Mrs J. E. M. Curtis, F.I.A. (Chair): I would like to extend a warm welcome to our guests this evening: Ken Kneller Chief Actuary, GAD, Scotland; Christine Scott, Director for Charities and Pensions, Institute of Chartered Accountants in Scotland; John Hibbert, co-founder of Barrie & Hibbert; Zahir Fazal, Director, BES Trustees plc; Dr Colin Holmes, Barrie & Hibbert; and Dr Steven Morrison, Barrie & Hibbert.

I would also like to extend a very warm welcome to Falco Valkenburg, Chairman of the Pensions Committee of the Groupe Consultatif Actuariel Européen, who will be chairing the discussion part of tonight's meeting.

The Chair, Mr Valkenburg, Mr C. J. Turnbull F.I.A. of Barrie and Hibbert, introduced the papers in essentially the same way as at the London meeting. The content has not been duplicated – the discussion of the London meeting precedes this discussion in the British Actuarial Journal.

Mr F. R. Valkenburg: I should like to open the floor for technical questions. We will postpone the full debate until later, until after the second presentation.

Dr C. A. Donnelly, F.I.A.: About your assumption that the employer will not make good a deficit only if the employer itself defaults. Are there scenarios where this may not be a reasonable assumption? One example, might be where the funding level is so low that it is simply too big a deficit for the employer to fund? Another example might be where the pension scheme may be too large relative to the size of the employer at some point in the future?

Mr Turnbull: In the work I basically assumed that the pension fund is a creditor of the sponsor, just like any other. Therefore, it essentially has a legal right to the funding to the extent that the company can afford it unless some other arrangement has been made.

I imagine that this is an issue that could vary quite significantly in terms of the geographical jurisdiction of the pension fund. It is certainly an assumption that makes the work easier because you can directly use market prices of corporate debt treating the pension fund exactly like a creditor. It is an interesting question whether a further generalisation of that assumption is required for it to be appropriate in all cases.

Mr M. Wright, F.F.A.: I work with a number of large last man standing pension schemes, which are unusual and specific to the world of pensions. You can have hundreds of employers potentially dependent on each other, and with interdependent sponsor covenants. Would your approach be extendable to that scenario?

Mr Turnbull: Yes. I think it would. Again, certainly in the corporate case, where we can attach some market-based default cost or probability for each of these different sponsors. The thing you would have to do is make an assumption about the correlation of the events across the sponsors. In principle, the framework could easily extend to adopt that, albeit more assumptions would be required to get an answer.

Mr J. Hastings, F.F.A.: I wondered whether there were circumstances, where a scheme is in deficit, the deficit repair contributions are insufficient, but nevertheless this scheme would not be breaching any Pension Protection Fund (PPF) limits which would cause it to move into assessment. Therefore it might be able to manage its way through a deficit and you would not want the scheme to be put under pressure that would put the company into liquidation if it is capable of being able to manage its way through the deficit, and maybe give it latitude for some time.

Mr Turnbull: I think this all comes down to what question we are trying to answer, what information we can get from these market-consistent valuations, and what actuaries or policymakers do with that information. It is difficult to say a particular thing about what we view as an objective measure of the security of the pension fund or at least the sponsor's ability to fund that pension fund.

I am sure there certainly would be cases where it would be possible that the sponsor may recover, but equally there might be a probability that the sponsor will not recover. I suppose it is all down to the level of risk with which our policymakers are comfortable. It is ultimately a matter of policy to decide about what trigger points or action should be prompted by the information.

At this point Mr Setchim (non-member, PwC), Mr H. E. Piperdy F.I.A., and Mr C. A. Sword (non-member, PwC) introduced the PwC part of the paper. The introduced the papers in essentially the same way as at the London meeting. The content has not been duplicated – the discussion of the London meeting precedes this discussion in the British Actuarial Journal.

