario but a little worse than a best case scenario.

I would not necessarily say that the playing field is uneven, but the insertions of prudence, if they are coming in, and they obviously are, based on what has been written about ENID, are necessary because such events do not exist within the data itself. Perhaps such prudence is entering via different ways.

Mrs K. A. Morgan, F.I.A.: I just wanted to make one small point in response to Mr Courchene’s comment. One of my colleagues, Paul Sharma, who was instrumental in designing the Solvency II regime, is leaving the Prudential Regulation Authority. He is doing a series of lectures to disperse all his knowledge before he goes.

One of the points he made recently was that, in his experience, any implicit prudence in a technical provision is generally used up three times. He is quite cynical about the extent of prudence, particularly implicit prudence. I consider that is one of the good things about Solvency II technical provisions. Any prudence is explicit.

Mr Paul: I have been on the non-life working group on the Groupe Consultatif Solvency II project.

I have just a few thoughts about how advanced or not Continental actuaries are with this. We have had for a number of years now meetings of about 12–15 actuaries, representing the substantial member states in Europe. It is probably true to say that the people who come and work on Groupe Consultatif are quite advanced.

To echo comments already made, apart from a few of these individuals who have possibly worked in the London market, I think that it is true to say that those on the Continent are quite a long way behind.

There seems to be in some countries a tradition of booking a figure that the accountants come up with and then doing the triangles. Then you have a liability adequacy test, rather than what we are accustomed to, which is to start with booking the triangle number. There is quite a gap there.

One thing that seems to me a weak link for the United Kingdom is the Association of British Insurers Statement of Recommended Practice (ABI SORP) and what is the reasonable margin, or buffer. What we have all been booking under ABI SORP has, I think, and I hope others would admit as well, not been very clear. In that sense, perhaps we are not so great compared to the Continent.

We have sent this paper out to the Continental group of actuaries in Groupe Consultatif, and we are hoping to get a debate going about it inside the group, particularly about ENIDs.

Mr C. M. Smerald, F.I.A.: Before this began, I had a couple of conversations about technical provisions, and one thing that came up was complexity. Reserving, communication, and liability are ever more complicated.

I will add that, as part of the technical provisions guidance, there is an expectation to choose the most appropriate method. I do not remember the exact words.

The question is: “How can we assure others that we have used the most appropriate method?”. I wonder how the working party approached this matter.

That ties into the other themes of potential liability. You can start to be in trouble if you say, “Well, the answer could be this instead of that for these reasons, and here is why I chose what I did. But here is what I could have chosen”. You are trying to explain something more complicated. What I am trying to find out is how we approach these best practice questions as part of the study and keep ourselves safe.

The Chairman: Mr Orr, is the herd instinct a topic being covered by one of your working parties?

Mr J. B. Orr, F.F.A.: That is right. Within general insurance, certainly among actuaries, there has been the suggestion that bad news is gradually recognised and there tends to be a drift towards a shared view.

An example would be how the recognition of the costs of the Asbestos, Pollution and Health liabilities was first brought up and then recognised and reflected in businesses’ numbers. It was speculated that it might be useful for us to set up a behavioural experiment. You have heard of people like Daniel Kahneman, the author of “Thinking, Fast and Slow”. We see behaviours that emerge from groups. That is not what an objective, intelligent individual might do but what emerges from a group where you have things like agency risk and availability of data and the appetite of business as a whole to recognise cost. Perhaps it is implied in the design of Solvency II that we try to move towards greater objectivity in our approach. Part of that is looking at ourselves and the way in which we make decisions. So yes, Chairman, thanks for the prompt. We are trying to set up the working party, which is called “Quantifying the Herd Effect”, as part of GIRO 2014.

Mr Paul: Just a quick answer to the question about the choice of methods. The Groupe Consultatif non-life group did a paper in either 2008 or 2009 that was motivated by trying to put the record straight on some elements that we felt were going on and with which we were unhappy. When I say “we” I mean a proper representative group of Continental actuaries, not just UK actuaries. At that time CEIOPS was tending to veer off on a tack of thinking that non-life reserving could be mechanical. There were times when I think they thought we could identify the Solvency II line of business and that would then tell us whether it was Bornhuetter–Ferguson, chain ladder, or whether we should use “incurred” or “paid”, or whatever.

