Skip to main content
Loading…
Chamber and committees

Scottish Parliamentary Pension Scheme Committee, 11 Mar 2008

Meeting date: Tuesday, March 11, 2008


Contents


Scottish Parliamentary Pension Scheme Inquiry

The Convener:

Under agenda item 2, we will take evidence in our Scottish parliamentary pension scheme inquiry. I welcome Grant Ballantine, who is our sole witness today. He is a senior consulting actuary in the Government Actuary's Department.

We will move straight to questions, of which we have many. I shall ask the first. What, in general, does the Government Actuary's Department do? To whom do you give your wisdom?

Grant Ballantine (Government Actuary's Department):

Our main role in the Government Actuary's Department is to provide actuarial advice on the Scottish parliamentary pension scheme. Normally, such advice would be provided to a set of trustees or, in the absence of trustees of unfunded schemes, to the managers of those schemes. I suppose that technically, our SPPS client is the Scottish Parliamentary Corporate Body, but in practice we deal with matters through the Executive. We see the SPCB once every two or three years when we produce reports. With respect to the advice that we provide, our main function is to carry out a full actuarial valuation every three years, which determines the contribution rate that the SPCB is to pay. In addition, we have on-going work in supplying factors and providing financial advice on the running of the scheme.

Obviously, you provide advice on a range of schemes other than the Scottish parliamentary pension scheme—you deal with other similar Government schemes.

Grant Ballantine:

GAD's main function is to provide actuarial advice to public sector bodies, particularly on pensions. We also do a bit of insurance work, for example, but we deal mainly with pension schemes. We provide advice on most of the main unfunded public service schemes, such as those for Scottish teachers and the Scottish national health service. We also advise the Westminster Parliament, National Assembly for Wales and Northern Ireland Assembly bodies. We provide advice on the parliamentary contributory pension scheme at Westminster, the National Assembly for Wales's scheme and the Northern Ireland Assembly's scheme. We fulfil much the same function for them as we fulfil for the SPPS.

When any of those bodies is considering changing the rules, as we are, is your advice simply a matter of saying, "This is what you can do and it will cost you such-and-such"?

Grant Ballantine:

Essentially, that is it. Our advice has two aspects. Our basic duty is to provide costings for changes or possible changes that our client—the trustees, the Assembly or whoever—initiates. We are also sometimes asked more generally for advice: what would be sensible, what the options are and what other bodies do in similar situations.

In your first answer, you mentioned the three-year actuarial valuation. You also produce an annual report. What is the difference between the two?

Grant Ballantine:

By the annual report, do you mean the costings that are produced for the SPCB accounts?

Yes.

Grant Ballantine:

That function is rather different. What the SPCB does in that respect is what private sector companies are required to do by the Accounting Standards Board, which lays down reporting requirements for United Kingdom companies. About six, seven or eight years ago, the board issued financial reporting standard 17, which in effect required private sector companies to report annually their pension liabilities and assets and prescribed the basis on which that had to be done. Public sector bodies such as the SPCB have decided to do something analogous, although they are not required to do so—not by the ASB, anyway. The annual publication to which you referred contains a numerical calculation that we do, consistent with what the ASB requires of private sector companies. It is therefore made on a somewhat different basis from the on-going valuation.

The Convener:

In the valuation to 31 March 2005, to which you referred, you recommended an increase in the employer's contribution rate to 0.8 per cent to cover the accumulated deficit. I understand that that recommendation was accepted and implemented, yet the deficit that was reported in the accounts was £4.86 million. Does that represent a contradiction?

Grant Ballantine:

The calculations are made on different bases. The approach that the ASB prescribes for private sector companies, which we follow for the Parliament in the annual financial report, is driven by a bond-type approach. The discount rates are driven by the yield on corporate bonds and the ASB requires liabilities to be discounted using the yield on AA-rated corporate bonds. That is not as conservative as a gilt rate, but it is much more conservative than what most pension schemes would hope to earn. From a funded pension scheme that has substantial investment in equities and other alternatives such as property and infrastructure, trustees would hope to generate extra returns on top of bond returns.

In an on-going actuarial valuation, it is quite common to take some credit for the expected outperformance of equities and other assets relative to bonds. That justifies the use of a somewhat higher discount rate than the discount rate of AA-rated corporate bonds. The main reason for the different results is that the liabilities on the annual accounting requirement are discounted at a lower interest rate than is assumed to be justifiable for an on-going valuation.

For those of us who are worried about the health—or otherwise—of our scheme, are you saying in effect that we should not look closely at the annual accounts but that we should worry, or not worry, about the three-year valuation?

Grant Ballantine:

Yes. To say that you should not worry might be an overstatement, but the three-year actuarial valuation is the more important determinant of the SPPS's long-term wellbeing. That is what drives the cash contributions from the employer, or the SPCB in this case.

Naturally, the annual financial reporting is of some relevance. It is an answer that is received on a particular basis and it is, depending on how interest rates change relative to equities, likely to be extremely volatile from one year to the next. Members might have seen press reports about assessments of private sector funding schemes, which within the past 12 months have fluctuated from a surplus of £30 billion to an overall deficit of £90 billion. If you take such a short-term approach to assessment of liabilities, you can get very big swings in a relatively short time.

The Convener:

That is helpful.

As you might be aware, the Review Body on Senior Salaries has carried out some work on various matters relating to our Westminster colleagues and has, with regard to pensions, suggested that cost sharing be introduced to cover future shortfalls. How would such a system operate if it was introduced in the Scottish Parliament?

