The Moderator (Mr G. A. Bradley, F.I.A.): Hello everyone. My name is Glyn Bradley. I am an actuary at Mercer and a member of the IFoA’s Pensions Board, where I chair its Pensions Research Subcommittee. We set up and oversee the member-led working parties, and we are always looking for new ideas and members. For example, there is currently a Liability Driven Investment (LDI) working party looking for new members. Please contact me if you are interested in helping us, or alternatively you can find all the vacancies on www.actuaries.org.uk. All our working parties aim to generate papers and events, such as tonight, for our members and the profession more widely. It is always satisfying for me to see a working party reach this stage of the process. I would like to thank the Commutation Rate Working Party for their paper and for joining us tonight.
I would like to introduce tonight’s topic: “Considerations for actuaries when advising on commutation rates.” The paper arose from the recommendation of a thematic review several years ago on actuarial factors used to calculate benefits and pension schemes.
We are joined tonight by three of the Working Party members. Jonathan Hilsden has worked over 10 years at Willis Towers Watson (WTW), advising both pension scheme trustees and their corporate sponsors for schemes ranging from £100m to £20bn. He is a member of WTW’s specialists on transfers and commutation factors. He leads WTW’s annual commutation factor benchmarking survey, and he also leads a team which provides guidance to trustees on scheme factor reviews. He has been the chair of the commutation research working party since it started in September 2021.
Kerry Lindsay is a scheme actuary and a senior consultant with Hymans Robertson. She has over 12 years of experience in the pensions market where she advises a range of pension scheme trustees and companies with scheme sizes ranging from 20m to 16bn. Her experience includes supporting trustees in developing their funding strategies, advising on scheme mergers and benefit changes, and supporting trustees as they secure and transition their scheme liabilities on their journey towards the end game.
Judith Fish is a qualified actuary, with over 20 years of post-qualification experience and is an accredited professional trustee who works on a wide range of schemes. She joined Dalriada Trustees Ltd in July 2019, having previously worked in-house on a multi-billion-pound pension scheme. Prior to this, she worked for several large consultancies, both as a scheme actuary and a corporate advisor. More recently, she has acted as the chair of trustees for a scheme with an overseas sponsor, helping them navigating the additional difficulties this brings.
I am also joined tonight by David Gordon. He is the senior review actuary at the IFoA, and he leads the thematic review programme. In his 2020 report: “Pensions: actuarial factors used to calculate benefits in UK pension schemes,” it recommended the research that is being discussed in the meeting today. Before David (Gordon) joined the IFoA as a staff member in 2020, he worked as a Defined Benefit (DB) scheme actuary for 30 years. During this time, he had a leading role within his firm on regulatory matters.
Mr J. P. Hilsden, F.I.A.: Thank you Glynn (Bradley) and welcome everybody to the sessional meeting for the Commutation Rate Research Working Party. Kerry (Lindsay) and I will spend around 20 minutes setting the scene and summarising some of the key recommendations from the working party: in the areas of actuaries advising on setting the commutation rates. We will then be joined by Judith (Fish) and our guest David Gordon to discuss any questions from the audience.
As Glynn (Bradley) already mentioned, the December 2020 thematic review covered actuarial factors, in particular commutation rates and transfer value factors. The working party was formed because of the further research recommended from the thematic review. I will also convey my thanks to David (Gordon) to be our guest on the panel discussion and for his support of the thematic review.
Throughout this presentation, where myself, Kerry (Lindsay) or other working party members refer to “We” or “Our”: we are referring to the working party and our views as a working party. There are six working party members in total, including myself. Most of us are from a range of pension consultancies, and we have one professional trustee in Judith (Fish). As a trustee, Judith was able to present a different perspective to the rest of us on the working party. Particularly in areas such as how trustees preferred to view actuarial advice in commutation rates presented. We met regularly, starting from September 2021, to discuss the brief set out by the thematic review, which we will cover on the next figure.
The recommendations from the thematic review which we covered were: the appropriate allowances to make for selection risk, market volatility and other common criteria in setting commutation rates. We also looked at the frequency of commutation rate reviews and how those reviews should be presented. We did not cover all the areas in the thematic review. As an example, we did not cover the recommendation for industry-wide benchmarking on commutation rates.
Our paper is published on the Institute and Faculty of Actuary’s (IFoA) website and the link was also added to the registration invitation last week. The paper provides more information than we can cover in this short presentation. If you were to take one thing from today, it would be that our paper and presentation was designed to be thought-provoking and stimulate further discussion amongst actuaries when advising and setting commutation rates. We recognise, as a working party, that there are many different views and opinions which we hope will be raised in the Q&A panel discussion later.
As already highlighted earlier, we understand that are many different considerations on the topic. We felt it was therefore important to set out at the outset what we did not consider during our work. As well as the already mentioned industry-wide benchmarking, we did not consider any other uses of commutation rates, for example in trivial commutation calculations. We also did not consider commutation rates in public sector pension schemes, as this was outside of the experience and expertise of many on the working party. Some could argue that there could be more regulation around commutation rates and whilst this may be true, this was also not considered by the working party as again we steered clear of any legal arguments in which we are not experts ourselves.
I am sure we all appreciate how fast-moving the pension industry can be, with the most recent examples during the September to October 2022 events or the more recent Spring budget announcements. It is important to clarify that we have considered commutation rates at the current point in time. If you are listening to a recording (or reading a transcription) of this session in the future, then things may have developed or changed further. However, we hope that many of the points raised in this session (and in our paper) will still be helpful to actuaries advising on commutation rates in the longer term.
Ms K. Lindsay, F.I.A.: The way that we approached this review was to first think about what the theoretical starting point might be for setting commutation factors, and then secondly, the reasons in which you might deviate from this position. In practice, we did note that the current commutation rates are often used as a starting point by actuaries in their advice. Whilst we do believe that the current commutation rates may come into your advice at some point, the working party believe that it is important to start with a blank sheet of paper when formulating your advice to trustees.
Unlike transfer values, there is not a prescribed methodology to start from. However, as the transfer value is designed to represent the best estimate value of the benefits that members are giving up, we believe that the transfer value principles are the most appropriate place to start for commutation factors as well. This can ensure that members are given a fair value for pension benefits they are exchanging for cash and that terms are consistent across the different options that members have.
When we refer to transfer value principles, we mean that assumptions should be set consistently. Examples include: there should be consistency in the approach when deriving your discount and inflation rate assumptions, whether or not one is allowing for an inflation risk premium and the setting of the various longevity and demographic assumptions. We would also suggest consistency in terms of whether yield curves or flat rates are used, whether terms are unisex or sex-dependent and whether different factors are used for different tranches of benefits. However, the working party do note that there are valid reasons why actuaries might recommend deviating from this theoretical starting point and we are going to cover a number of these in today’s session.
Mr Hilsden: Following on from the theoretical starting point, as explained by Kerry (Lindsay), we came up with several different reasons that actuaries could justify moving away from that theoretical starting point and some which we did not believe to be good reasons. Figure 1 sets out the main groups of the areas that we considered.
Figure 1. Reasons to deviate from theoretical start point.
We do not have time to cover all of these, so we have picked out a few we felt were most relevant to raise in today’s presentation. More information on all these areas is provided in our paper.
By selection risk, we refer to members in poorer health who are expected not to live as long as those in normal health. These members may select against the scheme by taking a larger pension commencement lump sum (PCLS). In our view, selection risk should not be used as a justification for moving away from the theoretical starting point, unless it is quantified and can be justified by looking at the data. We did explore some potential industry-wide data sources but found that the information needed to do this analysis was not captured in those sources. Therefore, our view is that schemes would have to use their own data to justify selection and the appropriate reduction to their theoretical starting point. We believe that any evidence used to justify the adjustment for selection should be applied in the mortality assumptions. However, in our view, it may be difficult to justify reducing commutation rates for selection. If the majority of the members take commutation, how can everyone be in poorer health? An example as set out in our paper was: if a scheme is making a 10% reduction to commutation rates for selection, using the typical mortality assumptions and assuming 80% of members take on the commutation, this is equivalent to assuming that life expectancy is more than 10 years lower for those who commute compared to those members who do not commute and equates to a mortality scaling factor of around 200% for members who do commute versus 45% for members who do not.