In the paper, we tried to get back to the idea that you needed judgement. I do not know whether we were therefore responsible for the phrase “expert judgement”.

You can see why regulators do not like all this “free form” approach, realising that you could come up with three or four different answers, depending what choice of method you use. We have tried to say at times that in most of non-life reserving you are going to find out in 2 or 3 years’ time whether you were right. That much at least is a safety valve.

I have one other comment. That paper was written to try to put the record straight on other things. I hope that this paper was not written to try to put the record straight on some of the things that the Groupe Consultatif said 4 years ago.

The Chairman: Does anyone have a view as to where we go from here? What are the next areas to be explored, if not by this working party then by other working parties, regarding aspects of technical provisions, either calculations or the way they operate in practice?

Mr Ryan: This codification is quite useful. Solvency II applies to many relatively small insurance companies, captive insurance companies, and small specialist mutuals. I must be careful in light of some of the governance issues that have been raised recently.

Raising some of the standard issues and some of the ENID issues forces them to think about things that they were not thinking about before. They have not put the same effort into doing the reserves and valuations in the same way as many of the big companies and the companies in which most of us here are involved. Therefore, the codification, as laid out in the paper, is valuable to such an organisation. It can have a checklist. I suspect the biggest problem is for some of the regulators – not necessarily in the biggest companies but the other ones. Trying to identify where things go wrong is generally what those companies do or not do because they think that they are on top of everything. This is useful from that point of view.

The Chairman: Do you think that there is something that smaller companies should do differently? Or should they be paying attention to particular aspects that are different to the larger organisations?

Mr Ryan: The smaller organisations tend not to worry about the contingency reserves, the margins and the ENIDs in the same way. They tend not to have access other than in a peripheral way. Therefore, laying it out in a clear way, as in this paper, is valuable for such organisations. Often, there are many knowledgeable people, but they are in a specialised area.

Personally, I think this is a valuable document.

Mr Smerald: This question is for those who are on the working party. I am thinking about this as a joined-up process. I was on the GIRO Committee and I voted for the paper and I have read it. I was struck by how much individual insight was coming through in the different sections.

What I like about the paper is that it could not have been written by one person because there was so much to know. This begs the question: can only one person do all of it? But also, having seen or lived through it, did it feel like Solvency II is joined up in the end? You could see that it is complicated.

You have to do the 20 steps, whatever they are, but I can see my way of getting through them all. Or are there parts where you think that it may be hard to navigate between three steps out of the 20? If you group the 20 into three sub-groups, does it work? You cannot tackle it all but you can tackle three big chunks. So, do you have any insights, having lived through it? What can we expect when we try to do it by ourselves in practice?

Mrs Dreksler: Yes, I think one person can do it all. I think the challenge in writing a paper is thinking about everything and remembering to incorporate it all. You understand the business for which you are trying to do it. You just have to make it work for your business. You consider things but you do not necessarily have to write down every single one.

There is a lot of documentation, though, so, realistically, it must be tough if you are a one-man-band.

When I have done it, it has been a team effort. I think you can understand it all but it may be hard work to get your head round everything. For example, I remember sitting the general insurance examination years ago where discounting reserves was theoretical. You learned the theory but you never thought about how you would do it in practice.

Having to work out in your head what will work – what is technically robust – is challenging. However, I do not think that it is beyond a single person to do it and to understand all the different aspects.

Having said that, I am glad that you appreciated the collaborative effort. I encourage everyone to get involved in a working party. You learn such a lot. It is to be hoped that you produce something that will help others as well.

Mr Kirk: I wanted to bring out that someone can do it from start to finish, but often reserving gets undersold. It is easy to forget that it is difficult to do reserving properly. Sometimes, maybe, in the profession, and certainly outside of the reserving spheres in companies, people can think reserving is easy and therefore it is overlooked. It could partially be because we have been in a cycle where there have not been so many reserving issues in the last few years. Corporate memories are quite short term.