Grant Ballantine:

That is a very good question. I am not sure that anyone knows exactly how such a system might apply.

I think that that has been said about one or two of the senior salaries review body's recommendations, but never mind.

Grant Ballantine:

Of course, it is not only the SSRB that is going down that road: the UK Government has already implemented cost sharing in the main public service schemes. The teachers' system has already been agreed in principle, the NHS and the civil service are following closely behind and a consultation document has been issued on the local government scheme.

It is pretty clear what the SSRB is getting at. Any variation in the assessed costs above or below a certain base level—which is, as far as I am aware, pretty vaguely defined as far as the parliamentary contributory pension fund is concerned—is shared 50:50 or 40:60, as agreed through an adjustment of contribution rates or benefits. It has also been suggested that with cost sharing up to a certain level there should be an absolute cap on employer contributions. With the civil service pension scheme, for example, there is a proposal to cap employer contributions at 20 per cent—the SSRB has proposed the same for the PCPF—and I believe that the schemes for teachers and the NHS have a 14 per cent cap on employer contributions.

That is a broad outline. An awful lot of detail has to be sorted out before we can really get to the nitty-gritty of any likely impact.

So, simply talking about cost sharing does not really tell you very much at all. It all depends on how the costs are shared.

Grant Ballantine:

I suppose so. It tells you something, but you have to establish first the baseline and secondly whether, for example, past service deficits, investment fluctuations or fluctuations in pensioner liabilities that impact on active members will be taken into account. Once you have done all that, you have to decide whether costs will be shared 50:50 or 40:60; whether cost sharing will be applied through an increase in contributions; whether it will affect the future service of active members, which means that pensioners and deferreds will get away scot free if costs increase or will benefit if there is a surplus; whether there is an option for the parties involved to decide between a mix of benefit reduction and contribution increases; and how any cap will be applied. The SSRB has suggested that, if costs break the 20 per cent cap, the scheme in question should be fundamentally reviewed.

As I have said, a lot of detail has to be sorted out before we can say exactly how such a system will apply. However, the principle is that, depending on the baseline, costs might be split 50:50 between the member and the employer up to a certain level, after which the member contributes 100 per cent.

The Convener:

Because our scheme is relatively young, it was decided that there should be a pooled fund arrangement for investing our assets. Is such an approach still appropriate? Unless there have been sudden changes in the stock market, I believe that our assets are approaching £18 million.

Grant Ballantine:

Yes, that is still appropriate at that level, although that is not to say that you could not go down another route if you so wished. The SPCB's current investment manager, Baillie Gifford, has indicated that it would not normally consider a pooled fund for anything less than £30 million, but it is prepared to make an exception for the SPPS. You can still take a mix-and-match approach, using different pooled funds and asset strategies. The pooled fund for a relatively small scheme is generally about economies of scale and benefits of diversification that would be difficult to replicate with a segregated direct portfolio within the UK equity market, for example.

Okay. We now have some questions on funding.

David McLetchie (Edinburgh Pentlands) (Con):

As you may be aware, the committee is considering giving members the option of accruing pensions at one fortieth of their salary for each year of service in addition to continuing with the existing increments of one fiftieth for accruals. It is intended that, if a member chooses the accrual rate of one fortieth, that will be fully funded by the member's own contributions and not in any way by employer—that is, taxpayer—contributions. What additional contribution would be required of scheme members over and above the existing 6 per cent if they were to move from an accrual rate of one fiftieth to an accrual rate of one fortieth?

Grant Ballantine:

We have provided a costing on that point and have recommended an extra contribution of around 5 per cent of payroll for the extra benefit.

David McLetchie:

That is higher than the additional 4 per cent that was charged of members of Parliament at Westminster when they similarly changed from an accrual rate of one fiftieth to an accrual rate of one fortieth. Can you explain, perhaps by reference to the underlying schemes or the composition of the members, why the extra cost is necessary for this scheme?

Grant Ballantine:

Yes. The main reason for the difference is the wretched subject of retained benefits. As you will know, retained benefits are benefits that the member has acquired through membership of a pension scheme in previous employment. Under the rules of both the Westminster scheme and the Scottish scheme as they existed before the Finance Act 2004, members' benefits that were accrued in the SPPS or the PCPF were subject to an overriding revenue limit that, in broad terms, could not exceed two thirds of their final salary less retained benefits from other pension schemes. A lot of the members at Westminster have retained benefits from other schemes, and some have substantial retained benefits in relation even to a Westminster MP's pay level. For example, former barristers can have personal pension pots of £1 million, £2 million or £3 million. The fact that they have significant retained benefits has an impact on the benefits that they can draw from the PCPF.

It is generally quite time consuming to get information on previous benefits. One has to ask the member to provide the information with respect to retained benefits, which is quite a time-consuming operation. The member does not usually want to be bothered with digging into the past and so on. The trustees at Westminster had to mount quite an exercise to get members to tell them what retained benefits they have. They got information back from about half the members, which enabled us to make an assessment of the level of the retained benefits. The existence of those retained benefits has an impact on the costing of the Westminster scheme and it has an impact on the costing of the option to move from an accrual rate of one fiftieth to an accrual rate of one fortieth.

Obviously, if you are accruing benefits at one fortieth rather than one fiftieth, you will hit the limit of two thirds of final pay for fewer years of service than otherwise—broadly speaking, it would be 26 and two thirds years compared with 33 and a third years. Of course, the member might opt for the one fortieth option initially without expecting to have a long career as a member of Parliament but find out, at the end of the day, that he has been paying money for no benefit, because the retained benefits kick in.