Allowing for de-risking caused much debate amongst the working party, and it is likely to cause much debate amongst the audience. We believe that it is appropriate to reflect de-risking to date. However, there may be exceptions where there has been a specific agreement between the Trustees and sponsor, when it is appropriate to not reflect de-risking to date in the setting of commutation rates.
Regarding allowances for future de-risking, where this is theoretical (so not formally documented or is contingent on future events) then it may be reasonable not to reflect this in the commutation rates. However, since this future de-risking might not be reflective in the cash equivalent transfer value basis, which as Kerry (Lindsay) explained is the theoretical starting point, you may not be deviating from that theoretical starting point. Where the de-risking is agreed, formally documented and not contingent on future events, then we believe it could be allowed for in the commutation rates.
We consider that if the scheme is under-funded on a commutation basis, then the commutation rates could be reduced to reflect this. This is like what would occur for transfer values in under-funded schemes. The company’s covenant should also be considered before reducing the commutation rates, in the same way that it is considered for transfer values. However, while members can transfer at any point up to one year before normal retirement age, or some schemes allow members to transfer out later, in practice they may have less choice as to when they retire. The decision-makers should therefore consider alleviating the impact of any stepped changes in commutation rates. We will touch on stepped changes in commutation rates later in this session.
Ms Lindsay: The next area we considered was intergenerational fairness. We believe it is important for actuaries to retain the same principles over time. For example, if you have always kept factors with reference to the transfer value basis, then it would not be appropriate to move away from that only at times where it is more favourable to the member. We also, as Jonathan (Hilsden) has just alluded earlier, believe it is important to avoid cliff edges in rates. It might therefore be appropriate to move closer to the theoretical start point in increments rather than in one step. Importantly, we did not think the argument of, “We have always kept rates low in the past, so we should keep them low now so that everyone gets the same per value over time.” is a reasonable argument from our members’ point of view.
The thematic review noted that market practices were commonly cited by actuaries as a reason for setting factors the way they are. Indeed, we all had experiences of seeing benchmarking data used to help justify certain commutation rates. Importantly, we are not completely against the use of market or other industry data, and we believe that this could be appropriate if you are comparing to similar schemes. However, we do not believe that benchmarking data should be used as a justification for keeping commutation rates low. Benchmarking data is, by its design, historic. There could be factors that have been in place for many years that may be coming up for review where the changes have not occurred yet, for example. It also often does not reflect the individual characteristics of specific schemes: for example, are they set by trustees or by the company, does the actuary have to certify the rates, and what are the pension increase types? Therefore, while it may be relevant in some cases, we believe that when using market data within your advice, its limitations should be made clear.
The next area we considered was the practicalities and process of members commuting pension for cash. We noted that some schemes, especially those schemes in the private sector, can be complicated and have a large numbers of different benefit tranches. In such cases, it may be sensible to simplify the number of factors used and combine factors for tranches where the pension increase rates are similar. This is particularly true where schemes are seeking to manage administration costs. However, most administrators now automate their retirement calculations and so they are capable of handling different factors impacting different tranches of benefits. In such cases, we therefore believe that administration simplicity should not be a driver in changing (or not changing) factors.
In providing their advice, actuaries should always be mindful of the balance of implementation costs against the impacts on member benefits. Where relevant, we believe actuaries should err on the side of the members. Finally, as Jonathan (Hilsden) has already touched upon, there were several other areas that the working party considered, the relevance of which varies by a scheme’s specific circumstances. These are covered further in our paper, and we would also be happy to touch upon any of them in the panel’s Questions and Answer (Q&A) at the end of this session.
Mr Hilsden: We agree with the thematic review that three years between commutation rate reviews should be the absolute maximum and not the default. We believe that actuaries should review commutation rates more regularly and a good practice would be to at least perform a high-level check on commutation rates annually. Given recent experience, pre-agreed triggers or circumstances which would lead to review could be beneficial, particularly in times of material movements in market conditions. Consulting actuaries would have to be mindful of the time and cost of these reviews as well.
We also wanted to touch on an area which we felt deserves more attention than it is perhaps currently receiving, which are market-related commutation rates (see Figure 2).
Figure 2. Market-related commutation rates.
We now have a poll question “Do you advise schemes (or are a trustee of a scheme) which have market-related commutation factors?”
The results of the poll show that 27% of the audience say “Yes they do advise or are involved in pension schemes with market-related commutation factors” and 73% say “No.” That is roughly a quarter to three-quarter split. There seems to be many schemes that do use market-related commutation factors, but also a number of schemes that do not. We understand that market-related commutation factors are more common with insurers who provide bulk annuities and schemes that are close to a full buy-in and are moving towards insurer rates - particularly where those buy-ins relate to deferred members. Market-related commutation rates may simplify valuation negotiations during the periodic rate review cycles and will likely reduce the potential of stepped changes between the commutation rates (although rates will be changing more regularly). This can make it more difficult for members to plan for retirement if the commutation rate is changing more regularly and could be more challenging to administer and manage the scheme. Some schemes may decide to guarantee the commutation rates for a short period of time (for example, it could be guaranteed for three months in a similar way to how transfer values are guaranteed).
Of course, the technical actuarial standards should be followed when providing advice (to trustees or corporates) on commutation rates.
A lot of the points on Figure 3 were highlighted in the thematic review, but we wanted to highlight and emphasise a few further points.
Figure 3. Presenting advice.
We believe that actuaries should set out their theoretical starting point and their reasons for moving away from this. Judith (Fish) on our working party also commented during our discussions that she likes to see in the advice from her scheme actuaries the impact on members’ benefits from the change in commutation rates. We think that the impact on funding, the actual basis used and a description of the decision-makers involved in setting the commutation rates should also be disclosed. As set out in the thematic review, if the scheme actuary has the authority to certify the commutation rates, we suggest this authority is provided in writing.
Ms Lindsay: To summarise the working party’s key conclusions: firstly we looked at an appropriate starting point for setting commutation rates and the allowances that should be made for selection risk, market volatility and other criteria. We believe that the theoretical starting point should be the best estimate, based on a scheme’s transfer value. In our paper, we do highlight several valid reasons why actuaries may advise trustees to deviate from the theoretical starting point. However, we did also note that several of the common reasons that we observed are being used without adequate justification. It is important that if you are moving away from the best estimate starting point, that you justify it for your scheme’s specific circumstances.
Secondly, we believe that three years should be the maximum time between commutation rate reviews. A more frequent high-level review may be appropriate, especially in times of significant market volatility or if there is a lot of changes happening in your scheme. We believe that there are many good reasons why schemes should consider setting market-related factors and we do expect this to be more common over time, particularly as schemes get closer to the insurance end game. We believe that it is appropriate for commutation rates to be reviewed, ideally during or immediately following a triennial actuarial valuation.
In presenting advice, actuaries should ensure that it is clear and as concise as possible, which I appreciate can often be a challenge. Any recommended changes should be highlighted. When you are proposing a change in factors, the impact on the benefits that appear in members’ retirement quotations should be highlighted explicitly.
Finally, as Jonathan (Hilsden) has mentioned earlier, if required under scheme rules, actuaries should provide their certification of the reasonableness of scheme factors in writing. We will now pause and pass over to David Gordon, who is going to cover more around the wider thematic review programme.
Mr D. A. Gordon, F.I.A.: Good afternoon, everyone. Can I first thank Jonathan (Hilsden), Kerry (Lindsay) and the rest of the working party for their informative and illuminating research. I joined the IFoA three years ago to lead the thematic review programme and, as we have been discussing, this research stemmed from my first thematic review report. Thank you, during the presentation today, for name-checking the report and indeed endorsing many of the recommendations.
