Conclusion

So what started out as a labour of love has ended with a call for more focused and detailed research. As I said at the start of my remarks, it has been a pleasure to help prepare this paper and I look forward to the discussion.

The Chairman: To lead us in our discussion we have Colin Ledlie. Colin is the Chief Risk Officer (CRO) at Standard Life, and is well known for his contribution to the study of risk management.

Mr M. C. Ledlie, F.F.A. (opening the discussion): I would like to open my remarks by thanking the authors for an excellent paper which I thoroughly enjoyed reading. It gave me great insight into the operation of an important element of the financial landscape in the United Kingdom.

There are two ways in which you could view the paper. Firstly, you could view the paper as a detailed analysis of the operation of the Pension Protection Fund and an opportunity to open up the management of the Fund to external challenge and scrutiny from a well-informed professional audience. Secondly, you could view the paper as a case study in Enterprise Risk Management applied in the context of a complex, distinct and challengingly different type of entity. I will seek to cover both these aspects in my opening remarks but will concentrate more on the second of these perspectives, given my own professional role as the CRO of a large financial services company.

Firstly, the paper provides a comprehensive review of the financial management of the Fund, covering, in particular detail, aspects such as funding, risk and investment strategy. I was impressed by the thorough analysis presented in the paper and the open and transparent approach adopted by the PPF of making details of the operation of the Fund available on its website for scrutiny and review.

To some extent, the Fund has been dealt a hand of cards and a framework in which it must operate. The PPF was established as a statutory corporation and is run by a board that is independent of Government. The main reason for this structure was to ensure the Fund would not have to be underwritten by the Government and, ultimately, taxpayers. I found this surprising, as it would seem highly unlikely that the taxpayer would not step in to support the Fund if it came into financial difficulty, just as happened in the case of the protection of bank depositors in the recent financial crisis. A different assumption in this area might have led to a different financial management approach.

A key difference between the PPF and a commercial insurance entity is that it does not maintain a significant capital buffer over and above a reasonable estimate of the liabilities. An insurer would be required to hold sufficient excess assets to be able to withstand at least a 1 in 200 year adverse event and still be able to meet its liabilities. I contrast this with the PPF which has had a funding ratio hovering around 100%. I wonder how consistent such funding levels are with the objective of not having access to additional government support.

A further concern is in relation to the public's understanding of the limitations of the Fund and, in particular, the financial cap on compensation at an individual level. Are the 90% and £30,000 (approx) limits widely appreciated? Also, is an 80% chance of meeting the financial objective over the next 20 years aligned with what I imagine is a public expectation that the Fund would meet all its commitments as they fell due. I would be interested in any research the PPF has done in relation to public expectations and understanding.

In paragraph 2.12 the objective is stated to be to be fully funded by 2030 with no further risk to the balance sheet at that point. This felt like a challenging aim and is, perhaps, unrealistic.

The authors recognise that there are different ways of viewing the Fund. They suggest that it may be a combination of a credit insurer and an annuity writer. The analogy that came to mind for me was that of a permanent health insurer where the insured benefit was a combination of an insurance component followed by the payment of an income once a claim was made on the scheme. The management of the Fund felt as if it had been developed from the heritage of pension fund management rather than from an insurance perspective. At least in the short to medium term, the characteristics of the Fund appear to me to be much more like an insurance business with the rate of claim inception being the key element of uncertainty. I felt that this aspect was somewhat underplayed in the paper. A particular challenge for the Fund is that the claims experience of different schemes, particularly in adverse scenarios, are likely to be highly correlated. There could be a very severe downside scenario of many employers going insolvent simultaneously just when scheme funding positions are weakest.

Subject to these remarks, I was very impressed by the coherence and clarity of the financial management approach adopted for the Fund.

My second theme is to view the paper as a case study in Enterprise Risk Management. In this context, my first difficulty is that the paper does not set out any particular framework or model for ERM against which it is seeking to implement or measure the Fund. A range of ERM standards and models are available, which would provide a useful focus and assessment criteria. For example, one framework that could have been used is the International Actuarial Association Practice Note on ERM published in 2008. This includes a useful ERM maturity assessment in Appendix 2. The good news is that the approach adopted by the PPF covers the majority of the criteria in that document.

For reasons of personal familiarity and simplicity, I am, however, going to use the ERM framework that my own firm uses to guide my comments on the PPF operation. This ERM framework has five inter-related components: risk culture; risk and control processes; risk and capital models; emerging risk management; and strategic risk management.