Dr Donnelly: The Swiss Solvency Test defines security as another insurance company being willing to take over an insurance company, which means a (market) consistent valuation. Barrie & Hibbert value the employer covenant as a European put option that the employer sells to the pension scheme, give or take some default probabilities. Then the value of that European put option is placed on the balance sheet of the pension scheme. If you agree with this financial economics approach, the company should be willing to give the value of that European put option to the pension scheme. If they were willing to do that, we would have a consistent and more objective way of comparing the funding strength of pension schemes

I feel this range of figures is just a way of fudging the situation to improve the funding position of the scheme rather than trying to come up with a consistent framework that we can use to compare pension schemes within the UK. The way that pension liabilities and assets are valued should be comparable across schemes.

Mr Turnbull: In terms of the methods the two papers developed, basically they gave the same answer. The B&H results would give the same answer as the PwC paper.

Essentially, we assume that the sponsor made good the deficit. The only difference between the two methods is in the B&H paper we are allowing for the possibility that the sponsor does not pay into the fund tomorrow and before they do get round to paying they go bust.

In terms of the put option point, I think it is an interesting point that what the market-based approach does is give an indication of the cost to the pension fund itself to hedge the risk of that sponsor default event occurring. So something you can do with these numbers is to understand how that default risk can be taken away from the pension funds subject to other assumptions in the valuation process. That is certainly technically output from the modelling process.

Mr Setchim: We will have come across, I think, as being quite pragmatic in our approach. It is a pragmatism grounded in a lot of experience of seeing how sponsors cope with the funding of their pension schemes.

But the big point that struck me, which will be apparent if you looked at Section 1.1 of our part of the paper (the schematic of the holistic balance sheet), is that, in the new world of EIOPA, what we currently refer to as technical provisions move up to something much closer to the buyout level of funding, and you do not stop there. You put some risk buffers on top, what I describe to colleagues as the Rumsfeld layers or the unknown unknowns. (I know it is a more technically precise concept than I have just described!)

The commercial animal in me does rather assume that if a sponsor was obliged to fund something to a higher level than buyout, then they would sell out as soon as they possibly could to a friendly insurance company who would be willing to take it on. The big difference, of course, is that defined benefit (DB) schemes are not insurance companies. That is recognised by EIOPA. The sponsor covenant asset which is intended to offset these risk buffers, is a rather inchoate concept. In English law it is something much more than the contractual commitment under the contribution schedule. There are some territories, including our own before pensions law changed around 2006, where people could walk away from the promise. It was more of an aspiration than a promise. Now the section 75 debt, which crystallises on the insolvency of a sponsor, makes it as close to a promise as you can obtain, albeit in an unsecured way.

What we have sought to do is to find a methodology that recognises the long-term capacity of the employer to meet those obligations until a scheme reaches self-sufficiency which is defined as no longer requiring the sponsor covenant.

What we have done, with help from valuation colleagues, is to suggest that it is possible to do that using tried and tested techniques that are being used daily. The business valuations that Mr Sword and others give are based on professional judgement and built on sound techniques.

That is all rather different from asking what amount you put into the fund in order to treat it like an insurer.

Mr Sword: I think there is a practical perspective. For many of the companies with which we are dealing, it would be difficult to derive the required inputs on an option-type basis.

The approach we are taking is market consistent, albeit with a degree of judgement involved, in terms of the assumptions that go into deriving elements like the discount rate and the multiples.

It is not that easy to find appropriate data in the marketplace. Thus it is very difficult to avoid using professional judgement.

Mr Hastings: When I think of the economic value of a company, I think of it as its replacement cost because, if a company is over-valued, I could build a similar company and compete with it. If it is undervalued, it should be able to raise its margins to bring them in line with economic cost. My question is in a ‘deals context’. Do you think that the way you value a company using EBITDA multiples equates to that replacement cost or is there any risk that you are actually imputing an investor risk preference which overvalues or undervalues the company on that basis?