As a Royal Charter profession, the IFoA has a duty in the public interest to regulate. This is carried out by the Regulatory Board, which sets the Actuarys’ Code, professional standards, and guidance materials. The thematic review programme (see Figure 4) reports to the Board, and this enables us to monitor the work (in specific areas) of actuaries. This in turn then informs the Board’s work.
Figure 4. Objectives of Thematic review programme.
The objectives of the programme are to showcase the best practices and to further improve the quality of actuarial work. It also provides evidence to refine standards, guidance, training and education. From the perspective of organisations and members, the benefits are simple. It is an opportunity for us to provide you with independent and confidential feedback on your approach and work. It also supports the IFoA to ensure that its regulatory work is based on evidence and reinforces our credibility as a member-led regulator in the public interest (see Figure 5).
Figure 5. Actuarial factors thematic review.
Now turning to the thematic review itself, I do have a thumbnail of the report cover. The report was published in December 2020, and we had a series of recommendations besides the one on which this research was based. I know that some of these recommendations have been highlighted already in the presentation tonight. We should not forget that the overall finding was that the actuarial advice we reviewed was generally of a very good standard. The recommendations we had were mostly centred around the benchmarking evidence we had of commutation rates being well below transfer values. We called on actuaries to continue following the existing standards fully, but in particular to explain why the two actuarial factors differ (especially when the differences are large) which is a Technical Actuarial Standards (TAS) requirement. The advice on the two sets of factors (commutation rates and transfers values) is often provided at the same time after the actuarial funding valuation, but the advice is often not tied together, but rather two distinct sections or even two reports. This should not take away the need to explain why they differ if they do indeed differ at all.
On reviewing rates regularly, this has clearly been well rehearsed this evening. In our review, we mostly saw triennial review cycles, and it is interesting from the poll earlier that a quarter of you have seen market-related rates, which is probably higher than what I observed in the thematic review.
On certification, around half the submissions we observed had commutation rates that needed to be certified by the actuary, yet we only saw one or two positive affirmative statements as required under the standards. I suspect practice on this has moved on since 2020, and I would continue my general recommendation to communicate assumptions and results better.
These member recommendations are simply to encourage actuaries to follow existing standards and guidance fully. I am pleased that so many of them are being endorsed earlier. The role of the thematic review was not to provide a critique on the technical aspects of setting commutation rates, hence the recommendation to set up the research tonight. In answering questions later, I will not be (even though I probably do have a view) expressing technical views. I will try to express my answer from the perspective of existing standards and guidance.
Finally, can I also commend our latest report, on corporate pensions advice to Defined Benefit (DB) pension scheme sponsors. To find out more about that report, we have a webinar next Thursday and you can find the QR code on the screen, which should hopefully allow you to register for the event. If that does not work, the event is still linkable from the front page of the IFoA website.
In conclusion, the quality of the advice we observed was good. We had several findings, as you can imagine, about some standards not being followed, many of which were aligned to the themes of the commutation rates review today.
To close, the thematic review programme relies on members and organisations to take part. In our six exercises to date, we have had over 50 organisations taking part. I would like to thank you all for your help with this, because we do rely on the organisations and members.
Mr Hilsden: We will now move onto the panel discussion where Kerry (Lindsay) and myself will be joined by Judith (Fish) and David (Gordon) and invite the audience to submit any comments or questions and Glyn (Bradley) will assist to filter and moderate the questions. So, over to you Glyn.
Moderator: I should begin by mentioning that the working party paper is available on the IFoA website. You can also find (on the events page) the details of that corporate thematic review next week that David (Gordon) mentioned earlier.
We have had a few questions come in, but if there are any more, please do put them in the Q&A panel and I will attempt to group them before putting them to the panellists.
To get the ball rolling, the focus of the paper advocated a theoretical point of view of best estimate rates. You have said that is not the only view one should recommend but that is where you would start as an actuary in your advice. I wondered, Kerry (Lindsay), what were the other options that the group had considered and why were they discounted?
Ms Lindsay: Yes, we did think about various options for the theoretical starting point whether it was: the transfer value basis, technical provisions basis, the insurer basis, self-sufficiency basis or something weaker than best estimate. There is a whole range of possibilities. Where we kept coming back to is the transfer value basis as a best estimate that offers fair value to members who are taking the option. Best estimate by definition is a 50% chance of those assumptions being realised. Members who take the option get fair value, but also importantly, members who do not get the option get fair value as well. If you had gone with something that is more prudent for the scheme, then that could cause a funding strain that weakens the funding position for any members who do not commute cash. But also, if members are commuting pension for cash, they are only commuting 25% of the value, so 75% of the value of their benefits is still in the scheme with a potentially worse funding position.
Moderator: I will take a different line now, unless anybody else on the panel wishes to come in on that. We have had a very pro-member stance on some of the topics in the paper. I am going to ask a pro-corporate question to Judith (Fish) who I think might be well positioned to respond. Commutation is seen by some as an option and a benefit design issue. If that was the case, why shouldn’t the employer be the ultimate arbiter of the commutation rates?
Mrs J. A. Fish, F.I.A.: For some schemes it is clearly part of the benefit design. You have schemes that have a pension and a cash sum written into the rules. But for most of the schemes that came out of thematic review, the commutation factors were decided by the trustees, either on the advice of the actuary or certified as reasonable by the actuary. I think for most cases, you are having to look at the actuary and the trustees to set those factors. Whilst the employer might have some strong views on what those factors should be, we must also remember the members probably do not understand the value of the pension they are giving up. For most schemes, we are seeing most members take pension commencement lump sums on retirement and therefore not selecting against the scheme. I think you must be very careful to look at the rules of the scheme in question when setting the commutation factors, to make sure that what you’re doing is reasonable and in line with the trust deed and legislation.
Moderator: Thank you. Anybody else on that point? Otherwise, I will try my question from a different angle. David (Gordon), you kicked all this off really with your thematic report, which I have recently read again and well worth a look at on the Institute website. There is a lot in there about advice on factors and not just relating to commutation. Generally, you mentioned the advice on factors was of a high standard but there was some variability, specifically in relation to commutation factors. If nothing else, when our members leave this webinar, what would you want them to do differently?
Mr Gordon: I would suggest the members follow your example by reading the report after they have read the research that we are talking about tonight. I think Jonathan (Hilsden) mentioned earlier the objective of today’s paper is designed to be thought-provoking and stimulate debate. Hopefully it does just that within your organisations. I think that would be fantastic. I will leave others to reflect on exactly what the right adjustments should be, but I very much support the conclusion that you have reached today.
Moderator: I will start with a couple of questions. I think a point was mentioned earlier about selection being a possible reason to adapt or to depart from the best estimate and someone said, “for consistency with transfer values, shouldn’t transfer value assumptions make some allowance for possible selection risk?” Who wants to take that one?
Mr Hilsden: Transfer values were not under the specific remit of the Commutation Rate Research Working Party, so we did not consider how to set them. It comes back to whether there is evidence that members who take transfer values have shorter life expectancy. You are guided by the transfer value regulations, which are very different to the commutation rate regulations. From that point of view, you would be setting your transfer value basis in relation to the characteristics of your scheme and of the mortality of your scheme. I do not know whether that is an answer which skirts round the issue. It was not really part of the commutation rate research, but similar principles might be considered.
Moderator: Just while we are on Transfer Value (TV) factors, somebody has asked how consistency would work where a transfer basis is set using a dual discount rate approach.
Ms Lindsay: I am happy to answer that one. Where you have dual discount rates, in relation to commutation rates, it is only the post-retirement discount rate that matters because the pre-retirement discount rate does not come into play.
Moderator: We have a couple questions around the theme of moving towards buyout, whether we would see schemes moving closer to insurer buyout terms. And if that’s a long-term objective, whether it would be sensible to set commutation factors using an insurer’s proxy basis.
Mr Hilsden: I think there is probably a journey towards buyout, both in terms of the scheme, but also in terms of the commutation rates that are set. It depends where the scheme is on that journey. One of the key considerations is whether the scheme is going to reach buyout when there are still deferred members to buyout. You might have more consideration if you have a deferred buy-in which the insurer will set commutation rates in a certain way and whether you should move your commutation rates in line with that. But if you have a process of securing buy-ins for pensioners after they have reached retirement, then there might be less of an impact on commutation rates.