First, as a foundation to the overall embedding of ERM, is “risk culture”. Without appropriate buy-in and leadership from the top, any ERM framework will be little more than window dressing. In the case of the PPF, it is clear that risk management is taken seriously with strong and clear leadership and ownership at Board level. This starts with providing clarity on business and risk strategy and on the risk appetite of the entity. This is not a trivial exercise for the PPF, where the unique nature of the entity means a great deal of thought will have been required to agree upon the right approach. The risk appetite of the Fund is expressed in qualitative terms in section 6 of the paper. Having set out clear risk metrics in section 4 of the paper, I was surprised that quantitative risk appetites were not also set, allowing the Board to express its appetite for risk in both qualitative and quantitative terms.

The governance structures of the entity are a key part of the risk culture and are set out in section 5 of the paper. This sets out key board and executive level committees. Something the PPF should think about is whether the committee structures support an holistic view of enterprise-wide risks or whether they risk encouraging a siloed approach with internal control and operational risks being managed separately from investment and financial risks. Following the Walker Review, and the widespread adoption of risk committees, there may be an increasing trend towards risk committees taking responsibility for all risk types.

The PPF has a CRO who is responsible for the management of financial risks. What is not clear to me is the extent to which the PPF has adopted a “three lines of defence” model for risk management that provides clarity on the split of responsibility between the risk-taking first line and the oversight and challenge role of the second line. The paper also makes no mention of the third line of defence internal audit function and the role of independent assurance providers in the detailed operation and implementation of the framework.

The paper gives little clue as to the risk culture that pervades the staff of the PPF organisation. I suspect that this is, perhaps, of less relevance to the PPF than to a large financial institution where a key role of the risk function will be training and development to ensure that good risk management is understood and supported at all levels in the organisation.

The second ERM framework element is “risk and control processes”. There is good evidence of effective control processes operating in the PPF through the descriptions of the investment process. The paper, however, gives little focus to operational risks and the potential costs of ineffective operation of the framework. These are often aspects that actuaries will ignore, but which I believe are critical for a well-rounded ERM framework.

The third framework element is “risk and capital models”. This is well covered by the paper, and usefully a comparison with the Solvency II requirements of an internal model is set out.

The fourth framework element is “emerging risk management”. These are the processes adopted by the entity to identify and then manage new or rapidly changing risks. The PPF includes within its processes the opportunity to review and discuss such emerging risks. Extreme scenario analysis is also an important element of emerging risk management. It is now a regulatory requirement for insurance companies to conduct reverse stress testing exercises whereby they look at how extreme a scenario would need to be to make the entity unviable. I think this is an exercise that the PPF might also find useful.

Finally, the fifth framework element is “strategic risk management”. In the latter part of section 6 there is a good example of the PPF using its risk framework to guide its decision-making to optimise the risk-adjusted returns achieved by the Fund – a key strategic output.

Overall, the paper provides a great example of an entity effectively using ERM tools and techniques to manage its risks and I would again like to again thank the authors for sharing in such detail their experiences.

The Chairman: Everyone here, whether they be members or guests, is encouraged to continue the discussion, which is now open to the floor. I am sure that this will be an interesting and stimulating debate because there is something in the paper for pensions actuaries, for insurance actuaries and for those involved in risk management.

Mr K. Barker F.I.A.: I think the model that has been adopted, which is that of an insurance company, is entirely inappropriate for a government body. Generally speaking, governments do not choose to go down the insurance model route for the risks that they take. So, for example, government buildings are not insured against fires, and so on; nor are schemes, such as the deposit protection scheme, based on an insurance model.

I wonder whether that has led you to a completely erroneous interpretation of risk, and has ended up with you investing in what I think are not particularly helpful investments for a government scheme. Effectively, you have one government body, the pensions protection body, lending to another part of government, which is the Treasury, by the buying of gilts.

I wonder, given that you have a very long time frame, and you have a sponsor who could deal with virtually any economic shock dealt to it, whether you should have done something more useful to society, and invested in equities, preferably through active management.

Mr S. J. Richards, F.F.A.: I am glad I am speaking second because I am going to say something in contradiction to the previous contributor.

I should like to thank the authors for the interesting background paper on the PPF. I found the section on international comparisons particularly interesting. The first thing that struck me was how late the UK was in setting up something like the PPF. Both the German and United States equivalents seem to have been founded a full 30 years prior to the PPF, which led me to wonder what the UK pensioners did before the PPF was set up in 2005.

The second thing that struck me about the international comparison was that the predominant European model seems to be that of a non-life insurer which then buys out necessary annuities with life insurers. In contrast, the US model, which is like the PPF model, is akin to a combined non-life and life insurer that both insures against the deficits and pays the annuities. On paper, either model seems valid, but the history of the Pension Benefit Guaranty Corporation in the US appears to have suffered rather badly due to politics and lobbying, as evidenced, perhaps, by its rather substantial deficit. I was left wondering what the reason was for the PPF following the US model and not the European model.