As it turns out, we are one of the only working parties to come up with practical solutions to what is the fundamental calculation under Solvency II. How long has Solvency II been in the making? 18 years? Maybe not that long but it has been a long time. There are so many different discussions on the capital modelling side where the fundamental technical provisions calculation is overlooked. That is something which I think is to the detriment of reserving and I do not think that there is enough recognition of the difficulty involved. As somebody said, if it makes boards think and realise that there are many considerations and that people doing the reserving should be given credit for it and given more air time, that is one good thing that will come out of Solvency II. I do not think all members will necessarily read our papers, but if one of the consequences of Solvency II is to make boards think and to recognise reserving more, that will be a positive outcome.

Mr Smerald: So far, documentation has been mentioned as a particular hurdle. I imagine there are more. Technical Actuarial Standards (TAS) have not been mentioned in this context. That is making people worried. Are we doubling the time to determine the technical provisions? What sort of change in time are we expecting or experiencing from doing this? What are the other things that seem relatively straightforward before you start and then turn out to be difficult? You mentioned discounting as an example. How many difficulties are there where it is easy to underestimate how much time and effort and perhaps collaboration is needed?

One of the problems with Solvency II is that it makes us talk to other people and dig in more when we did not before. That takes time.

Mr J. Instance, F.I.A.: I heard TAS mentioned so I thought that I had better say something. Boards are responsible for provisions under Solvency II and under GAAP, and actuaries provide them with information to enable them to make decisions about what those provisions should be.

I think this paper is to be commended in terms of thinking about the issues. Communicating these issues to boards is important. One of the slight concerns that I have is this. Andrew Smith chaired an interesting forum 2 or 3 weeks ago when Andy Haldane talked about risk-concerning models. There is no single best estimate of a non-life provision. There will be a range of answers. There will also, probably, be a range of models.

It is important that the reserving actuary, the reserving specialist, points out the risk of uncertainty to the board so that they can make good decisions.

Mrs Morgan mentioned Paul Sharma’s view about margins being used several times. When the Board for Actuarial Standards was developing the TAS, they wanted some explicit discussion around prudence. Boards need to know what the prudence is in provisions so that they can count out the number of times it is used.

Another thing that we need to think about is that discounting is definitely coming in. It is in Solvency II. It will be in GAAP provisions probably in 2018–2019. Thus, we have to become accustomed to discounting and thinking about the cash flows underlying provisions. We need to talk to our users and boards about the uncertainty in those cash flows, both the amounts and timing. Some consider that TAS have caused reserving actuaries issues. We have recently completed a review of TAS. One of the things that came out of it was that they have not caused a significant increase in cost.

Mr D. Simmons, F.I.A.: As the paper helpfully reminds us, the technical provisions are made up of three components: the claims provision, the premium provision, and the risk margin.

I often have a problem with the risk margin. I can understand the claim and the premium provision – how understanding all the components of the cash flows, and when we think they might arise, is going to add value to the clients that we serve. But I struggle to see how the risk margin helps our firms run their businesses, particularly when you look at the formula of the sum of discounted solvency capital requirements (SCRs). You try to think how that will equate in the real world to the difference between what you would have to pay someone to take the liabilities off you against the best estimate.

If the working party struggles to see the benefits of the risk margin to firms, should we as a profession be saying that for risk margin we will do what we have to but no more, but put our efforts, thought, and intellect into the cash flows of the best estimate part, either claims or premiums.

Mr Wheatley: The short answer is, “Yes”. I think the risk margin calculation that most people carry out is mechanical. I think that there is a strong argument that it bears no resemblance to the principles underlying Solvency II. Probably, anything that is more theoretically robust may give a higher number. Focus on the cash flows will help the organisations. And the profession could do much more.

As I said earlier, they have a get-out-of-jail-free card with the yield curves being low. Certainly, in the commercial market I struggle to come up with any examples where I think, yes, those are good representative cash flows.