We had some hard information in relation to retained benefits in Westminster. That enabled us to say that members would, in general, be provided with lesser benefits if they opted for a fortieth than one might expect from simply looking at the difference between a fortieth and a fiftieth. That is the major difference between the 5 per cent and 4 per cent costing.

In Scotland, we have not got any information on retained benefits. It is not necessary to get that information in relation to the young scheme. As the convener mentioned, the retained benefits do not impact until members retire.

We have costed the 5 per cent figure for the SPPS without any allowance for the impact of retained benefits.

Can you explain why the basic employer contribution rate at Westminster is lower than the equivalent rate here, when the additional amounts to cover previous scheme deficits are excluded?

Grant Ballantine:

If you exclude the deficits, you are looking at something like 19.5 per cent against something like 18.5 per cent or 18.1 per cent—there is a difference of about 1 per cent or just over. That is primarily due to the different demographic assumptions that we have made. The PCPF valuation was carried out on the same date as the SPPS valuation—effectively, 31 March 2005. In both cases, we used similar discount rates, which were, broadly, 3.5 per cent in excess of price increases and 2 per cent in excess of earnings increases, both for the SPPS and the PCPF. There is no difference there; the difference is entirely on the demographic side.

Looking at the comparison, the biggest single factor—which does not account for the whole difference, but might account for half of the difference or a bit more—is the assumed pattern of retirement ages, which is the age at which the pensions commence. Again, because the Westminster scheme has existed for quite some time, we have good evidence that a significant proportion of members of Parliament work beyond the normal retirement age of 65, or 60, on unreduced benefits with 20 years of service. Some of them, as you probably know, continue to be MPs into their 70s, and even into their 80s. Based on that evidence, we were able to make a credible assumption that a proportion of members will continue working beyond normal retirement age.

We do not have that evidence in relation to the SPPS as yet, although it might eventually emerge. For the initial costings of the scheme, we have adopted a prudent approach of assuming that members will retire when they reach 65, or at the end of the parliamentary session after they reach 65, and that some will take advantage of the early retirement provisions and go at 60 on unreduced benefits, once they have 20 years of service.

The main difference between the two costings is the earlier retirement age that has been assumed in relation to the SPPS and the somewhat later retirement age in relation to the PCPF.

David McLetchie:

Speaking of older members—one of whom I will become at some point—people's life expectancy is increasing, particularly males. What impact do you expect that to have on scheme funding?

In addition, there is a difference between the proportions of females and males in the Scottish Parliament and the proportions in Westminster. The Scottish Parliament has a far higher proportion of female members than the Westminster Parliament has. Given the different compositions of the Parliaments, what assumptions do you make about life expectancies?

Grant Ballantine:

You are right that women are more expensive than men.

I am glad that you said that, although it is borne out by my personal experience.

We certainly could not comment.

Grant Ballantine:

The higher proportion of women is a minor factor. This Parliament has a higher proportion of women, so there is a slightly more expensive mix of members here than there is at Westminster, but it is not a major issue.

Mr McLetchie's question on life expectancy is highly relevant. One might say in a one-line answer that the impact of life expectancy is potentially worrying. There seems to be no end to the pace at which life expectancy is increasing. Old codgers are constantly defying the actuarial expectations.

For the 2005 valuation of the SPPS, we used the latest tables published by the actuarial profession—the 1992 continuous mortality investigation bureau tables. Those consist of pension tables for males and females, and built into them are improvement factors to allow for future improvement. For the 2005 valuation, therefore, we used the 1992 tables projected forward to 2005 with the built-in improvement factors. We also applied the same improvement factors to all future years. So we used the tables in 2005, and incorporated the improvement factors that were considered appropriate when the 1992 tables were drawn up. The improvement factors vary by age and sex, but, by and large, the improvement factors are less than 1 per cent per annum, and involve mortality rates dropping by 1 per cent per annum, or a bit less, indefinitely in the future.

Experience of the past five years or so, looking at wider bodies of experience than is available from either the SPPS or the PCPF, such as big pension schemes or the population as a whole, indicates that, at some ages, mortality improvement is up to 3 per cent per annum rather than 1 per cent per annum. In other words, mortality improvement at the older ages of 65 and above is happening significantly faster than it was even 10 years ago. In that context, since the 2005 assessment, the Office for National Statistics, which takes some advice from GAD on the projections, has produced 2006-based population projections that have built in to them significantly greater assumed mortality improvements than any previous set of tables has had. In addition, the actuarial profession has worked on updated tables that were produced in 2000, although they do not have a specific assumed allowance for future improvement.

Also, much work has been done on causes of death and on how improvement might be affected in the future. Of course, nobody knows how improvement will be affected—it is a guessing game—but all the recent evidence suggests that improvement will accelerate even faster than in the past. If that happens—good heavens—we will be paying pensions for 100 years, if we are not careful. There is therefore a real worry about longevity.

Another factor is that the Pensions Regulator has just issued a consultation document on the minimum improvements that it would be looking for.

Having regard to all that, it is beyond doubt that we will have to incorporate a larger allowance for greater assumed improvement at the 2008 assessment than we did at the 2005 assessment, moving from 1 per cent to not necessarily 3 per cent, but something a lot bigger than 1 per cent. In an extreme scenario, 2 per cent of pay could easily be added to pension costs in a longer-term assessment.

So 2 per cent of salaries could be added to costs as a result of improvements, although such an amount would be borne between employer and employee.