The European model seems to have a number of desirable features, one of which Mr Ledlie touched on, which is that the requirement to buy annuities forces market prices for risk to be used. It also distributes the risks around the market, around the various insurers that are participating, instead of concentrating them in a single entity. This led me to my last question which perhaps opposes the previous contribution. The UK has a particularly large and competitive market for bulk annuities. I was wondering whether the PPF has given thought to using this for its risk management by buying out the pension liabilities with insurers.

Mr Clarke (replying): I would like to comment on these points. First, I will address the comment that the insurance model is inappropriate because, after all, we have government sponsorship. I think if the PPF were to be sponsored by the government then maybe the insurance model would be inappropriate. That is not the basis on which the Fund is operated. In the paper we say the Fund was set up by an Act of Parliament, but it is actually an independent corporation. It is funded by payment of a levy from defined benefit pension schemes and is not subject to a government guarantee.

As Mr Ledlie suggested in his opening remarks, were that to be the case, then a different financial model would apply. One feature might be that the Chief Investment Officer of the Fund would be able to get away with more risk taking and adventurous investments, and indeed behave like an investment bank, which is, after all, underwritten by government – or at least our recent experience has proven this to be the case.

Equally, those from whom we seek to charge a levy for the protection would be, perhaps, a little less interested in paying up given the fact that, if they did not pay, somebody else would pay on their behalf.

So the basic premise on which the Fund is set up is as a self-funded vehicle, effectively a mutual insurance product for the defined benefit pension schemes that are eligible for it. The premise is as set out in the paper.

Mr Richards mentioned the fact that we were late, or the UK government was late. I cannot comment on that. No doubt members of defined benefit pension schemes that were wound up prior to the PPF becoming available and whose assets were insufficient to fund the benefits would have suffered as a consequence. Indeed, the government more recently than 2005 recognised this by increasing the levels of the Financial Assistance Scheme, which was a PPF forerunner, to be more or less equivalent to PPF levels. Again, I cannot comment on the policy debates at the time.

You raise the interesting question of the insurance model that is prevalent in European jurisdictions and contrast it with that in the US and the UK where a different set of conditions has applied.

I think that it is fair to say that the competitive market for bulk annuity business did not exist in 2004. The emergence of new providers in that market, the interests of capital markets and indeed investment banks, is a phenomenon that has grown during the lifetime of the Pension Protection Fund. Certainly, one of the fears of the legislators at the time was that a requirement to buy out would be thwarted, not just by the lack of availability of private sector resources to take on these risks, but also by the demographics of the defined benefit pensions universe, where there are many very small schemes. So, some sort of aggregation model has to apply, otherwise the so-called competitiveness of the marketplace would not allow the model itself to capture the efficiencies of scale.

On longevity risk transfer, we set out in the paper, in section 4, the scenarios that we run to judge whether a longevity risk transfer-out of the PPF to the private sector is feasible. If the diversified longevity risk that the PPF carries – and we say it is diversified because the PPF carries other risks – is capable of being reduced by outsourcing to another provider, then obviously that is something that the Fund would have to consider. To date, it has not reached that point, but one can imagine a situation in five or ten years, or maybe later, when the Fund is dominated by its annuity risk and is not able to diversify that risk, that a different scenario might apply and the Fund might have a greater appetite for buying the risk externally. The Fund might be very large at that point, of course, so there is also a question of market capacity.

Mr A. C. Martin, F.F.A.: I should like to make two points in the area of strategic risk, the first around the word “compensation”.

The PPF does not suffer – and I use the word “suffer” deliberately – from Section 67, (Pensions Act 1995) and the compensation provided is not guaranteed and can be changed in future. I think that that is a huge cushion for the PPF. We have seen the change in CPI only slightly moving the goalposts. Has it been appreciated yet?

I wonder if future modelling should include, for example, moving the commencement date for compensation to something akin to future changes in state pension age. I think that this would be a positive design feature.

The other area under strategic considerations is the fact that the PPF inherits schemes after the Pensions Regulator has done its work, and the Pensions Regulator is a risk-based regulator. It can act as a significant buffer in terms of the pension schemes coming through. I will deliberately say that they will control the bigger schemes coming through. The smaller schemes, I think, do not hit its radar as much, and certainly are not such an issue from a risk perspective.

Mr C. R. Barnard, F.I.A. (in a written contribution that was read to the meeting by the Chairman): I have read the discussion paper ‘Financial Management of the UK Pension Protection Fund’ with great interest. I hope my comment will add to the discussion and I would like to raise some important questions about the possible interaction in the future between pension schemes and State pension protection schemes.