Mr Sword: I think it is important to look at a valuation from as many different approaches as possible. So, if you were to rely solely on EBITDA multiples, for example, there is a risk that you could be just incorporating current market sentiment and investor preferences to a certain degree. Looking at other approaches, such as the asset approach and discounted cash flows, forces you to say “Economically, can the business support this valuation?”

There is a degree of judgement around the valuation. I have occasionally decided that the market is not providing a sensible valuation. That can work both ways with the market being too pessimistic and too optimistic about companies. So, looking at it from various different angles allows you to make that assessment. Replacement cost in an economic context involves replacing not just the tangible assets of the business but replicating its market positioning. That is something that forward-looking approaches are designed to do. For a lot of businesses it would be very difficult to sit down and just replicate them from scratch. Their market position has taken a long time to build up. The best way of assessing the value of such an intangible is by capturing its. That is also how people look at it in a deals context. You are right to say that you have to take a step back and ask what is the market telling me, and do I believe that to be correct?

Ms M. Bruevich (student): The Barrie & Hibbert approach is really flexible when a company credit rating drops, while the market approach of PwC is based on past information. If the risk profile changes, or credit rating drops, it is not clear how you adjust your valuation with the sponsor covenant in this case. I can see the advantage of the flexibility of the Barrie & Hibbert approach.

Mr Sword: I think it is worth saying that whilst the multiple approach may be based on historic financial information, it can also be based on projected financial information. If you are using public companies, for example, we look at their forecast earnings, not just historic earnings, to understand what the market is assuming about the growth rate for business by comparison to the growth rate that might be applicable to the company at which we are looking. The changes of the nature that you were discussing in a market valuation sense would be reflected in the multiple that is being applied to those earnings. If, for example, a business has become more risky, or its earnings have declined, or it is changing its credit rating position, then that would affect the multiple that is applied to those earnings. If we were looking at it from a discounted cash flow perspective, it would also potentially affect the forecast future cash flows of the business or the volatility of those cash flows which may then also feed into the discount rate.

While the valuation may use historic information to some extent, it is very much a forward-looking exercise. Credit rating agencies are very often trying to do the same thing. A lot of the situations that they face have lots of historic information. The degree of forecast information that they have will vary depending on the company at which they are looking. Credit rating agencies have been criticised in the past for not reacting quickly enough to market information. In that sense, I do not think credit ratings are particularly more reactive than the approaches that we are suggesting, assuming that they are applied properly.

It is also worth mentioning that you can observe the credit rating of large public companies which have issued debt. But many of the companies that we have to look at do not have an assessed credit rating, so we have to develop a synthetic one to evaluate their credit risk.

Mr Setchim: The key question is the purpose of the holistic balance sheet? The answer is not entirely clear.

There was a really interesting question earlier about comparability and consistency. We completely understand that in the insurance industry, where people move their books around through the Part 7 process, you really need to know the effect of moving the liabilities around.

As an accountant, we put different figures in the accounts for pension scheme liabilities because the purpose is to enable investors to compare results, hence there has to be a comparable way of valuing them. If you are looking at the pension promise from the standpoint of the member, which is what is stated as an objective, then I believe that you must have a methodology which enables a view to be taken about a specific sponsor.

Holistic Balance Sheet (HBS) was defined as a supervisory tool. The results of the Quantitative Impact Studies (QIS) for the UK, that the Pensions Regulator has produced, really do not give a supervisor any help at all in targeting their resources. I know roughly what proportion of UK corporates can afford their pension schemes, and what proportion cannot, and that view is not derived by the QIS methodology. I am interested in security for scheme members with specific sponsors and our method is designed accordingly.

Mr Valkenburg: I should like to open the general debate and invite you all to participate and comment.

Dr B. T. Porteous, F.F.A.: The two papers are very different. I would say that the B&H paper is “high rigour”, whereas the PwC paper is “high realism”. Having the two papers side-by-side worked really well.

I should also like to give EIOPA and the EC a little bit of credit. I think they have opened up an important debate. The security of defined benefit scheme members is not often discussed.