Grant Ballantine:

That is right. I have not done any calculations, but the figure would be of that order.

David McLetchie:

I would like to move on to other changes that would impact on costs. At the moment, there is no provision for unmarried partners of members. It has been proposed that the scheme could be expanded to cover not only members' spouses but members who are in civil partnerships and members who are cohabiting in a long-term relationship but are not formally married or in a formal civil partnership. From your experience of other schemes, what costs are associated with extending the categories of surviving partners and spouses?

Grant Ballantine:

As you say, the scheme provides for legal spouses' benefits. Since 2005, I think, it has also covered civil partners, to some extent.

The costs of extending the scheme to cover unmarried partners largely depend on what is and is not included in the definition of an unmarried partner. If there is a wide and lax definition, any member who does not have a spouse or civil partner will almost have the option of selecting any individual and claiming that they have an unmarried partner. The resulting costs can be significant—they can be 3 or 4 per cent of pay.

Most schemes that have gone down that route have opted for a fairly tight definition of an unmarried partner, and have tried to limit the scope of the definition to those who are in relationships that are similar to those that spouses or civil partners are in. In other words, there must be a long-term or permanent commitment that has been established for some time. If the scheme is structured in such a way and is policed fairly regularly, the costs can represent less than 1 per cent of pay.

David McLetchie:

Currently, many people in society are in cohabiting relationships as opposed to married relationships—I am thinking for the moment about people in heterosexual relationships. In the past, a member would have the choice of being married or single. From an actuarial standpoint, is the previous proportion of people who were married now split between formally married people and cohabitees? Is the former married total the equivalent of the present married total plus the present cohabiting total?

Grant Ballantine:

That is a good question, but there are not enough data on small schemes to get a credible answer to it. However, we should consider the population statistics as a whole. It is true that the proportion of younger people who are married has declined and that the trend is extending, so that the proportion of people in their middle ages—even those in their 60s and so on—who are married is declining. It seems that the trend in the past 20 or 30 years is that lower proportions of people are marrying. By extrapolation, that might be expected to impact on people of all ages in the next 30 years or so. It is reasonable to postulate that those who are not married to their partners will fill the gap resulting from the drop in the proportion of people who are married, although some partnerships will be caught and brought under the new definition that did not come under the old definition. The falling proportion of those who are marrying will be offset somewhat, but the offset will not be complete.

David McLetchie:

At the moment, the pension given to a surviving spouse or civil partner terminates when they remarry or enter into a new civil partnership. Would removing that rule and continuing to pay a pension to a remarried or repartnered spouse or civil partner be a particularly expensive burden on the scheme? Is that regarded as significant?

Grant Ballantine:

It is not significant in relation to past experience. If behaviour does not change and people do not manipulate the system, the cost is relatively small—of the order of a quarter of a per cent or one half of a per cent of pay. However, you have to be careful about what happens if you remove the cessation of a spouse's pension upon remarriage, particularly if you include unmarried partners, as the cessation of an unmarried partner's partnership is difficult to identify without being intrusive and seeking a lot of evidence. That situation can be difficult to operate, which is why one or two schemes did away with the cessation provision when they opened up to unmarried partners.

However, if you remove the cessation provision, there is a danger that you might encourage people to identify an unmarried partner or marry a partner as they approach the end of their lives—it is a free hit against the scheme, particularly if you do not have to marry anyone but can simply nominate someone. To try to limit that risk, some schemes reduce the level of the spouse's pension if there is a great age gap between the member and the partner. For example, if the age gap is, say, more than five years, the level of the spouse's pension is reduced by 2.5 per cent for each year of the age difference, which means that with a 40-year age gap, the spouse's pension would be practically nothing. That is common in the private sector, and one or two public schemes have started to do it as well.

From the celebrity pages, we can all think of relationships that have such a wide age gap, but I do not know if we are the sort of people who are likely to fall into that category.

Speak for yourself, Mr McLetchie.

David McLetchie:

This is a lot more entertaining than I thought it was going to be.

What would be the cost of increasing the age limit for a child's pension from 22 to 23 and of introducing a dependant's pension for children who become incapacitated at age 23 or over?

Grant Ballantine:

The cost of paying a child's pension for an extra year, for those in full-time education, is minor. It does not alter the funding rate. It is way under 0.1 per cent or something.

The question of extending the coverage to dependent children is somewhat more problematical. There are some similarities to what I said before about selection against the scheme. The issue is identifying whether the child was dependent. It is, perhaps, not too difficult for a member to claim that he is supporting a child who just chooses not to work, for example. You need to be pretty confident that you can establish that a real illness or disability prevents the child from working and makes them dependent on the member. That is particularly the case if we say simply that the child must be dependent on the member at the date of the member's death, which sort of writes off what happens between the child being 22 or 23 and being 45, if that is when the member dies. Having a reasonably tight definition or criterion for what qualifies, such as the child being permanently and totally disabled from the age of 23 until the member's death, should ensure that costs are kept relatively small.

So the focus should be on long-term or near-permanent incapacity to minimise costs.

Grant Ballantine:

That is right.

The scheme's current practice of reinsuring members' death-in-service benefits apparently costs £48,000 in premiums. Should we continue that practice or should the scheme take on its own burden in that respect?

Grant Ballantine:

It was sensible to insure the lump sum death benefit in the scheme's early years, as it protected the fund's position when there were very few assets. However, now that the fund stands at £18 million, the scheme's asset base is probably of sufficient size that the death benefits should be self-insured, assuming, of course, that the benefit stays roughly at the current level of three times pay.