The Chairman: Going back to one of the points Mr Smerald made earlier on, how much longer does this exercise take to do than the existing reserving. How much extra work is required? I would like to hear from practitioners who have gone through this exercise as to how much extra work was involved compared with standard GAAP reserving.

Mr Paul: I was at an event a week ago where people sat round tables. I was at a table with six or seven financial controllers. They unanimously said that they were going to have to do November cut-offs. This was partly fear of the Quantitative Reporting Templates rather than a discussion about Solvency II technical provisions. I was a bit taken aback by how strongly they felt.

As I say, they were financial controllers. They were saying it because they were relying upon other people, people like us, and we were all probably saying this is terribly difficult to do.

To go back to my previous point, I still think that people are doing the GAAP, and then doing another layer of calculations after the GAAP reserving. It seems to me that we will have to get away from that. Companies will have to do everything at once and produce two numbers in a simple process.

They were talking then about the adjustments that you would need if there were a significant event during December. It seemed to me to be making it more and more complicated unnecessarily.

Mr D. Martin, F.F.A.: I am a pensions actuary and so I cannot give any examples from my own experience on Solvency II because so far pensions actuaries have not been subject to doing work on Solvency II. But that may well change.

The main reason for my being here is that I chair a committee of the Groupe Consultatif, which looks after professionalism and standards and has spent some time on the standard for the actuarial function holder’s report. That work is still in progress. One of the problems has been formulating it alongside EIOPA, which is producing all the detailed rules.

There has also been some discussion about whether there should be any other standards produced by this European actuarial organisation.

I should therefore be interested to hear more this evening about communication. Although there has been some discussion about it already, I should like to hear whether there is an issue about actuarial function holders’ reporting, whether anybody thinks that the draft standard that we have produced is of use, and whether any further standards might be of assistance.

Mr Smerald: I am heading up a working party that is starting to look at communication in report writing – the good actuarial report.

One of the themes that has come out is that context is generally missing from actuarial reporting. We just provide numbers or show a chart and say what we have selected. We do not tell the story about what is going on. Personally, I think the direction of better reporting is providing that context and helping people appreciate the story of why the figures or charts were chosen, and where they might go.

Mr T. R. Webb, F.I.A.: I will speak as another pensions actuary with a little practical experience in Solvency II within Lloyd’s.

I am intrigued by the dichotomy that I experienced. Will it take longer? Yes, probably in the short term, I would perceive. From my involvement with it, it seems to me as though the main reason it will take longer within Lloyd’s is because people have not collected the data in the right format in the past. Once it is all re-formatted, it is just a case of following a process through.

My general concern with Solvency II is that it is dumbing down the input from professionals such as ourselves. We are overseeing, and maybe communicating about, the process. But the process is becoming more and more mechanical, more and more prescribed. Indeed, if you look at the spreadsheets for SCRs and so on, there is no scope for input. I wonder whether there are any views on that. It seems as though we are being squeezed out from adding value and just doing a little communication at the top end.

The Chairman: To add to that, as I was looking through the paper, I realised how many assumptions at which you are meant to look. I already look at many assumptions when doing standard reserving. There are more to consider when you are determining technical provisions. Personally, I find it quite difficult to look at every single assumption afresh. It is tempting with some assumptions to think “I looked at that last time and it seemed okay then, so I’ll leave it alone now”.

My concern is that, when technical provisions are prepared, there may be a lot that goes through almost on the nod. To keep it fresh and to make proper decisions and selections about all of the assumptions, I think, could be tricky – certainly if there is only one person doing it. If there were a team of people, it may be easier.

Maybe in that respect it does dumb it down. Maybe it gives a false sense of exactitude.

Mr Newman: To address the points that have just been made, I think it takes longer to do Solvency II technical provisions than it does to do GAAP numbers, and that is possibly because people did not necessarily do cash flows for expenses. It was a broad-brush approach. Now we are doing it properly or, to be more precise, more precisely, rather than necessarily more accurately. So it takes longer.