This difficult issue needs to be addressed. At times the debate around this topic can become quite emotional, as we have seen in the press recently, and, most definitely, refinements are going to be needed going forward.

I have an interest because I and a couple of colleagues did some research on this topic recently. We looked at a real defined benefit pension scheme and we quantified the economic capital for that scheme. We found that the economic capital requirements were around 50% to 100% of best estimate liabilities. They were rather large, therefore. The reference for this work is: “Economic Capital for Defined Benefit Pension Schemes: An Application to the UK Universities Superannuation Scheme”, Journal of Pension Economics and Finance, Volume 11, Issue 4, October 2012, pp. 471–499. Irrespective of the arguments about the technicalities, I think that we can conclude that this work shows there is significant risk in these schemes that affects member security.

Moving on to the two papers. I found that in the Barrie & Hibbert paper the justifications for the market consistent and risk neutral approaches could have been fuller and more rigorous in defining what these concepts are and why they are appropriate. The paper makes the assumption that government bonds are risk-free throughout. One thing we have learned from the sovereign debt crisis is that this is probably not true, at least for some European Member States. The physical rationale for the model in the Appendix was not really given. I accept that this may not strictly be necessary for the purposes of illustrating a relatively simple example. However, it would be helpful to see models with genuine physical rationales that can be tested against data.

I wonder how plausible it is, in paragraph 2.2.1, to assume that default rates depend on all relevant explanatory variables only through a firm's equity return. That assumption is a little bit heroic and is one which, in my view, could be tested against market data.

Continuing on the Barrie & Hibbert paper, in terms of the saleability of this material to stakeholders, I wonder what a typical pension scheme trustee will make of phrases like: risk-neutral has no “direct economic interpretation” (2.2.1) and that illiquid liabilities have “market prices which are not observable” (2.4). I also wonder about the coherence of mixing of real-life management actions and risk-neutral cash flows (2.4.1). There is a challenge there in explaining this to pension scheme trustees.

On balance, however, I think that the Barrie and Hibbert paper is a very good paper. This is probably the first time that the sponsor covenant has been valued with real rigour. It is a very good starting point. But I do see some stakeholder management issues going forward.

Coming to the PwC paper, I found much on which to agree in principle. The economic definition of value seems sensible (Section 3.1). I agree that there are times when market pricing is dysfunctional and can be misleading. We definitely need rigour in doing this work. However, we need an appropriately flexible approach that is able to take account of market dysfunction (Section 3.1).

The PwC interpretation of market consistency in Section 4.2 felt legitimate to me, even although it is a lot less formal than the Barrie & Hibbert definition. However, I do feel that we need more rigour for valuation purposes than the simple market ratios we heard about earlier on this evening. I wonder whether such market ratio approaches are really too simplistic.

A few closing observations: these comments are probably coloured by my Solvency II experiences over the years. Firstly, in finance, we still do not seem to have any real consensus on how to value and discount uncertain financial liabilities. I am thinking very much of Solvency II here where we are debating concepts like illiquidity premiums, matching adjustments, market consistency and so on. Secondly, I think that we do need valuation rigour and it needs to be appropriately flexible. We must avoid dogma and partisanship, which is one of the pitfalls into which Solvency II has fallen, especially on market consistency. We need to be a bit more pragmatic and sensible. Finally, I feel that we have an interesting journey ahead. The defined benefit pension scheme genie seems to be emerging from the bottle and we should not forget scheme member security. I do not believe this issue is going to go away.

Mr Turnbull: Thank you for those considered comments. I will pick up on a couple of the key observations that you made on the B&H paper. I agree on the point about exploring and defining your market-consistent techniques and pricing measures. The paper is very light on that topic. We had to make a decision early on about how much general education one needs to put in the paper and how much we had to assume was pre-existing knowledge which was well-documented elsewhere. Maybe we moved a bit too much to the latter end of the spectrum.