So do you recommend that the scheme should review that matter imminently, annually or during the triennial review?

Grant Ballantine:

It should not be necessary to review it annually, but it could be considered in conjunction with or immediately after the triennial review.

Hugh O'Donnell:

Under the current scheme rules, office-holders who are not MSPs are excluded from a number of benefits, such as buying added years and the enhancement to survivors' pensions on a member's death in service. What extra costs would be involved in giving those office-holders the same rights and entitlements?

Grant Ballantine:

I want to go back a stage to postulate the rationale behind that approach, which, of course, follows the Westminster approach. Indeed, the initial provisions in the SPPS were almost a carbon copy of the Westminster provisions.

It was felt that office-holders' earnings were often temporary, transient and volatile; sometimes they went up and sometimes they went down. Moreover, ministers might get sacked, come back, get sacked again and come back again. The nature of the job is different to that of a permanent, full-time MP or MSP.

Because of that volatility, you could get some capricious results with benefits. It can be down to happenstance, for example, whether someone dies very high up or very low down the Cabinet ranks, and one might be faced with difficult cases in which a junior minister just promoted to Cabinet minister appears to get a windfall benefit while a Cabinet minister who is demoted loses out entirely. That capriciousness was the reason for excluding benefits that, depending on pay at a particular point, might be substantial. The pay of an MP or MSP is, on the other hand, fairly stable. Of course, you could extend those rights and entitlements to office-holders, as long as you were prepared to accept the volatile nature of the potential outcomes.

As for the enhancements that you mentioned, I have to say that there have not been many cases of death in service or ill-health retirement. If the incidence of such cases were to remain low, the costs in relation to office-holders would be quite modest.

Added years are a trickier issue. The ministerial salary tends to be temporary and to fluctuate, so the concept of an added year has limited relevance. However, it might be possible to allow office-holders to buy an added year of MSP service.

Hugh O’Donnell:

You will know that we are considering other changes. What financial savings could be made if the First Minister and the Presiding Officer did not have special pension arrangements and were instead treated as office-holders in the current SPPS scheme?

Grant Ballantine:

We were asked to provide costs for that, on the basis of the First Minister and the Presiding Officer continuing in their roles for a four-year parliamentary session. On current pay levels, the net saving from a change from the current position to the position that you describe would be round about £700,000 for the First Minister and about £270,000 for the Presiding Officer.

That is nearly £1 million.

Grant Ballantine:

Yes, over a four-year period.

On a different topic, what would be the cost of removing the earnings cap but retaining the two-thirds pension limit?

Grant Ballantine:

The cost would be virtually nothing in the medium term. I understand that, at present, no one would be above the earnings cap, which I think is about £112,000 at present. The First Minister might just be getting there.

The extra contributions would apply to earnings above the cap. There would of course be a cost—the member and the employer would pay—but the impact on the overall percentage contribution rate would be minimal. There could, however, be longer-term costs—compared with the current situation—if earnings were to increase significantly faster than the earnings cap.

There is some merit in the proposal, as all relevant earnings would be pensionable. The earnings cap might be difficult to justify 10 years down the road. It will not be imposed by HM Revenue and Customs but will be self-imposed by the scheme. Unless you had a pretty good rationale for keeping it, removing it might be a convenient simplification.

Can you confirm that increasing the maximum commutation limit to 25 per cent of pension in line with the revenue limits would be cost neutral to the scheme?

Grant Ballantine:

Yes—it would be intended to be cost neutral, or as near to that as we could make it in terms of the factors that would be used to convert pension to lump sum. As with any option, there is always a risk of possible selection. Converting pension to lump sum is a good idea for pensioners who are in poor health, because they are not likely to live as long as the average. Practice has shown that a very high proportion of the membership takes up the lump sum option because of the tax advantages. The scope for selection is therefore fairly limited, and we regard the measure as pretty well cost neutral.

If a scheme member buys added years, subject to the current limit in the scheme of 15 per cent of salary, is there a cost to the scheme?

Grant Ballantine:

There is not intended to be a cost. The added-years factors are intended to be cost neutral, with full costs to the member, not to the employer. However, the actuary does not always get his sums right.

That is a bold admission to make on the record.

Grant Ballantine:

You might have deduced it from my earlier comments about life expectancy increasing.

With the best will in the world, nobody can predict the future of human demography and people's financial circumstances. All that the actuary can do is make reasonable assumptions. The added-years facility is not expected to have a cost, but there would still be a risk to the employer—the SPCB in this case. If it turned out to be more expensive, the SPCB would have to pick up the bill. For that reason, most employers who offer the added-years facility try to limit the scale of the option in order to limit the scale of the risk—it is more a limitation of risk than of cost. If it were a completely unlimited option, the pension scheme would be converted into something that is like an insurance company that offers benefits on pseudo-commercial terms, and that is not really the scheme's purpose.

Given those comments, am I right in saying that if the cap were removed completely and we retained only the two-thirds pension limit, would the scheme be exposed to a higher level of risk?

Grant Ballantine:

Yes, although you might want to consider having tighter limits for the 59-year-old who has done only one term—or been employed for four years—and, if he can afford it, wants to buy 20 added years.

Knowing what each year costs to buy, I see that as an extremely unlikely scenario.

Have you any suggestions for what the limits might be, Mr Ballantine? What would be reasonable?