There is a flipside to the issue raised about revisiting or not revisiting the assumptions you made last year. For example, in consulting environments the new junior gets the data and does projections on it and someone else looks at it and revises the projections and then it comes up to you. You have a look at them again.

You end up with exactly the same numbers or you are making exactly the same corrections and judgement calls as you did the previous year. Actually, what you have done is spent a lot of effort not really getting anywhere. I think that there is a case to be made for “If it ain’t broke, don’t fix it” and if there is no reason to change it, then do not automatically go and do it again from scratch.

In regard to the idea that we are adding less value and just communicating more at the top, I would argue that communicating more at the top is adding more value.

Mr Courchene: I want to follow-up on a couple of the points that were just made. They seem germane to the chapter on validation in the paper.

It seems like it will be much more work to produce a validation report that is compliant with the validation standards. There are a significant number of assumptions that, as was so aptly put, were looked over recently and are not expected to have changed.

One of the benefits of Solvency II and the validation standards is that these assumptions should be somewhat documented in a way such that the new junior who picks it up can begin to understand the thought process involved in arriving at the assumption that was made.

One challenge, of course, is that you cannot prove whether many of the assumptions are good or bad. You are in this situation where the best you can do is to say: “I have made the assumption. It seems like it reasonably reflects our observations”. That is a challenge.

A second challenge that I see is when a validation report shows a number of paths that did not lead to the final conclusion. I wonder if we, as actuaries, are putting ourselves in a position where a higher level of scrutiny from lawyers involved after, let us say, insolvency occurs, can pick up a validation report and say, “You quickly went down this path and gave up. Why did you not go further?”.

Perhaps because we feel we are opening ourselves up, the validation reports that have been produced in the past have not been to the level that is expected to occur under the Solvency II regime.

The Chairman: Bearing in mind the time, I am going to call upon Mr Orr to close the discussion.

Mr Orr (closing the discussion): It is fair to say it is early days with Solvency II: we only now have what we might call a definite date for the start of Solvency II, 1 January 2016, with transposition of those powers into international regulation from March 2015.

I am not talking on behalf of the regulator tonight: I am talking from a personal position as an actuary.

I think that it is fair to say, if you look at some of the things that have been agreed within Omnibus II, in particular, the transitional arrangements and the idea of transitioning to a new regime with some 16 years in which to make those transitions, it is going to be a new experience for all of us.

Certainly, I think that we should not assume that we know all the answers to all the questions immediately from day 1.

What has been great about this paper has been that people have said “Okay, let us talk about this”. There has been strong collaboration within the group. That is largely why it is such a good paper and a good experience for the people who were engaged.

I was lucky enough to talk to some of the people that were on the working party. I remember that the working party had some important principles put in place to allow shared working, with some of the initial thinking being captured in worked examples in spreadsheets. We agreed: “We do not know the answer. We are not going to produce guidance but here are some thoughts”.

I wonder if one of the challenges for the audience in discussing this paper is that the paper is trying to grapple with a large number of issues.

The paper introduces a number of new challenges and there are three in particular that I want to mention. Those of you who are close to discussions within the regulators will know that ENIDs are a topic close to our hearts. People should be thinking more broadly than just the data that is in front of them. Comments were made about a mechanistic approach and its dangers. Data is limited: you only have what you have observed.

There is a theme of moving towards objectivity and thinking about what are the actual risks, recognising that there is a great deal of uncertainty. As for discussion about validation turning into a purely auditable calculation process, I do not think that is really what validation is about. Validation is doing the best job you can in the time available. We have to remember proportionality. We have to remember resource constraints.

We need to recognise that all that we see is not all that there is and ENIDs are an important part of that.

The second issue I wanted to mention was expert judgement. The fact that we end up exercising expert judgement is entirely appropriate. It is also right that actuaries are being asked more about why they make a particular decision. It is not because there is a sure-fire justification for every decision that we make, but thinking forward to that day in court, say, it is useful to think in advance about the lines of argument that we will use to justify our decision making. How did you make that decision? Why was that a good one? Not “the right decision”, because you cannot know the right decision when you are looking to the future and the unknown. The risks are unmeasurable and unobservable. Risk is just something that exists in the future.