Mrs Fish: I should add that commutation rates from an insurer potentially might not be better. You might find that as schemes have de-risked so much, you have increased your commutation rates to a point where they are higher than the insurer’s rates. But clearly, intergenerational fairness should also be considered so that you don’t get cliff edges.
Mr Gordon: In the thematic review, there was little discussion on buyout and annuity insurance terms. I suspect that was because it was completed during 2020 and based on evidence a year or two before in many cases.
Moderator: Okay, thank you. Now we have a volunteer to go live. If we can bring in Chris O’Brien please, to ask his question in person. Indeed, if anybody else wants to come on camera and as their question, please indicate this in the Q&A.
Mr C. D. O’Brien, F.I.A.: My question was partly answered by an earlier question from yourself. I was concerned that the paper did not include references to the alternatives to transfers value as the theoretical starting point. In particular, the number of references in literature to the use of bond yields to set commutation factors, for example, as is required by law of the Pension Protection Fund. In some cases, the scheme will be wholly invested in bonds and that might be the strategy from which the commutation factors calculated. In many cases it won’t be.
Moderator: Okay, thank you Chris. Sorry, your audio was breaking up a little bit, so I am going to repeat the question for the benefit of others. I think in summary, you were referring to some of the actuarial literature on setting factors using bond yields and you were asking the authors of the paper whether they had given more consideration to setting factors that way?
Ms Lindsay: We did not include in our paper, for want of space, is all the different options that we ruled out. We did discuss a number of options over and above the transfer value basis as a starting point. Jonathan (Hilsden) mentioned earlier the points around allowance for de-risking, and obviously bond yields can become more relevant there depending on the scheme’s investment strategy. As mentioned previously, transfer values were not part of the remit of the Working Party, but I know there can be lots of discussions when it comes to members’ transfer value basis and whether the value that a member receives should vary depending on how the scheme is invested. Indeed, the same principles would apply when you are talking about commutation rates, and whether one scheme should provide different commutation rates because it happens to be in a different investment strategy. This comes back to the point of what we are trying achieve, which is to provoke further discussion. We have come up with a theoretical starting point, but acknowledge that there may well be scheme specific reasons that justify moving away from it. Indeed, the specifics of the scheme’s investment strategy might be one of those.
Chris O’Brien: Thank you.
Moderator: Thank you very much for your question, Chris. I have got another challenge here from Neil Wharmby. He makes the point that the paper is putting forward the viewpoint to start theoretically with the transfer value, which is closely prescribed in legislation on how they are to be performed, which therefore forces you into a best estimate position. Commutation factors on the other hand are not subject to the same legislative environment. Is this a question of, whether we are trying to compare apples and pears when we start with that view? Further, he asks, what is the rationale for adopting a market-based set of assumptions? It is in fact a brilliant question because you have three questions in one, so perhaps different people may want to chip in. The third question is whether intergenerational issues lead one away from a market-based approach? So, who wants to tackle at least part of that first question? Perhaps Judith (Fish) you have a view from a trustee perspective?
Mrs Fish: The first bit around whether we start with the transfer value basis. We noted that there were legislative differences around the transfer value basis, for which there is legislation, and the lack of legislation around commutation factors. We started with the transfer value basis, not because of its legislative background, but because it is meant to be a best estimate basis. We felt that it was the right starting point when converting members’ benefits to an alternative form. So, when you are doing other exercises where you are converting member benefits, you might often start with the transfer value basis Although I thought in the thematic review it was very interesting to see the comparison between the transfer and the commutation values (which were significantly lower). I am sure that most members of pension schemes won’t realise the huge differences in potential value between a transfer and commutation value. Somebody else said earlier that most members take commutation, but far fewer members take transfer values. Commutation is something that members do value as an important part of their benefits.
Ms Lindsay: I can take the second part of the question. It is important to know that we are not necessarily suggesting you have to go down a market-based set of assumptions. We are more highlighting the fact that it should be considered whether a market-based or fixed basis is appropriate for a specific scheme in the time horizon that you are working in. Especially as schemes get closer to insurance annuity purchase, if the insurance factors are market-related, then this might give more justification for moving towards a market-related basis to avoid future volatility in your funding. We are, however, not necessarily of the view one way or the other. The point around intergenerational fairness is interesting and I agree there is an argument that, just because one person is retiring at a point when bond yields have fluctuated, should they get a higher or lower lump sum depending on the market fluctuations? It is a challenging question and it is a similar discussion that happens with the topic of transfer values as well, especially now that you have freedom and choice. The significance of transfer values was around them being a best estimate rather than, as Judith mentioned, the legislative impact. So, it is not about trying to play one set of members off against another.
Moderator: Whilst we are on the topic of intergenerational fairness, we have had a question similar to what was raised earlier on why should those that retire later get a better deal because they have retired later? The other way round you could pose the question, is why should you hold members (or people that have left before) back in order to be fairer? Does holding down rates for the future make it better for the people who’ve perhaps had low rates in the past? Kerry (Lindsay), do you want to elaborate more on that tension?
Ms Lindsay: I think it is quite a challenging one. The holding down of rates in the future because people may have or have not had them in the past. Just because rates have been one way for one group of people does not mean you should not be doing the right thing for your current members. Ultimately, trustees have a duty to act in the best interests of members. I think that’s a really important point. Generations, never mind rates, do experience differences. Longevity improves over time. Someone retiring 30 years ago and taking cash might have been expected to live for a much shorter period than someone retiring now and taking cash. That in itself would lead to a different lump sum. The point around de-risking as schemes get closer to the insurance end game, is an important one, and as Jonathan (Hilsden) alluded to, we had a lot of debate within the working party around how best to allow for this. I do not think we had a conclusive view.
Mr Hilsden: The de-risking journey was an area we discussed a lot, allowing for the fact that schemes are maturing, maybe towards lower risk investment strategies, and whether that should (or should not) benefit commutation factors. This caused a lot of debate and I think one that continues to do so.
Mr Gordon: A reminder of the TAS requirement to explain the rationale for differences between assumptions used for different factors. I would say that it is critical so that the trustees understand the implications of what they are being advised.
Moderator: One questioner said, as factors are generally moving upwards, and actuaries are being asked to certify that, what difficulties might that present for actuaries in relation to certifications in the past where rates were much lower?
Mr Gordon: My observation is that actuaries were not certifying in the past. That is not a reason, just as holding down rates to be fair for people in the past is not either. I do not think not certifying now because you didn’t certify before is a good explanation. Certification is about the facts and circumstances at the time, just as if you certify in the past, it would be about the facts and circumstances at that time.
Moderator: We have time for one more question. This one has come up quite a lot, related to where people leave a deliberate margin, say 10% below the value either on the best estimate or technical provisions basis. How does the panel feel about that as an alternative way of setting factors? Is that building on what you’ve done or is it in contradiction to what you are suggesting?
Mr Hilsden: There are probably two things in the question. An assumption for how many members will commute and what basis those members will commute on is a funding assumption and is required to be prudent. Setting that rate as something like a percentage of technical revisions seems sensible. It will allow for your commutation rates to change without there necessarily being a funding strain, because the technical provisions will not change. I do not think it necessarily means one should set commutation rates in the same way. It is a funding assumption versus the commutation rates that determine how much members get by way of their pension commencement lump sum. You could still use the theoretical starting point of a transfer value, which is a best estimate basis, and which might be less prudent than the funding valuation assumption.
Moderator: It is coming up to the end of our time so I would like to thank everybody on the panel, and also everybody that has joined us today, and particularly those who have put forward questions. I am sorry we were not able to cover them all. I would like to remind everybody to have a look at the paper on the IFoA’s website which you will find on the events page for today. If nothing else, the next time we all author or peer review a factors paper, hopefully we will make sure to go back and challenge ourselves that, besides ticking all the boxes, could we do more to help our clients and members understand our advice. With that, I would like to thank you all very much again and close the session.