The European Insurance and Occupational Pensions Authority (EIOPA) published a consultation paper last October on the response to the call for advice on the review of Directive 2003/41/EC (the IORP directive). This paper discusses introducing a Solvency II type regime for occupational pension schemes, which would allow for certain particular features of pension schemes. One of these features is the existence of sponsor covenants and, more importantly for our discussion paper, of pension protection schemes. The call for advice mentions: “The possibility to restate the value of assets in the IORP and liabilities of the sponsoring undertakings into a single balance sheet, including the possibility to recognise sponsor covenants and claims in pension protection schemes as an asset similar to reinsurance.”

My own view of this matter is that I would caution against aggressively including pension protection schemes in a pension scheme solvency framework for the following reasons:

• - Valuations would be less comparable between pension schemes with differing levels of pension protection;

• - The valuation of the pension protection is subjective in the tail conditions that would be expected to apply when the protection should be called upon;

• - The valuation of the pension protection could become a multi-agent problem, as pension protection schemes apply to multiple pension schemes. This is different from sponsor support, where the only other agent to consider is the sponsor itself;

• - Their inclusion could lead to systemic issues in the sense that all the pension schemes covered by the same pension protection scheme would presumably hold lower assets and capital. This could lead to a greater possibility of systemic call on the pension protection scheme in times of financial distress; and

• - The above result, i.e. pension schemes systemically holding lower assets and capital almost implies that the pension protection scheme is a “joint sponsor” of the pension schemes, rather than last-resort protection.

However, should this possibility come to fruition, I would like to ask the speakers and all the participants in this session the question: shouldn't the UK Pension Protection Fund, for example, therefore be required to carry a corresponding liability, mirroring any such assets, in its own balance sheet? And, in general, do you think that potential future claims on the Pension Protection Fund should be already explicitly valued within the liabilities of the Pension Protection Fund?”

Dr B. T. Porteous, F.F.A.: I think you have done an excellent job in managing the risks of the PPF, particularly because you started from scratch with very limited guidance. The one thing that did surprise me, however, was the lack of reference to Solvency II, which has taken a long time to develop – at least 10 years up until now. But it has been running alongside your PPF developments and there is much you could learn from it.

For example, there is more than just the “use” test. There are the systems and governance requirements, for example, which potentially the PPF could use. In fact, in 3.5.9 you distance yourself from Solvency II and say it is possibly appropriate for insurance companies, but not necessarily for an organisation like the PPF.

I think that there is also good thinking on counterparty default risk and concentration risk in Solvency II that you can potentially pick up and use.

Moving on to a related point of the written question, we saw the final advice published by EIOPA last week on the IORP Directive. This now suggests that there is a high likelihood that Solvency II is going to apply to defined benefit pensions within Europe in at least some shape or form. I would be interested in your reaction to this and what it is going to mean for the PPF, and also for the European pensions industry.

My final point is about chart 8.1. It seemed to imply that there was systematic under-pricing relative to a more economic capital-type pricing approach, which passed without comment. I would be interested in your views on this.

Prof R. S. Clarkson, F.F.A.: The last speaker suggested paying more attention to Solvency II. My advice would be to go exactly in the opposite direction, and run away from it.

There was one paragraph that I thought was brilliant: 7.2.6. Mr Clarke mentioned this in his opening remarks. He is sceptical about financial and economic models and feels that some people place too much reliance on them. I had never seen better words that sum up my practical experience: economic models are not infallible.

Paragraph 7.1.8 states: “The value of liabilities at any particular time step is expressed in terms consistent with the contemporaneous market parameters (such as interest rates and inflation assumptions) which underlie the market value of assets.” The chart above this paragraph is Frank Redington's expanding funnel of doubt. This is in Einstein's Brownian Motion paper.

My strong view is that, in financial markets, it is not uniform diffusion over time. From time to time you have violent shocks. In terms of Solvency II, one of the main assumptions is that human behaviour is robotic. It anticipates everything that can be foreseen, and all rational calculations are taken into account. That, I believe, is not the case.

A lot of references that I thought could have been included are not listed. There are none before the year 2000. One particular piece of work which I think deserves more attention in terms of modelling extreme events is David Wilkie's stochastic model.

He follows the robotic method as described in 7.1.8, and I think his is possibly the best model using these general assumptions.

A number of years ago, for the second AFIR international colloquium, I wrote a paper about a non-linear model for inflation which brings in random shocks. I believe that some sort of random shock which upsets the system and generates far more volatility is the best way forward for future modelling.