Grant Ballantine:

An easy limit would be to control the contribution input rather than the number of years, in addition to having the two-thirds limit. You could also relax significantly the existing contribution limit—at something like 20 per cent of pay—without exposing the scheme to any great risk. It should look reasonable in relation to a member's pay. You do not want to get to the stage at which a member who has other resources contributes 60 per cent of pay. I suggest that a limit of something like 20 or 25 per cent of pay would be reasonable.

Hugh O’Donnell:

Thank you for your patience with me; this is my final question. With the removal by the Finance Act 2004 of the maximum amount of pension—two thirds of final salary—that can be accrued, is there any reason to restrict the amount of pension payable from the fund?

Grant Ballantine:

Yes, one could advance some reasons. First, the more pension provided, the more the cost. Initially, the cost might not seem very large, particularly in a very young scheme such as the SPPS, because it is not likely that many members will reach the two-thirds limit for many years, whereas a significant minority of the Westminster scheme members are above the limit.

Secondly, the reason why the SSRB and others involved were prepared to recommend an accrual rate a bit above the average—either a fiftieth or even a fortieth—was to provide a faster accrual of benefits for members who are exposed to career interruption in a way that a normal employee is not. Members of Parliament and MSPs are likely to have their careers interrupted for reasons that, in the main, are nothing to do with them.

A nice caveat.

Grant Ballantine:

Shorter-than-expected average service was one reason that justified the higher accrual rate. If you justify the higher accrual rate on the basis of shorter-than-expected service and then give that higher accrual rate to those who are lucky enough to stay in for the long term, they might get two bites at the cherry.

Peter Peacock (Highlands and Islands) (Lab):

I must confess that the triennial valuation report has not been high on my reading list, but I took the trouble to read it a couple of weeks ago and found it fascinating, as I am finding your evidence fascinating.

It is clear from the report and your evidence that managing the funds requires you to take a long-term view to ensure that the funds are sufficiently buoyant over a long time to meet all their obligations. In fact, you said earlier that you apply improvement factors to "all future years", which seems to mean eternity. You also said that because the SPPS scheme is new—it is immature, in that sense—there is not the same actuarial experience underpinning assumptions as there is with the Westminster scheme.

I want to apply some of that to our situation to get a complete grasp of it. My understanding of the historical and current situation is that members' contributions are fixed and that any variation in the fund's buoyancy is dealt with by the employer's contribution, which maintains the fund. However, if the fund outperformed the actuarial assumptions, there could be a reduction in the employer's contribution such that the fund then underperformed in relation to the assumptions—or a variation could cause an underperformance—meaning that the employer's contribution might have to increase again.

I understand that you use the triennial report to check whether the actuarial assumptions remain valid, so you make a judgment only every third year about whether a pattern is emerging. For example, the long staying-on rates experienced at Westminster, which are beginning to show themselves here, mean that you must make certain actuarial assumptions. Is that the basic process? Every third year, you check the actuarial assumptions against performance.

Grant Ballantine:

That is absolutely the case.

Peter Peacock:

Another point that comes out from the report is the huge number of factors that affect the fund. You have touched on some of them: market performance; real interest rates; the balance between men and women in the fund; the fund's assets; life expectancy; age of retiral; and transfers in and out. In your experience, does one factor dominate decisions about changes in the employer's contribution every third year? Does market performance dominate everything else, or are the factors evenly spread?

Grant Ballantine:

On the assumptions that the actuary makes at the three-yearly assessment, the critical factor is the discount rate—the assumed investment return against the real investment return, and the assumed investment return against excessive price increases.

Is that affected by a factor of many times?

Grant Ballantine:

Yes, the weighting is about 50 or 60 per cent.

The second most critical factor is the longevity assumption. If you sort that and the discount rate, the rest of the factors are unlikely to have a huge impact, although they will have some impact.

In terms of experience, anything can happen. However, the biggest factor that impacts on experience from one three-year period to the next is generally investment performance. If it is a good time for the markets, that will have a favourable impact on the fund; if it is a bad time for the markets, that will have a bad impact on the fund. In most cases, that is the single biggest experience factor.

Peter Peacock:

I noted that, although the report refers to the issues that I want to come on to—early retiral, ill-health retiral and widow's and widower's pensions—none is mentioned as a specific factor in its own right. I assume that that is because it is assumed that members of the Scottish Parliament will retire, on average, at 64 rather than at 65. Does that assumption embrace the actuarial assumptions about early retiral, ill-health retiral and widow's and widower's pensions? Is my assumption broadly fair?

Grant Ballantine:

Yes. Dealing first with the early retirement point, the assumptions that we have made are broadly equivalent to assuming an average retirement age of 64. That is an amalgam of people whom we assume will retire at the normal retirement age of 65 and a few whom we assume will carry on until the end of the parliamentary session, when they will be 64, 67 or whatever. We also take into account that individuals could retire at 60 with, for example, 20 years of service. That would give rise to extra costs. We said that 64 was the equivalent average age. However, we valued the pensions for those who could qualify for an unreduced pension at 60 as if they would retire at 60.

Peter Peacock:

The headline average figure is 64 for the Parliament, but various assumptions are caught up in that. Under the scheme, nobody currently qualifies for early retirement because of the requirement for 15 years of service. If someone has not served 15 years—and none of us could have—they will not qualify. However, as the years go by, more and more people will meet the 15-years requirement, will have reached the age of 60, and will have made sufficient contributions. Ten years from now, more people will be eligible for early retirement. Do the costings and the average figure of 64 under the current scheme take that into account? Is that a correct assumption on my part?