You have
Access
Open access
The Moderator (Mr G. A. Bradley, F.I.A.): Hello everyone. My name is Glyn Bradley. I am an actuary at Mercer and a member of the IFoA’s Pensions Board, where I chair its Pensions Research Subcommittee. We set up and oversee the member-led working parties, and we are always looking for new ideas and members. For example, there is currently a Liability Driven Investment (LDI) working party looking for new members. Please contact me if you are interested in helping us, or alternatively you can find all the vacancies on www.actuaries.org.uk. All our working parties aim to generate papers and events, such as tonight, for our members and the profession more widely. It is always satisfying for me to see a working party reach this stage of the process. I would like to thank the Commutation Rate Working Party for their paper and for joining us tonight.
I would like to introduce tonight’s topic: “Considerations for actuaries when advising on commutation rates.” The paper arose from the recommendation of a thematic review several years ago on actuarial factors used to calculate benefits and pension schemes.
We are joined tonight by three of the Working Party members. Jonathan Hilsden has worked over 10 years at Willis Towers Watson (WTW), advising both pension scheme trustees and their corporate sponsors for schemes ranging from £100m to £20bn. He is a member of WTW’s specialists on transfers and commutation factors. He leads WTW’s annual commutation factor benchmarking survey, and he also leads a team which provides guidance to trustees on scheme factor reviews. He has been the chair of the commutation research working party since it started in September 2021.
Kerry Lindsay is a scheme actuary and a senior consultant with Hymans Robertson. She has over 12 years of experience in the pensions market where she advises a range of pension scheme trustees and companies with scheme sizes ranging from 20m to 16bn. Her experience includes supporting trustees in developing their funding strategies, advising on scheme mergers and benefit changes, and supporting trustees as they secure and transition their scheme liabilities on their journey towards the end game.
Judith Fish is a qualified actuary, with over 20 years of post-qualification experience and is an accredited professional trustee who works on a wide range of schemes. She joined Dalriada Trustees Ltd in July 2019, having previously worked in-house on a multi-billion-pound pension scheme. Prior to this, she worked for several large consultancies, both as a scheme actuary and a corporate advisor. More recently, she has acted as the chair of trustees for a scheme with an overseas sponsor, helping them navigating the additional difficulties this brings.
I am also joined tonight by David Gordon. He is the senior review actuary at the IFoA, and he leads the thematic review programme. In his 2020 report: “Pensions: actuarial factors used to calculate benefits in UK pension schemes,” it recommended the research that is being discussed in the meeting today. Before David (Gordon) joined the IFoA as a staff member in 2020, he worked as a Defined Benefit (DB) scheme actuary for 30 years. During this time, he had a leading role within his firm on regulatory matters.
Mr J. P. Hilsden, F.I.A.: Thank you Glynn (Bradley) and welcome everybody to the sessional meeting for the Commutation Rate Research Working Party. Kerry (Lindsay) and I will spend around 20 minutes setting the scene and summarising some of the key recommendations from the working party: in the areas of actuaries advising on setting the commutation rates. We will then be joined by Judith (Fish) and our guest David Gordon to discuss any questions from the audience.
As Glynn (Bradley) already mentioned, the December 2020 thematic review covered actuarial factors, in particular commutation rates and transfer value factors. The working party was formed because of the further research recommended from the thematic review. I will also convey my thanks to David (Gordon) to be our guest on the panel discussion and for his support of the thematic review.
Throughout this presentation, where myself, Kerry (Lindsay) or other working party members refer to “We” or “Our”: we are referring to the working party and our views as a working party. There are six working party members in total, including myself. Most of us are from a range of pension consultancies, and we have one professional trustee in Judith (Fish). As a trustee, Judith was able to present a different perspective to the rest of us on the working party. Particularly in areas such as how trustees preferred to view actuarial advice in commutation rates presented. We met regularly, starting from September 2021, to discuss the brief set out by the thematic review, which we will cover on the next figure.
The recommendations from the thematic review which we covered were: the appropriate allowances to make for selection risk, market volatility and other common criteria in setting commutation rates. We also looked at the frequency of commutation rate reviews and how those reviews should be presented. We did not cover all the areas in the thematic review. As an example, we did not cover the recommendation for industry-wide benchmarking on commutation rates.
Our paper is published on the Institute and Faculty of Actuary’s (IFoA) website and the link was also added to the registration invitation last week. The paper provides more information than we can cover in this short presentation. If you were to take one thing from today, it would be that our paper and presentation was designed to be thought-provoking and stimulate further discussion amongst actuaries when advising and setting commutation rates. We recognise, as a working party, that there are many different views and opinions which we hope will be raised in the Q&A panel discussion later.
As already highlighted earlier, we understand that are many different considerations on the topic. We felt it was therefore important to set out at the outset what we did not consider during our work. As well as the already mentioned industry-wide benchmarking, we did not consider any other uses of commutation rates, for example in trivial commutation calculations. We also did not consider commutation rates in public sector pension schemes, as this was outside of the experience and expertise of many on the working party. Some could argue that there could be more regulation around commutation rates and whilst this may be true, this was also not considered by the working party as again we steered clear of any legal arguments in which we are not experts ourselves.
I am sure we all appreciate how fast-moving the pension industry can be, with the most recent examples during the September to October 2022 events or the more recent Spring budget announcements. It is important to clarify that we have considered commutation rates at the current point in time. If you are listening to a recording (or reading a transcription) of this session in the future, then things may have developed or changed further. However, we hope that many of the points raised in this session (and in our paper) will still be helpful to actuaries advising on commutation rates in the longer term.
Ms K. Lindsay, F.I.A.: The way that we approached this review was to first think about what the theoretical starting point might be for setting commutation factors, and then secondly, the reasons in which you might deviate from this position. In practice, we did note that the current commutation rates are often used as a starting point by actuaries in their advice. Whilst we do believe that the current commutation rates may come into your advice at some point, the working party believe that it is important to start with a blank sheet of paper when formulating your advice to trustees.
Unlike transfer values, there is not a prescribed methodology to start from. However, as the transfer value is designed to represent the best estimate value of the benefits that members are giving up, we believe that the transfer value principles are the most appropriate place to start for commutation factors as well. This can ensure that members are given a fair value for pension benefits they are exchanging for cash and that terms are consistent across the different options that members have.
When we refer to transfer value principles, we mean that assumptions should be set consistently. Examples include: there should be consistency in the approach when deriving your discount and inflation rate assumptions, whether or not one is allowing for an inflation risk premium and the setting of the various longevity and demographic assumptions. We would also suggest consistency in terms of whether yield curves or flat rates are used, whether terms are unisex or sex-dependent and whether different factors are used for different tranches of benefits. However, the working party do note that there are valid reasons why actuaries might recommend deviating from this theoretical starting point and we are going to cover a number of these in today’s session.
Mr Hilsden: Following on from the theoretical starting point, as explained by Kerry (Lindsay), we came up with several different reasons that actuaries could justify moving away from that theoretical starting point and some which we did not believe to be good reasons. Figure 1 sets out the main groups of the areas that we considered.
Figure 1. Reasons to deviate from theoretical start point.
We do not have time to cover all of these, so we have picked out a few we felt were most relevant to raise in today’s presentation. More information on all these areas is provided in our paper.
By selection risk, we refer to members in poorer health who are expected not to live as long as those in normal health. These members may select against the scheme by taking a larger pension commencement lump sum (PCLS). In our view, selection risk should not be used as a justification for moving away from the theoretical starting point, unless it is quantified and can be justified by looking at the data. We did explore some potential industry-wide data sources but found that the information needed to do this analysis was not captured in those sources. Therefore, our view is that schemes would have to use their own data to justify selection and the appropriate reduction to their theoretical starting point. We believe that any evidence used to justify the adjustment for selection should be applied in the mortality assumptions. However, in our view, it may be difficult to justify reducing commutation rates for selection. If the majority of the members take commutation, how can everyone be in poorer health? An example as set out in our paper was: if a scheme is making a 10% reduction to commutation rates for selection, using the typical mortality assumptions and assuming 80% of members take on the commutation, this is equivalent to assuming that life expectancy is more than 10 years lower for those who commute compared to those members who do not commute and equates to a mortality scaling factor of around 200% for members who do commute versus 45% for members who do not.