My assessment of David Wilkie's model could be summed up in just a few words – three, in fact: “far too tame”.

You have unpredicted shocks from time to time. Once they get into the system, government reactions such as quantitative easing, upset the equilibrium even further.

Let us look at more complicated models than we have now to bring in shocks on a random basis. An adaptation of a non-linear model of mine could be of use in setting reserves and prudent regulatory standards.

What I want is that, when assets are overblown and you are at the top of the equity market, red signs will flash. However, the typical financial economics models follow the trend and no one sees anything wrong at the top of the cycle.

I thoroughly enjoyed the paper. My general point is to look at models that bring in some sort of shock and upset the equilibrium balance.

The Chairman: Perhaps I might be permitted to ask a question of my own, which relates to the horizon of 20 years from 2010 that was chosen. I think, from my reading of the paper, that it was based on the fact that we have, at the moment, a substantial slowing down of the DB pensions market, and that people are closing schemes to new entrants, and sometimes also to future accrual for existing members.

However, the Pensions Minister has talked about sharing risk between the employer and the members in ways other than by either the conventional defined benefit or defined contribution ones, where the risk all lies either with the sponsor or with the member.

If there is an arrangement such as that, perhaps along the lines that the ACA was suggesting a year or two ago, or something of that nature, then it might well be that an arrangement that does have an element of sponsor risk involved might be included in the PPF.

This seems to me a political risk which is probably impossible to measure but could well be one of those incidents that might knock some models askew. I was wondering whether that can be factored in, in any way.

Mr Clarke (replying): I will pick up some points relating to Solvency II. I will try to answer your question as well, Mr Chairman. Then I am going to let one of my co-authors, Ian McKinlay, speak on non-linear random shocks in our models.

I think that the first rule of modelling is to be suspicious of models, but I would not like to navigate a world without some kind of framework that models can provide. But I will leave the point to Ian McKinlay.

The written comments that were sent in advance talked about the potential for a Solvency II-like regime to be inflicted upon us in some way or other, and some of the questions which arise from that are those of measurement of any protection cover that you might have from the likes of the PPF or others.

One thing I should mention, and Mr Porteous mentioned it as well, is that this was a call for advice. At the present time, the response from the committee that has responded to the call for advice is really only that, a response, and so there is a long way to go on these issues.

One point in particular that was fed back as part of the response was the need to have a full impact assessment of any change in measures before they are enacted. I believe that the Commission itself has said that it does not want to have inadvertent consequences from its reactions. Nevertheless, for an organisation like the PPF, the question of a Solvency II-type regime for pensions does bring about a bout of schizophrenia. On the one hand, one might have greater visibility, possibly even a strengthening, of the funding regime. On the other hand, if it goes too far, it may cause the unintended consequences of mass entry into funds like the PPF.

That is why I said at the outset that I think this is an area worth further study and I hope some of the quantitative impact study will cover areas such as this. I do not think sufficient has been done at a macro level to understand the relationships between pension funding and corporate survivability.

Mr Barnard also mentioned the question of Solvency II for the Pension Protection Fund, and how difficult that might be given the fact that the PPF is the protection regime for all the funds. Who is looking after the contagion risk? And what about the crossovers with other territories? I am afraid that I do not have ready answers to these questions, other than to remark that the reason for wanting to recognise protection regimes in the holistic balance sheet is to recognise in the funding obligations that those funds that are enjoying the protection have to pay for it through levies.

So it is one of those questions that is difficult to answer. Nevertheless, you can see why people want it answered because not to recognise a pension protection regime would imply that the fund itself is not getting the benefit of the protection and ought to be funded to a higher level.

Mr Martin was kind to say that the switch to CPI has made a difference to the PPF, which is true. As part of the policy development, we were able to use our model to inform some of the thinking, or at least to quantify some of the thoughts. We think that, from a PPF perspective, that is more our role, rather than to set the policy itself. Having a framework which can balance the investment risks, the levy strategy and the options about what might happen to compensation is quite important, because we are able to feed back into the system some quantitative impact of our own in terms of the policy options that are available.

For example, the Minister has gone on record as thinking aloud on such things as you suggest, such as deferring retirement age or linking it with longevity, or potentially excluding spouse's benefit from the pensions obligations and, by implication, from our rate of compensation. We can feed into those thoughts some quantitative work based on our models, which I am sure would be valuable.

In terms of risk-sharing ideas and potential options, it is indeed a regulatory risk but the PPF can only deal with the known environment. There are various ideas floating around. One option might have the effect of removing some of the risk altogether, in which case there may be less reason for the Pension Protection Fund to continue. But if there is a new model that emerges that still has some element of insolvency risk on the part of the sponsor, there would be some scope for a protection regime to support that risk.