In terms of the difficulties of interpreting these results for pension fund members I absolutely take the point that was made. Ultimately, how what we are trying to say is “given the way in which the sponsor is committed to making good this deficit, and the quality of their credit, what is the market price that someone would independently pay to receive those sets of cash flows?” That is fundamentally the number at which we are trying to arrive. At the heart of all this is communication to members around the security of the benefits. Clearly, the pension protection fund, etc, will complicate that picture further. It is not completely clear to me that communication to pension fund members about security of benefits is working particularly well in the absence of market-consistent techniques today. It is an interesting challenge for policyholders and the profession regarding the role they will play in that type of education.

The technical point about real life management actions and residual cash flows is an important point. It is one that can cause a lot of confusion. Actuaries have had a lot practical experience of this issue over the last 10 years in the context of the valuation of with-profits policies and similar types of liabilities. For these liabilities there are a whole range of potential or real future management actions that are incorporated into the valuation model. This approach takes place today in valuing with-profits around Europe and variable annuities in North America and Asia.

Mr Valkenburg: The question for you is: a nice flexible method is fine, but is it too simple, perhaps?

Mr Sword: I think the question is how much is enough? For certain companies a relatively simplistic approach may be sufficient in that it is obvious that the answer to the question is that there is plenty of headroom. For other companies that answer is not going to be so obvious.

I think it is worth saying that, for the purposes of the paper, the approaches and examples that we have set out are necessarily relatively simplistic. You can apply those approaches in a considerably more detailed and complex fashion if the need arises. For example, if you are looking at discounted cash flow, we can use things like Monte Carlo simulation of costs and revenues to understand how a business's profit margins may move. You can have multiple different scenarios. We can build quite complex models around the business. You can apply those approaches in varying degrees in terms of assessing point estimates and ranges around those estimates. For a lot of businesses that is not necessary or required.

There is also a practical question as to how much information will be provided to pension schemes by the sponsoring employers. Our experience is that quite often the sponsoring employer will provide quite a lot of information which enables quite a lot of analysis.

Mr Z. Fazal (Chair of PSC): As several speakers have already suggested we need to understand the purpose of a holistic balance sheet. Is it for funding? Is it for supervision and risk management for the regulator? I have no doubt that the actuarial and accounting professions can come up with an acceptable basis and will have many technical arguments. I am not sure that really achieves anything. I do think this central question of why the holistic balance sheet is being prepared needs to be answered.

The point was made, I think extremely well, about security of benefits. Is that what we are trying to measure?

There is a danger that we are becoming involved in the “how” without having answered the “why”.

Mr Valkenburg: What would your personal answer be to this question?

Mr Fazal: I will now speak as a trustee. I just despair at what we are trying to do. Trying to communicate this to members is less than objective. I do not think we have any chance of communicating this to members. I also wonder who will be responsible for preparing the holistic balance sheet. I am guessing, but I expect it will need to be agreed between the trustees and the employer.

There will be two sets of professional advisers arguing it out trying to prepare a holistic balance sheet which both parties agree. There is enough difficulty now in trying to prepare a qualitative assessment and agree the covenant, whether it is strong, weak or somewhere in between. I despair of trying to get two sets of professionals to agree precise numbers, methodologies and assumptions.

Mr Valkenburg: I think it is very valid to ask why we are doing this. I think of pensions as different in general from insurance. Pensions are, in my view, more of a social contract between employers and employees. So, there are two main stakeholders who define or agree the pensions promise, including elements such as: how safe is the promise? How will it be funded? How will it be dealt with? What will happen if the fund is in a difficult financial position?

I think that, on many occasions, not only here in the UK but also in other member states in Europe, we have not been very successful in being transparent about this issue. Employers and employees have not been clear about the nature of the real promise. Employees, of course, see it as “This is a guarantee that I have with 100% security” The employers know that this is not possible, but have not been very transparent in making this clear. I think the holistic balance sheet is an invitation for discussions between employers and employees about the nature of the pension promise and how it can be structured.

Mr A. C. Martin, F.F.A.: I am speaking as an advising pensions actuary, a former scheme actuary, an occasional expert witness and also an independent trustee.