Allowing for de-risking caused much debate amongst the working party, and it is likely to cause much debate amongst the audience. We believe that it is appropriate to reflect de-risking to date. However, there may be exceptions where there has been a specific agreement between the Trustees and sponsor, when it is appropriate to not reflect de-risking to date in the setting of commutation rates.
Regarding allowances for future de-risking, where this is theoretical (so not formally documented or is contingent on future events) then it may be reasonable not to reflect this in the commutation rates. However, since this future de-risking might not be reflective in the cash equivalent transfer value basis, which as Kerry (Lindsay) explained is the theoretical starting point, you may not be deviating from that theoretical starting point. Where the de-risking is agreed, formally documented and not contingent on future events, then we believe it could be allowed for in the commutation rates.
We consider that if the scheme is under-funded on a commutation basis, then the commutation rates could be reduced to reflect this. This is like what would occur for transfer values in under-funded schemes. The company’s covenant should also be considered before reducing the commutation rates, in the same way that it is considered for transfer values. However, while members can transfer at any point up to one year before normal retirement age, or some schemes allow members to transfer out later, in practice they may have less choice as to when they retire. The decision-makers should therefore consider alleviating the impact of any stepped changes in commutation rates. We will touch on stepped changes in commutation rates later in this session.
Ms Lindsay: The next area we considered was intergenerational fairness. We believe it is important for actuaries to retain the same principles over time. For example, if you have always kept factors with reference to the transfer value basis, then it would not be appropriate to move away from that only at times where it is more favourable to the member. We also, as Jonathan (Hilsden) has just alluded earlier, believe it is important to avoid cliff edges in rates. It might therefore be appropriate to move closer to the theoretical start point in increments rather than in one step. Importantly, we did not think the argument of, “We have always kept rates low in the past, so we should keep them low now so that everyone gets the same per value over time.” is a reasonable argument from our members’ point of view.
The thematic review noted that market practices were commonly cited by actuaries as a reason for setting factors the way they are. Indeed, we all had experiences of seeing benchmarking data used to help justify certain commutation rates. Importantly, we are not completely against the use of market or other industry data, and we believe that this could be appropriate if you are comparing to similar schemes. However, we do not believe that benchmarking data should be used as a justification for keeping commutation rates low. Benchmarking data is, by its design, historic. There could be factors that have been in place for many years that may be coming up for review where the changes have not occurred yet, for example. It also often does not reflect the individual characteristics of specific schemes: for example, are they set by trustees or by the company, does the actuary have to certify the rates, and what are the pension increase types? Therefore, while it may be relevant in some cases, we believe that when using market data within your advice, its limitations should be made clear.
The next area we considered was the practicalities and process of members commuting pension for cash. We noted that some schemes, especially those schemes in the private sector, can be complicated and have a large numbers of different benefit tranches. In such cases, it may be sensible to simplify the number of factors used and combine factors for tranches where the pension increase rates are similar. This is particularly true where schemes are seeking to manage administration costs. However, most administrators now automate their retirement calculations and so they are capable of handling different factors impacting different tranches of benefits. In such cases, we therefore believe that administration simplicity should not be a driver in changing (or not changing) factors.
In providing their advice, actuaries should always be mindful of the balance of implementation costs against the impacts on member benefits. Where relevant, we believe actuaries should err on the side of the members. Finally, as Jonathan (Hilsden) has already touched upon, there were several other areas that the working party considered, the relevance of which varies by a scheme’s specific circumstances. These are covered further in our paper, and we would also be happy to touch upon any of them in the panel’s Questions and Answer (Q&A) at the end of this session.
Mr Hilsden: We agree with the thematic review that three years between commutation rate reviews should be the absolute maximum and not the default. We believe that actuaries should review commutation rates more regularly and a good practice would be to at least perform a high-level check on commutation rates annually. Given recent experience, pre-agreed triggers or circumstances which would lead to review could be beneficial, particularly in times of material movements in market conditions. Consulting actuaries would have to be mindful of the time and cost of these reviews as well.
We also wanted to touch on an area which we felt deserves more attention than it is perhaps currently receiving, which are market-related commutation rates (see Figure 2).
Figure 2. Market-related commutation rates.
We now have a poll question “Do you advise schemes (or are a trustee of a scheme) which have market-related commutation factors?”
The results of the poll show that 27% of the audience say “Yes they do advise or are involved in pension schemes with market-related commutation factors” and 73% say “No.” That is roughly a quarter to three-quarter split. There seems to be many schemes that do use market-related commutation factors, but also a number of schemes that do not. We understand that market-related commutation factors are more common with insurers who provide bulk annuities and schemes that are close to a full buy-in and are moving towards insurer rates - particularly where those buy-ins relate to deferred members. Market-related commutation rates may simplify valuation negotiations during the periodic rate review cycles and will likely reduce the potential of stepped changes between the commutation rates (although rates will be changing more regularly). This can make it more difficult for members to plan for retirement if the commutation rate is changing more regularly and could be more challenging to administer and manage the scheme. Some schemes may decide to guarantee the commutation rates for a short period of time (for example, it could be guaranteed for three months in a similar way to how transfer values are guaranteed).
Of course, the technical actuarial standards should be followed when providing advice (to trustees or corporates) on commutation rates.
A lot of the points on Figure 3 were highlighted in the thematic review, but we wanted to highlight and emphasise a few further points.
Figure 3. Presenting advice.
We believe that actuaries should set out their theoretical starting point and their reasons for moving away from this. Judith (Fish) on our working party also commented during our discussions that she likes to see in the advice from her scheme actuaries the impact on members’ benefits from the change in commutation rates. We think that the impact on funding, the actual basis used and a description of the decision-makers involved in setting the commutation rates should also be disclosed. As set out in the thematic review, if the scheme actuary has the authority to certify the commutation rates, we suggest this authority is provided in writing.
Ms Lindsay: To summarise the working party’s key conclusions: firstly we looked at an appropriate starting point for setting commutation rates and the allowances that should be made for selection risk, market volatility and other criteria. We believe that the theoretical starting point should be the best estimate, based on a scheme’s transfer value. In our paper, we do highlight several valid reasons why actuaries may advise trustees to deviate from the theoretical starting point. However, we did also note that several of the common reasons that we observed are being used without adequate justification. It is important that if you are moving away from the best estimate starting point, that you justify it for your scheme’s specific circumstances.
Secondly, we believe that three years should be the maximum time between commutation rate reviews. A more frequent high-level review may be appropriate, especially in times of significant market volatility or if there is a lot of changes happening in your scheme. We believe that there are many good reasons why schemes should consider setting market-related factors and we do expect this to be more common over time, particularly as schemes get closer to the insurance end game. We believe that it is appropriate for commutation rates to be reviewed, ideally during or immediately following a triennial actuarial valuation.
In presenting advice, actuaries should ensure that it is clear and as concise as possible, which I appreciate can often be a challenge. Any recommended changes should be highlighted. When you are proposing a change in factors, the impact on the benefits that appear in members’ retirement quotations should be highlighted explicitly.
Finally, as Jonathan (Hilsden) has mentioned earlier, if required under scheme rules, actuaries should provide their certification of the reasonableness of scheme factors in writing. We will now pause and pass over to David Gordon, who is going to cover more around the wider thematic review programme.
Mr D. A. Gordon, F.I.A.: Good afternoon, everyone. Can I first thank Jonathan (Hilsden), Kerry (Lindsay) and the rest of the working party for their informative and illuminating research. I joined the IFoA three years ago to lead the thematic review programme and, as we have been discussing, this research stemmed from my first thematic review report. Thank you, during the presentation today, for name-checking the report and indeed endorsing many of the recommendations.