That leads in some ways to the points that Mr Richards made about the fact that, in the European market, the protection model provides insolvency cover only, rather than also managing out the annuity risk.

I have previously mentioned the problems of the market dealing with small annuity books. Equally, there is a further problem, in comparison with European models, of the insurability of pension scheme sponsor. There is quite a wide divergence of credit-worthiness of pension scheme sponsors in the UK and some would be uninsurable. In some of the territories mentioned in the paper, pension schemes are only capable of being sponsored by companies, corporations or groups of corporations of a certain credit quality, which is why the provisions we have, where the PPF must insure everybody, make it more difficult to be selective in the risks that are taken.

Even if there was a new pensions regime under which there was more risk sharing, there might still be insolvency risk that needed to be protected against. If there is still some element of cross-subsidy, because some risks are uninsurable, then there would still be some disputes around the cost of the protection, and you would indeed get to the position, which one of the speakers has observed, in which our cost seems to be lower than the market equivalent. That was the point of the chart in section 8, which states that the PPF does not provide complete protection. There is this seeming gap in terms of the absolute certainty and guarantee that the Pension Protection Fund would be able to pay all its promises.

That brings me round to Mr Martin's comment, which referred to the fact that there is the safety measure of reducing compensation levels; but from the point of view of managing the finances of the business, the PPF works on the principle that such an outcome would be failure, so we deal with the regime as it is without any uncertainty over what levels of compensation or indexation might be in future.

Mr I. W. McKinlay, F.F.A.: I am grateful to Mr Clarke for dealing with the foregoing matters relating to insurance. These issues are complex and my own experience is mainly in pensions and investment.

On the matter of models, the paper cannot tell you everything that is going on at the coalface in the PPF; of course, we use models. I think it would be wrong to say that we rely on them to the extent that might be understood from the paper. What they are actually used for is guidance. The construction of investment portfolios, in particular, has to have an element of art, and perhaps even opportunism, to it rather than just applying a model. When I say “opportunism” I do not just mean return-seeking opportunities. I mean risk-reducing opportunities as well.

I am aware of investment models that have shocks built into them, including non-linear events. What happens is that the funnel just expands. In the business of constructing portfolios, you are seeking the relative improvement that you can achieve. Can you benefit from diversification? The funnel of doubt may have expanded but you still end up with a similar solution.

Having said that, I think you have to be mindful of using models in portfolio construction. We do shock the models, but perhaps a better way of describing it would be that we bend them to our will.

That may result in some anomalous outcomes. I think the three scenarios in the paper are only three of a group of scenarios. For example, we have considered the possible disruption which may be caused by a break-up of the Euro, hyper-inflation, Japanese-type depression, and so on. We supplement this analysis by considering where the models begin to break down, which is at the tail. We have attempted to put specific tail hedges in place and this is ongoing; we are always trying to refine the portfolio to make sure it is robust against the shocks that have been described.

The Chairman: Thank you very much for the many thought-provoking responses.

I will now ask Richard Muckart to pull the threads together for us before we ask Ian McKinlay to make a few final comments.

Mr R. D. Muckart, F.F.A. (closing the discussion): Thank you, Chairman, for this opportunity to close the discussion. I read the paper with great interest but, with an investment background, I had considerable scepticism over much of the modelling or perceived modelling. So I am slightly relieved that Ian McKinlay has stated that the model is only there as background and not followed as an investment model.

Mr Barker is a man after my own heart. I agree with him about this erroneous investment policy. I cannot see how buying gilts, especially at a time when interest rates are as low as they are, is going to do anything for either the long-term growth of the economy or attaining the objective of the PPF, which is to be fully funded by the year 2030.

I think that you should have active management. There seems to be a contradiction in terms of the investment of the Fund. On the one hand, they are taking out the interest rate and inflation risks. On the other hand, they are buying alternative investments, to some extent at the expense of equities. I would regard it as an opportunistic grab in the alternative market in the last few years because returns were seen to be available. But how available? We all know the appearance of returns is in the eye of the beholder and it is going to be some time before some of these materialise, particularly where alternative investments are concerned, be they private equity or property.

International comparisons are useful. Mr Richards highlighted the considerable difference in approach. It is probably not surprising that the Europeans use the approach that they do, given the background and nature of their financial services industry, and the way pension funds and the provision of pensions has developed, particularly in Germany.

I am surprised that the UK has followed the US model to the extent that it has, simply because of the point that Mr Richards made, that the Guaranty Benefit Corporation, if it is not close to going into administration, does have that problem and is always a political football.