Grant Ballantine:

Not quite. You mentioned earlier that the actuary has to take a long-term view. We look at the current membership and come up with a hypothesis of when they are going to retire—say in 15 or 20 years' time. For example, for a 50-year-old who already had five or six years of past service, we would say that he might be able to retire at 63 on an unreduced pension. We would factor that into our calculations. In a way, that is not quite considering the worst possible case but it is looking at the prudent or costly case. Therefore, unless the mix of membership were to change, we would not expect an increase in costs just because existing members have served an additional period in the Parliament, because such costs have already been built in. Of course, things might also go the other way: it is possible that, over time, people will reach 60 with 20 years of service but will want to continue being an MSP.

The built-in costing assumptions would be reviewed during a triennial review. If it were found that patterns were changing, would you have to consider changing the actuarial assumptions?

Grant Ballantine:

Yes.

Peter Peacock:

If, for example, you discovered in 15 years' time during a triennial review that more people were retiring early than you had assumed, would you consider that as a single factor in relation to what the employer's contribution might be, or, indeed, what the contribution of both employee and employer might be if costs were shared in future? Alternatively, would you balance the factor of an increase in early retirements against the factor of market performance during the same triennial period, or against other factors such as transfers or a shift in the balance of men and women in the fund? Would all the factors go into the melting pot to allow you to say, "Right, in the next three-year period, we will have to adjust the contributions by X?"

Grant Ballantine:

At the present time, we certainly take the latter approach: the rules do not allow contributions to be varied for one particular factor. All the factors are lumped together and it is the end result that applies. However, we consider each individual factor when we analyse the overall experience to see whether anything has to be changed. Within reason, we can identify the impact of individual factors on the financial position of the scheme. Heaven help us if, in future, we have to vary contribution rates to take account of every demographic element.

Peter Peacock:

Indeed.

Will you clarify one point for me? At Westminster, people tend to serve for longer than the assumed length of service here, and they therefore collect less pension over their life. Is the average retirement age in the Westminster scheme reckoned to be, say, 64.3 years, or 65.1 years? Is there a variance between the two schemes? Roughly, what effect would a variance of 0.5 per cent have on costs?

Grant Ballantine:

I do not have the average retirement age for PCPF members at my fingertips.

Perhaps you could find that out in the future.

Grant Ballantine:

Yes. A one year difference in the average retirement age might affect costs by around 0.5 per cent of pay.

So a 1 per cent shift in the average age—

Grant Ballantine:

I am talking about a one year shift in the retirement age. If the average retirement age is 64 rather than 65, the standard contribution rate would increase by around 0.5 per cent of pay.

Peter Peacock:

Okay. Thanks.

We have taken evidence from people in other parts of the pensions administration system on what is happening with other funds. What criteria are used in public sector schemes to determine qualification for early retiral? Is there a standard set of rules or do the rules vary, in your experience?

Grant Ballantine:

Most public and private sector schemes have a two-year qualifying period before a person becomes entitled to a preserved pension, but once they get past that period there is no qualifying period for normal retirement, early retirement or late retirement. That applies across the board. It is the qualifying period in the parliamentary scheme that is unusual.

What is the rationale for having a minimum qualifying period of 15 years before early retirement is available under the Scottish parliamentary pension scheme and the Westminster scheme?

Grant Ballantine:

That takes us back to the possibility that the service of an MP or MSP may be short or that their career may be interrupted. I assume that the rationale was to provide individuals who had given long service to Parliament with a favourable option of going early with an extra pension, but I do not think that the rationale for that facility exists nowadays. I will put things in another way. There is an opposite pressure nowadays. Given people's increasing longevity and the increasing proportion of the population who are likely to be aged over 65 in the next 30 years, people should be encouraged to stay at work and the system should be made to reward people who continue to work rather than those who retire early.

Peter Peacock:

We have heard evidence about people choosing to retire. You alluded to the fact that MSPs and MPs often do not have a choice about when they retire—would that things were different, but they are not. We have heard evidence about other schemes. Under the local government scheme, for example, if there is a reorganisation and people are made redundant—I am not talking about people in local government who want to change their lifestyle and leave early—there may not be an actuarial reduction in their pensions. I understand that the teachers pension scheme is broadly similar. Are there standard rules on actuarial reductions for people who voluntarily retire early?

Grant Ballantine:

Yes. In most voluntary early retirement schemes nowadays, actuarial reduction factors apply so that the scheme is not financially exposed by members' individual choices. Normally, there would be a full actuarial reduction.

Is that the norm in most public pension schemes?

Grant Ballantine:

Yes.

Okay, but we are assuming that the person has retired voluntarily.

Grant Ballantine:

Yes. The position is different in redundancy cases.

What would the actuarial reduction be if it happened under the SPPS?

Grant Ballantine:

Different schemes have different approaches, but by and large a reduction factor of between 4 and 5 per cent a year would be cost neutral. Quite a lot of schemes have a 4 per cent reduction factor. That would be fair under the SPPS.

Peter Peacock:

We asked a range of questions on ill-health retiral in our consultation. The difficulty lies in determining the severity of an illness and the ability or otherwise of the person to conduct any other comparable form of occupation. What is the broad experience in other schemes? Being a member of Parliament is not typical employment, but what is your view on the issue, in particular on what other schemes offer?

Grant Ballantine:

Ill-health retirement in the public services has been a thorny issue for some time. I am talking about the main public service schemes, rather than the parliamentary schemes. As you rightly say, the nature of employment is quite different for members. You will be aware that, at least four years ago, there was a substantial volume of press complaints about the scale of ill-health retirement, particularly in the police service and the fire service and, to some extent, in local government and other public services.