Another area that I wanted to highlight, which might be a little controversial, is the point made about risk margins.

The risk margin does look like something of an expedient compromise. There is no way that it can reconcile to the actual price at which you could pass on those liabilities. So I wonder if that is just the starting design requirement for Solvency III.

There were two other points mentioned during the discussion that I felt, as a profession, on which we would do well to ponder. One of those is our reliance on established methods. Established approaches are not necessarily the truth but accepted practice.

We were talking about technical provisions, perhaps, not being quite as sophisticated as we thought when we first saw the designs. People use traditional methods. A dominant traditional method is use of “incurred” data and the Bornhuetter–Ferguson method. If actuaries have accepted this, and there is evidence to suggest that they have, this is established. They use the incurred approach, and they begin to pay less attention to the “paid” data, which, in many ways, is less subjective. We will be exposing ourselves to margins being entirely used up and perhaps even taking credit for them more than once. We might find the experience within more recent ones, where questions around data and the credulity of actuaries, does expose us to greater risk.

Mr Kirk asked about whether we have perhaps been ignoring the most valuable thing that we do as general insurance actuaries, which is reserving. I share that view. I think we have, by looking at capital modelling for as long as we have, exposed ourselves to the risks of losing sight of the actual foundations and the real value that we can bring to businesses within the reserving process.

The final point that I want to make is in relation to the motivation for Solvency II. Believe it or not, there is quite a lot of excitement within the regulator that this negotiation has moved to a point where 1 January 2016 is a real possibility. Three areas were flagged up. One is a single market for insurance across Europe and the wider economic area. Another is growth, which is very much a theme at the moment when we consider that we are still in, it is to be hoped, the tail end of a financial crisis; and, third, policyholder protection within that. We should, as we do at the PRA, think about what objectives we are trying to serve.

A key element and purpose of Solvency II is policyholder protection. The best way to achieve that is getting businesses to understand, own, and manage their own risks. If actuaries play a role in helping businesses to understand the risks within their businesses and to better manage them, then that is the outcome for which we should be aiming.

As a regulator, we know that we should be careful not to create agency risk, perverse incentives, and unintended outcomes. If people look at the risks within the business and help businesses to make good decisions, couched in uncertainty but good decisions informed by risk, we would all do better in the long run.

ICAS had that as a principal aim – better managing risks – Solvency II also has that as a principal aim.

The Chairman: Thank you very much. Sue, would you like to make any final closing remarks from the working party’s point of view?

Mrs Dreksler: I am delighted that the paper has been so well received. When you are involved in the detail, all you see are the things that you have not done and the things you would have liked to have done had there been more time.

I agree with the fact that what we are trying to do here is fundamental to getting Solvency II to work in practice, and perhaps it does not receive as much attention, or has not had as much attention, as it should have had.

I hope you will find the paper useful, especially if you have never looked at this before and need to get up to speed, for the wide range of considerations that you have to take into account when you come to calculate Solvency II technical provisions.

Having said that, I think that the debate today has shown us that it does not cover everything and there is more for us, as a profession, to do in terms of thinking around the subject. The job that we ought to do for our employers is to make sure that the users of this information, be that a regulator or the board, are as well informed as we can make them in their decision making. Communication of uncertainty is important, and I think it is something that we can improve.

The other point that came out for me was the point about validation. This is something that has concerned me right from when I first started to chair this working party. I think that we can do more. This is about giving ourselves comfort, as much as everyone else, that what we are doing is reasonable. It worried me that we might get into a habit of not looking at things from one period to the next. I know that there is a danger in doing that.

Some of the things that we wrote about in the paper in terms of validation were sophisticated and will probably frighten some smaller organisations. But thinking about how you can use your modelling, and so on, to prod yourself to question things, and to make sure that you have proper “actual versus expected” calculations, even though, inevitably, they are going to be approximate because you have to adjust them, is important.