I think the two contributions tonight come into two distinct areas. The B&H paper could quite usefully be called a macroeconomic perspective, whereas the PwC paper is very pragmatic and at the microeconomic level.

My conclusion from the B&H paper is that it will be of most value to regulators, our Pensions Regulator, the European regulators, the PPF (and, dare I even say, Dun and Bradstreet). It may also help Her Majesty's government in considering some fundamental questions such as whether the current high level of pension debt repayment is stifling economic growth.

Two small points on the B&H paper: in paragraph 2.4.1, the principles and practices of financial management are outlined. That could be very useful, particularly if it is focused on extreme situations such as the biggest funds, the biggest deficits or the weakest covenants or some combination thereof. There are, in fact, some financial management plans already in existence. They are called business plans and they are being encouraged by our risk-based Pensions Regulator. In terms of taking them forward, I think Principles and Practices of Financial Management (PPFM)s would be very useful. Similarly, in paragraph 2.5.4, the correlation between investment risk and the sponsor is highlighted. I think that will become increasingly important as there are more investments in special purpose vehicles, with asset-backed contributions and indeed self-investment generally. Sadly, the sponsor balance sheet can look totally different in distressed conditions, for example when the security of high street property is at risk because the sponsor cannot sell their products from that high street.

The B&H analysis quite correctly addresses the issue of investment risk. I am sure that the audience will not be surprised by the current de-risking of pensions continuing at a much greater pace.

In terms of solutions, I think the PwC pragmatism is to be commended. I am a great fan of employer covenant assessment. I think, like audit, it should be compulsory. If it were complemented by a duty of care to third parties, i.e. the pension scheme member, then I think that would also be very useful indeed.

My conclusion from the PwC paper is that there are many different individual circumstances. For the 6000 schemes in the country there are 6000 different situations to approach. I would very much commend the approach of sifting out of the troublesome cases in a preliminary process.

Once things do go wrong, our risk-based Regulator, it is to be hoped, will have spotted them, and Figure 4 in the PwC part of the paper is very useful in terms of the demise curve. My experience of problem situations is that not only does debt repayment become difficult, but the value of everything on the balance sheet is re-written downwards, often by a factor of 10.

I would very much commend the pragmatism of the PwC paper to the hands-on user, which I think should be the Pensions Regulator.

Mr Hastings: Addressing the issue of what is the purpose of the exercise, I suppose, in essence, both methods are “average” risk-based approaches. One problem is that, although you can create a probability function for covenant strength, for the members concerned, the outcome will be binary, i.e. members will either get their pension or they will not get their full entitlement. There must be a risk of unintended consequences in the communication of covenant strength to members, particularly of a weak assessment. I agree with Mr Martin when he said that maybe this is more of a tool for a regulator. We would need to think quite hard about any message to be communicated to scheme members.

Mr Valkenburg: It is my task to close. These were very good papers contributing much to the debate. In London, representatives from EIOPA were in attendance. I know that they are taking this very seriously. I do not think that what EIOPA has drafted in the quantitative impact study of these questions is carved in stone. They are open to debate and for change. The debate with them will be ongoing.

It is also important to look at whether a pension is a guarantee, a promise or an aspiration? That is a basic discussion. If you do not know that, how do you put a value on it?

We should look at the issues from a broader perspective than just pensions. From the point of view of members the level of security can be considered as part of a total remuneration package. Pensions security can also mean a very good pension but no job any more.

I very much like the point that in the end we need judgement. It is one of our tasks, as actuaries, to bring judgement back to a very formal approach.

The Chair: Thank you, Mr Valkenburg. I would also like to express our thanks to the speakers tonight.

I also hope you will agree that the profession made the right decision in commissioning both pieces of research. We have only a limited pot from which to spend on research, but we took the decision to ask both Barrie & Hibbert and PwC to do the research and we think that the outcome has been particular worthwhile, given that they approached it from quite different angles. So we say thank you for that.