The Treasury set up a working party on ill-health retirement and made various recommendations to tighten up the qualification of ill-health retirement in public service schemes. A contrast was made with private sector schemes, in which the incidence of ill-health retirement seems to be much lower than it is in public service schemes. The main public service schemes have had a torrid time on the ill-health retirement front. However, the steps that were taken five or six years ago have been helpful in limiting the incidence of ill-health retirement. Ill-health retirement needs to be kept under control. If it is not administered properly, it can run away with itself.

For whatever reason, the parliamentary schemes have never really been subject to the same sort of difficulty. I think that that is because most parliamentarians are committed to their work and are not on the lookout for ill-health retirement in the same way that people in some other organisations might be. The incidence of ill-health retirement has been relatively low among members.

That being the experience at Westminster, presumably that is reflected in the actuarial discussion that we had earlier about the figures that are used for long-term projections.

Grant Ballantine:

Yes.

Peter Peacock:

A couple of weeks ago, we took evidence from the chair of the board of trustees to the Westminster scheme. He spoke about a review of the ill-health provisions there and discussed a tiered system, which involved an absolute inability to work in any occupation, an ability to work but not at the same level, and so on—I do not have the details at my fingertips. The chair of the board reckoned that such an arrangement could save the Westminster scheme about 0.4 per cent of payroll. If we were to introduce a similar arrangement, would it be fair to make a broadly similar assumption about the saving on payroll here relative to where we are now, starting off at the same position as Westminster?

Grant Ballantine:

That would be a reasonable assumption. The saving would probably be slightly smaller, however. At present, you have a one fiftieth accrual rate, whereas most of the Westminster MPs have a one fortieth accrual rate. The saving might be 0.3 per cent of payroll, rather than 0.4 per cent.

Peter Peacock:

You mentioned the need to keep tight control. In your experience, having observed how these matters work, might such a system be best achieved through having a clear tiered arrangement, or, rather than having clarity in that way, could the trustees of the fund—assuming that we move to a position of having trustees—be given significant discretion to review the payment of ill-health pensions and to seek continuing evidence of that ill health? What is your feel for the best way of controlling it?

Grant Ballantine:

It would be useful to be able to do both. A two-tier approach is being considered for the PCPF, and it is already in place, using similar arrangements, in the civil service pension scheme. It was a development of the attempt to adopt a more controlled process, with a tighter definition for ill-health retirement. Another factor has been the tighter definition that the Inland Revenue imposed through the Finance Act 2004. That approach has been initiated by those pressures. It is sensible to have a two-tier structure in which the most generous benefit is given in cases in which permanent incapacity is established very clearly and a lesser benefit is given to own-job disability, if we can call it that.

Peter Peacock:

In public life, there have been remarkable recoveries from conditions such as Alzheimer's disease. Do you argue that, notwithstanding the tier provisions that are put in place, we should maintain an ability to review the evidence on any condition from time to time?

Grant Ballantine:

Yes. It would be sensible for the trustees, or whoever, at least to have the power to review not only the member's state of health, but his earning capacity. If he can still fool the doctors that he is disabled, but he is earning megabucks, that may raise questions about whether he is a genuine ill-health case.

Peter Peacock:

You have touched obliquely on my next point. Is there any actuarial experience of—or, from observing funds and talking to people who are involved in the business do you know of—a link between ill-health provisions and early retirement provisions in that, if early retirement is difficult to get, the rate of ill-health retirement mysteriously rises? Is there any known or believed link or association between the two?

Grant Ballantine:

There is certainly anecdotal evidence to that effect. The joke in the local government context used to be that people first tried for ill-health retirement and if they did not get that, they tried for redundancy—and if they did not get that they were forced to accept voluntary early retirement. However, there has been no credible study to try to show that link.

Peter Peacock:

I understand completely your point that the decision of one member can be in effect a cost to all the other members, unless we are careful to get the balances right. However, to reflect other comments that you have made, it is to an extent the nature of any pension fund that some people who stay on longer in work than others, perhaps beyond their normal retirement age, in effect, pay for benefits that others will enjoy. So we men—we all happen to be men on this committee—in effect pay for benefits that women will enjoy for longer than we will. To an extent, we all pay for collective benefits that we all share. There is no way of not taking such a collective view, although I accept your point that we should not precipitate situations that have unnecessary consequences. Is that a fair comment?

Grant Ballantine:

That is absolutely right. One of the great merits of defined benefit pension provision is that there is a pooling of risks. The pensioner who dies below the average age in effect subsidises the pensioner who lives for a long time. The person who gets an expensive ill-health pension at the age of 40 and who lives to be 90 benefits from the other members of the scheme. There is a pooling of risk, but it should be done in a way that prevents individuals from getting an obvious hit against the scheme. There should be the pooling of unknown risks that are common to all, but individual in operation.

That has been very helpful.

The Convener:

As there are no further questions, I thank Mr Ballantine for his evidence. As somebody who many years ago had a brief and inglorious short career as an actuarial trainee, I was impressed by how succinct, comprehensive and interesting your evidence was. I think that all my colleagues were, too. We may be in touch with you by letter, e-mail or telephone to clear up one or two points but, in the meantime, thank you for your time.

Grant Ballantine:

I am happy to respond to any queries. Thank you for your kind comments.

We move into private session.

Meeting continued in private until 17:02.