One has to wonder how long it might be before the PPF becomes a political football if it were to achieve its goals and reach that solvency point early. If it starts to build up a reserve, what happens to that reserve? Does it sit there as a windfall or rainy day event fund? Does the PPF change its investment policy to get rid of it? I guess it is possible that the levies would drop away because they would not be needed to the same extent. But it seems to me that, while we have talked about how the fund is risk-managed at the moment, if it is successful – maybe I should say “when” it is successful – it has a new risk problem to solve.

There has been much talk about Solvency II and whether it should apply to pension schemes. I am no expert on Solvency II, but I cannot help feeling that the requirements that are being imposed of capital adequacy in effect are a step too far for most employers in the UK. They will baulk at the fact that they will have to put up more capital. I would see that as yet another reason why defined benefit schemes would close. In talking about defined benefits, and Mr Clarke talked about the smaller schemes, I would worry that, in the long term, the smaller schemes that are traditionally probably better funded than many of the larger ones will want to walk away. Why should they pay insurance for something that they probably are not going to take advantage of? At the moment, it is compulsory but you could see a groundswell of opinion with the opt-out, “Let us close our defined benefit scheme, go to DC and work it out. If the buy-out annuity market is big enough, then we can deal with it.”

So I worry that the schemes that are going to be left that might fall within the PPF's ambit are in fact those that are in a poorer state anyway, and therefore its claims experience could be worse than is predicted because the good schemes leave.

Professor Clarkson was provocative in his needing to introduce shocks into the model. Like him, I have a healthy scepticism of financial economics. I suppose when you have been a practical investment manager for 30 years, you worry more about the returns in the short term rather than in the long term and you are well aware that there are far too many outliers, that random walks do not happen on a consistent basis, and that the correlation of returns does exist.

Efficient market theory is, in my view, suspect, and I hope that the models that are in use do not involve too much efficient market theory.

I did see a reference to Sharpe ratios as a method of covering the risk. I reserve my judgement on that.

I loved Professor Clarkson's analogy of flashing lights at the top of the cycle. If only we could all find the top of the cycle in setting our investment policy. Hindsight is a wonderful thing. The trouble is we never have it.

It is on that point that I wonder whether the strategy at the moment has a possible problem. We have a bubble, as far as I am concerned, in interest rates. When long interest rates, as at the start of the year, are around 3% and inflation is running at a high level, are the derivatives that the Fund has in place doing what they should do? They must be significantly in the money, assuming that they were taken out some years ago. I am assuming that they are protecting downside interest rates and upside inflation. But maybe not. Maybe they are working in the real market.

My problem with derivatives, and the fact that they are counterparty and OTC, is that the providers of these must see the PPF coming.

If I am an investment bank, and I have the PPF as a client, I know it is a buyer. I know which way it is going. With an insurance company or a pension fund, there is always the possibility that it is going the other way. I would have felt that, with the PPF, it was a one-way street. I therefore wonder how it protects itself against the market, knowing which way its book is likely to be.

I found the paper interesting, although I thought the discussion might be more robust. There is one section I would like to comment on and that is 7.2.3, where the model, the economic scenario generator, appears to generate return on cash at a long-term average of around 5%pa to $$$5\tfrac{1}{2}\% $$$ pa with an average risk premium for equity investment of $$$3\tfrac{1}{2}\% $$$ pa to 4%pa. While I can live with the $$$3\tfrac{1}{2}\% $$$ pa to 4%pa equity risk premium, I find the cash return surprisingly high and I wonder if the economic scenario generator does not need some adjustment to more current day returns. The thought of a 9%pa return from equities in the long term is pretty attractive.

The Chairman: Thank you. We are now drawing towards the end of the meeting and I should like to ask Ian McKinlay to respond on behalf of the authors. Ian is the PPF's chief investment officer. Prior to that, he learnt his trade as an investment consultant with a series of different pension consultancies. Ian has already given us some of his thoughts.

Mr I. W. McKinlay: There was a lot of material in Mr Muckart's remarks and I will try to deal with them all.

There has to be a framework within which to make investment decisions. Modelling is part of the framework that we use. We are modelling not just our projected closed balance sheet but the emerging balance sheet. If we are not recognising the evolution of our balance sheet through time, including future claims, how can we sensibly construct an investment policy?

You talk about buying gilts, and I am going to come onto that in a minute. The point about buying equities – Ken Barker said that he would like to see some investment – is perhaps fair. But we have to recognise, at the PPF, our off-balance sheet risk is, at least in part, equity risk. There is a risk of our sponsors defaulting, which has to be correlated to the equity market. We also need to consider the way their pension funds are invested, and the purple book data that I saw was suggesting a 45% exposure to equity.