As a Royal Charter profession, the IFoA has a duty in the public interest to regulate. This is carried out by the Regulatory Board, which sets the Actuarys’ Code, professional standards, and guidance materials. The thematic review programme (see Figure 4) reports to the Board, and this enables us to monitor the work (in specific areas) of actuaries. This in turn then informs the Board’s work.
Figure 4. Objectives of Thematic review programme.
The objectives of the programme are to showcase the best practices and to further improve the quality of actuarial work. It also provides evidence to refine standards, guidance, training and education. From the perspective of organisations and members, the benefits are simple. It is an opportunity for us to provide you with independent and confidential feedback on your approach and work. It also supports the IFoA to ensure that its regulatory work is based on evidence and reinforces our credibility as a member-led regulator in the public interest (see Figure 5).
Figure 5. Actuarial factors thematic review.
Now turning to the thematic review itself, I do have a thumbnail of the report cover. The report was published in December 2020, and we had a series of recommendations besides the one on which this research was based. I know that some of these recommendations have been highlighted already in the presentation tonight. We should not forget that the overall finding was that the actuarial advice we reviewed was generally of a very good standard. The recommendations we had were mostly centred around the benchmarking evidence we had of commutation rates being well below transfer values. We called on actuaries to continue following the existing standards fully, but in particular to explain why the two actuarial factors differ (especially when the differences are large) which is a Technical Actuarial Standards (TAS) requirement. The advice on the two sets of factors (commutation rates and transfers values) is often provided at the same time after the actuarial funding valuation, but the advice is often not tied together, but rather two distinct sections or even two reports. This should not take away the need to explain why they differ if they do indeed differ at all.
On reviewing rates regularly, this has clearly been well rehearsed this evening. In our review, we mostly saw triennial review cycles, and it is interesting from the poll earlier that a quarter of you have seen market-related rates, which is probably higher than what I observed in the thematic review.
On certification, around half the submissions we observed had commutation rates that needed to be certified by the actuary, yet we only saw one or two positive affirmative statements as required under the standards. I suspect practice on this has moved on since 2020, and I would continue my general recommendation to communicate assumptions and results better.
These member recommendations are simply to encourage actuaries to follow existing standards and guidance fully. I am pleased that so many of them are being endorsed earlier. The role of the thematic review was not to provide a critique on the technical aspects of setting commutation rates, hence the recommendation to set up the research tonight. In answering questions later, I will not be (even though I probably do have a view) expressing technical views. I will try to express my answer from the perspective of existing standards and guidance.
Finally, can I also commend our latest report, on corporate pensions advice to Defined Benefit (DB) pension scheme sponsors. To find out more about that report, we have a webinar next Thursday and you can find the QR code on the screen, which should hopefully allow you to register for the event. If that does not work, the event is still linkable from the front page of the IFoA website.
In conclusion, the quality of the advice we observed was good. We had several findings, as you can imagine, about some standards not being followed, many of which were aligned to the themes of the commutation rates review today.
To close, the thematic review programme relies on members and organisations to take part. In our six exercises to date, we have had over 50 organisations taking part. I would like to thank you all for your help with this, because we do rely on the organisations and members.
Mr Hilsden: We will now move onto the panel discussion where Kerry (Lindsay) and myself will be joined by Judith (Fish) and David (Gordon) and invite the audience to submit any comments or questions and Glyn (Bradley) will assist to filter and moderate the questions. So, over to you Glyn.
Moderator: I should begin by mentioning that the working party paper is available on the IFoA website. You can also find (on the events page) the details of that corporate thematic review next week that David (Gordon) mentioned earlier.
We have had a few questions come in, but if there are any more, please do put them in the Q&A panel and I will attempt to group them before putting them to the panellists.
To get the ball rolling, the focus of the paper advocated a theoretical point of view of best estimate rates. You have said that is not the only view one should recommend but that is where you would start as an actuary in your advice. I wondered, Kerry (Lindsay), what were the other options that the group had considered and why were they discounted?
Ms Lindsay: Yes, we did think about various options for the theoretical starting point whether it was: the transfer value basis, technical provisions basis, the insurer basis, self-sufficiency basis or something weaker than best estimate. There is a whole range of possibilities. Where we kept coming back to is the transfer value basis as a best estimate that offers fair value to members who are taking the option. Best estimate by definition is a 50% chance of those assumptions being realised. Members who take the option get fair value, but also importantly, members who do not get the option get fair value as well. If you had gone with something that is more prudent for the scheme, then that could cause a funding strain that weakens the funding position for any members who do not commute cash. But also, if members are commuting pension for cash, they are only commuting 25% of the value, so 75% of the value of their benefits is still in the scheme with a potentially worse funding position.
Moderator: I will take a different line now, unless anybody else on the panel wishes to come in on that. We have had a very pro-member stance on some of the topics in the paper. I am going to ask a pro-corporate question to Judith (Fish) who I think might be well positioned to respond. Commutation is seen by some as an option and a benefit design issue. If that was the case, why shouldn’t the employer be the ultimate arbiter of the commutation rates?
Mrs J. A. Fish, F.I.A.: For some schemes it is clearly part of the benefit design. You have schemes that have a pension and a cash sum written into the rules. But for most of the schemes that came out of thematic review, the commutation factors were decided by the trustees, either on the advice of the actuary or certified as reasonable by the actuary. I think for most cases, you are having to look at the actuary and the trustees to set those factors. Whilst the employer might have some strong views on what those factors should be, we must also remember the members probably do not understand the value of the pension they are giving up. For most schemes, we are seeing most members take pension commencement lump sums on retirement and therefore not selecting against the scheme. I think you must be very careful to look at the rules of the scheme in question when setting the commutation factors, to make sure that what you’re doing is reasonable and in line with the trust deed and legislation.
Moderator: Thank you. Anybody else on that point? Otherwise, I will try my question from a different angle. David (Gordon), you kicked all this off really with your thematic report, which I have recently read again and well worth a look at on the Institute website. There is a lot in there about advice on factors and not just relating to commutation. Generally, you mentioned the advice on factors was of a high standard but there was some variability, specifically in relation to commutation factors. If nothing else, when our members leave this webinar, what would you want them to do differently?
Mr Gordon: I would suggest the members follow your example by reading the report after they have read the research that we are talking about tonight. I think Jonathan (Hilsden) mentioned earlier the objective of today’s paper is designed to be thought-provoking and stimulate debate. Hopefully it does just that within your organisations. I think that would be fantastic. I will leave others to reflect on exactly what the right adjustments should be, but I very much support the conclusion that you have reached today.
Moderator: I will start with a couple of questions. I think a point was mentioned earlier about selection being a possible reason to adapt or to depart from the best estimate and someone said, “for consistency with transfer values, shouldn’t transfer value assumptions make some allowance for possible selection risk?” Who wants to take that one?
Mr Hilsden: Transfer values were not under the specific remit of the Commutation Rate Research Working Party, so we did not consider how to set them. It comes back to whether there is evidence that members who take transfer values have shorter life expectancy. You are guided by the transfer value regulations, which are very different to the commutation rate regulations. From that point of view, you would be setting your transfer value basis in relation to the characteristics of your scheme and of the mortality of your scheme. I do not know whether that is an answer which skirts round the issue. It was not really part of the commutation rate research, but similar principles might be considered.
Moderator: Just while we are on Transfer Value (TV) factors, somebody has asked how consistency would work where a transfer basis is set using a dual discount rate approach.
Ms Lindsay: I am happy to answer that one. Where you have dual discount rates, in relation to commutation rates, it is only the post-retirement discount rate that matters because the pre-retirement discount rate does not come into play.
Moderator: We have a couple questions around the theme of moving towards buyout, whether we would see schemes moving closer to insurer buyout terms. And if that’s a long-term objective, whether it would be sensible to set commutation factors using an insurer’s proxy basis.
Mr Hilsden: I think there is probably a journey towards buyout, both in terms of the scheme, but also in terms of the commutation rates that are set. It depends where the scheme is on that journey. One of the key considerations is whether the scheme is going to reach buyout when there are still deferred members to buyout. You might have more consideration if you have a deferred buy-in which the insurer will set commutation rates in a certain way and whether you should move your commutation rates in line with that. But if you have a process of securing buy-ins for pensioners after they have reached retirement, then there might be less of an impact on commutation rates.
Mrs Fish: I should add that commutation rates from an insurer potentially might not be better. You might find that as schemes have de-risked so much, you have increased your commutation rates to a point where they are higher than the insurer’s rates. But clearly, intergenerational fairness should also be considered so that you don’t get cliff edges.
Mr Gordon: In the thematic review, there was little discussion on buyout and annuity insurance terms. I suspect that was because it was completed during 2020 and based on evidence a year or two before in many cases.
Moderator: Okay, thank you. Now we have a volunteer to go live. If we can bring in Chris O’Brien please, to ask his question in person. Indeed, if anybody else wants to come on camera and as their question, please indicate this in the Q&A.
Mr C. D. O’Brien, F.I.A.: My question was partly answered by an earlier question from yourself. I was concerned that the paper did not include references to the alternatives to transfers value as the theoretical starting point. In particular, the number of references in literature to the use of bond yields to set commutation factors, for example, as is required by law of the Pension Protection Fund. In some cases, the scheme will be wholly invested in bonds and that might be the strategy from which the commutation factors calculated. In many cases it won’t be.
Moderator: Okay, thank you Chris. Sorry, your audio was breaking up a little bit, so I am going to repeat the question for the benefit of others. I think in summary, you were referring to some of the actuarial literature on setting factors using bond yields and you were asking the authors of the paper whether they had given more consideration to setting factors that way?
Ms Lindsay: We did not include in our paper, for want of space, is all the different options that we ruled out. We did discuss a number of options over and above the transfer value basis as a starting point. Jonathan (Hilsden) mentioned earlier the points around allowance for de-risking, and obviously bond yields can become more relevant there depending on the scheme’s investment strategy. As mentioned previously, transfer values were not part of the remit of the Working Party, but I know there can be lots of discussions when it comes to members’ transfer value basis and whether the value that a member receives should vary depending on how the scheme is invested. Indeed, the same principles would apply when you are talking about commutation rates, and whether one scheme should provide different commutation rates because it happens to be in a different investment strategy. This comes back to the point of what we are trying achieve, which is to provoke further discussion. We have come up with a theoretical starting point, but acknowledge that there may well be scheme specific reasons that justify moving away from it. Indeed, the specifics of the scheme’s investment strategy might be one of those.
Chris O’Brien: Thank you.
Moderator: Thank you very much for your question, Chris. I have got another challenge here from Neil Wharmby. He makes the point that the paper is putting forward the viewpoint to start theoretically with the transfer value, which is closely prescribed in legislation on how they are to be performed, which therefore forces you into a best estimate position. Commutation factors on the other hand are not subject to the same legislative environment. Is this a question of, whether we are trying to compare apples and pears when we start with that view? Further, he asks, what is the rationale for adopting a market-based set of assumptions? It is in fact a brilliant question because you have three questions in one, so perhaps different people may want to chip in. The third question is whether intergenerational issues lead one away from a market-based approach? So, who wants to tackle at least part of that first question? Perhaps Judith (Fish) you have a view from a trustee perspective?
Mrs Fish: The first bit around whether we start with the transfer value basis. We noted that there were legislative differences around the transfer value basis, for which there is legislation, and the lack of legislation around commutation factors. We started with the transfer value basis, not because of its legislative background, but because it is meant to be a best estimate basis. We felt that it was the right starting point when converting members’ benefits to an alternative form. So, when you are doing other exercises where you are converting member benefits, you might often start with the transfer value basis Although I thought in the thematic review it was very interesting to see the comparison between the transfer and the commutation values (which were significantly lower). I am sure that most members of pension schemes won’t realise the huge differences in potential value between a transfer and commutation value. Somebody else said earlier that most members take commutation, but far fewer members take transfer values. Commutation is something that members do value as an important part of their benefits.
Ms Lindsay: I can take the second part of the question. It is important to know that we are not necessarily suggesting you have to go down a market-based set of assumptions. We are more highlighting the fact that it should be considered whether a market-based or fixed basis is appropriate for a specific scheme in the time horizon that you are working in. Especially as schemes get closer to insurance annuity purchase, if the insurance factors are market-related, then this might give more justification for moving towards a market-related basis to avoid future volatility in your funding. We are, however, not necessarily of the view one way or the other. The point around intergenerational fairness is interesting and I agree there is an argument that, just because one person is retiring at a point when bond yields have fluctuated, should they get a higher or lower lump sum depending on the market fluctuations? It is a challenging question and it is a similar discussion that happens with the topic of transfer values as well, especially now that you have freedom and choice. The significance of transfer values was around them being a best estimate rather than, as Judith mentioned, the legislative impact. So, it is not about trying to play one set of members off against another.
Moderator: Whilst we are on the topic of intergenerational fairness, we have had a question similar to what was raised earlier on why should those that retire later get a better deal because they have retired later? The other way round you could pose the question, is why should you hold members (or people that have left before) back in order to be fairer? Does holding down rates for the future make it better for the people who’ve perhaps had low rates in the past? Kerry (Lindsay), do you want to elaborate more on that tension?
Ms Lindsay: I think it is quite a challenging one. The holding down of rates in the future because people may have or have not had them in the past. Just because rates have been one way for one group of people does not mean you should not be doing the right thing for your current members. Ultimately, trustees have a duty to act in the best interests of members. I think that’s a really important point. Generations, never mind rates, do experience differences. Longevity improves over time. Someone retiring 30 years ago and taking cash might have been expected to live for a much shorter period than someone retiring now and taking cash. That in itself would lead to a different lump sum. The point around de-risking as schemes get closer to the insurance end game, is an important one, and as Jonathan (Hilsden) alluded to, we had a lot of debate within the working party around how best to allow for this. I do not think we had a conclusive view.
Mr Hilsden: The de-risking journey was an area we discussed a lot, allowing for the fact that schemes are maturing, maybe towards lower risk investment strategies, and whether that should (or should not) benefit commutation factors. This caused a lot of debate and I think one that continues to do so.
Mr Gordon: A reminder of the TAS requirement to explain the rationale for differences between assumptions used for different factors. I would say that it is critical so that the trustees understand the implications of what they are being advised.
Moderator: One questioner said, as factors are generally moving upwards, and actuaries are being asked to certify that, what difficulties might that present for actuaries in relation to certifications in the past where rates were much lower?
Mr Gordon: My observation is that actuaries were not certifying in the past. That is not a reason, just as holding down rates to be fair for people in the past is not either. I do not think not certifying now because you didn’t certify before is a good explanation. Certification is about the facts and circumstances at the time, just as if you certify in the past, it would be about the facts and circumstances at that time.
Moderator: We have time for one more question. This one has come up quite a lot, related to where people leave a deliberate margin, say 10% below the value either on the best estimate or technical provisions basis. How does the panel feel about that as an alternative way of setting factors? Is that building on what you’ve done or is it in contradiction to what you are suggesting?
Mr Hilsden: There are probably two things in the question. An assumption for how many members will commute and what basis those members will commute on is a funding assumption and is required to be prudent. Setting that rate as something like a percentage of technical revisions seems sensible. It will allow for your commutation rates to change without there necessarily being a funding strain, because the technical provisions will not change. I do not think it necessarily means one should set commutation rates in the same way. It is a funding assumption versus the commutation rates that determine how much members get by way of their pension commencement lump sum. You could still use the theoretical starting point of a transfer value, which is a best estimate basis, and which might be less prudent than the funding valuation assumption.
Moderator: It is coming up to the end of our time so I would like to thank everybody on the panel, and also everybody that has joined us today, and particularly those who have put forward questions. I am sorry we were not able to cover them all. I would like to remind everybody to have a look at the paper on the IFoA’s website which you will find on the events page for today. If nothing else, the next time we all author or peer review a factors paper, hopefully we will make sure to go back and challenge ourselves that, besides ticking all the boxes, could we do more to help our clients and members understand our advice. With that, I would like to thank you all very much again and close the session.