So, although there is some equity in the PPF, to double up on equity, potentially at a time when sponsors are defaulting and the equity markets are coming under stress, seems a challenging argument to put forward.

I will also talk about alternatives. I think that the PPF, and other large funds, have a competitive advantage. That competitive advantage is scale, and the ability to embrace complexity. We allocate resources accordingly. If we are not exploiting the opportunities that are in the alternatives markets, we are doing our stakeholders a disservice. We have some exposure to equities but it is global, and we have some specific protection in place against extreme events. But these sit alongside other assets where we can exploit our scale, such as private equity, infrastructure, distressed debt, and so on. This is an important way to make money in a risk-adjusted fashion.

We have some gilts for risk control. They are not held as a return-seeking asset. Quite a low proportion of the fund is in gilts, apart from that which is there for insurance purposes and to back our swaps as collateral.

We are trying to strike a balance between financing our liabilities and controlling risk. We must make an additional return in relation to the financing of our liabilities. We do that through holding equities, holding a portfolio of credit, and other such alternatives.

On the matter of seeing us coming, I take the point, and I am mindful of it. We are not just dumb buyers, I hope. When yields are as they are, we are able to take a view that those yields are extremely low. If we believe them to be low and we can make a profit from those yields rising in future, we can take what amounts to a short position, a not-fully-hedged position. It is not just a race to buy hedging instruments.

Our swaps are currently in the money which results in our being posted cash and gilts. When they are in the money we try to realise them and achieve the most out of them, so that there is not a drag on performance.

There were some questions about the specifics of model assumptions. Perhaps the paragraph referred to might have been better crafted. The return on cash, from memory, is something about $$$3\tfrac{1}{2}\% $$$ pa to 4%pa. It stabilises at 5%pa to $$$5\tfrac{1}{2}\% $$$ pa. In the long term, it reverts; it is not flat. We are not assuming that we are going to get $$$5\tfrac{1}{2}\% $$$ pa on cash forever.

I have tried to deal with the specific questions. I hope that my response, at least in part, deals with them. I am happy to have a discussion outside the meeting.

What I was going to say in closing is that, whilst the paper describes the management of the UK PPF, it is a discussion paper. I would emphasise this. We have never at the PPF claimed a monopoly on the best ideas. We like to think that there is a good degree of challenge within the PPF and that that challenge is sufficiently robust so that we end up with a robust solution in the management of our risks.

However, at all times, we aspire to engage with our stakeholders and expose ourselves to challenging new ideas. I have already heard quite a bit today and I noted the material on solvency, for example. I heard that we could make more use of the Solvency II framework and the use test which applies to insurance. I also heard good challenges on the matter of the use of models, and the construction of the investment portfolio.

In passing, I would say that the discussion that we have had today is not unlike the discussions that we have inside the PPF. Some people are opposed to the insurance approach; others are not. There is creative tension, which results in the management process and outcomes that are described in the paper.

There were some questions on the funding framework time horizon, which I should like to pick up before I close.

The reason we decided on 2030 was because we were expecting maturity of the pension fund landscape. It may be challenging to achieve that, but you have to set yourself a target. We have done that in a probabilistic framework. I wonder how many other funds have done that. There may be a lesson, not necessarily to adopt our investment policy, or the specifics of the framework, but to look at the funding framework and the governance that we use. Is there a lesson there that can be used to everyone's advantage?

To conclude, publishing this paper is part of the ongoing process of enhancing what we do at the PPF, and we are very grateful for the input which we have had today. We have another session in London next week. Whilst there may be an 80% probability of paying our pensions, you can, however, be guaranteed that we will take any ideas and make the best of them in the PPF.

Finally, I should like to thank everybody here for your time on behalf of the authors of the paper. We thank you for the quality of contributions; we welcome your input.

The Chairman: Thank you very much. I recall somebody said that the PPF was a long time coming in the UK. I know that it was a long time discussed and there was much scepticism as to whether it would work. It has perhaps become, in the eyes of most pension scheme sponsors, a necessary evil; something they love to hate.

I have seen it, however, in a new light tonight. It has made me realise that it is a subject which unites actuaries in the sense that we have had some pensions actuarial thought and some insurance company thought, and Solvency II has been mentioned. We have looked at various models, including investment models, and whether they work or not; we even had a touch of longevity in the paper.

I should just like to thank all the contributors tonight, but in particular I should like to thank the authors, most of whom are here this evening. I should also like to thank, in particular, Mr Clarke and Mr McKinlay for giving us their thoughts, and also to Mr Ledlie and Mr Muckart for opening and closing the discussion.