Legal & General Assurance Society Ltd v CCA Stationery Ltd

This is an appeal brought by Legal & General Assurance Society Limited (“L&G”) from a Determination by the Pensions Ombudsman dated 9 December 2002 (“the Determination”). The Determination arose out of complaints of alleged maladministration by L&G of the CCA Stationery Limited Pension Scheme and Assurance Scheme (“the Scheme”). The primary complaints were made by CCA Stationery Limited (“CCA”) which is the Principal Employer of the Scheme. The other complainant was a former employee of CCA. Neither of the complainants was represented before me or has supplied skeleton arguments in support of the Determination. It is convenient to start by setting out the factual background. In doing so, I draw heavily on the very full skeleton arguments and appendices prepared by L&G’s counsel, Mr Nigel Inglis-Jones QC and Mr Richard Hitchcock. I am particularly grateful to them for the help they have given me during this appeal.

In the 1970’s and 1980’s, L&G offered a type of with-profits insurance policy which was designed to be attractive to trustees of small to medium sized, final salary, occupational pension schemes. It was known as the AF80 policy. It provided some protection against insolvency problems on an ongoing basis. So long as trustees contributed to the policy at the rate recommended by L&G, there ought to be sufficient assets to cover accrued liabilities. L&G wrote somewhere in the region of 200 such policies. One of them (“CCA AF80”) was entered into by the Trustees of the Scheme. That is the policy in relation to which the complaints to the Pensions Ombudsman arise.

AF80 contracts form part of L&G’s long term business. As a proprietary Life Office, L&G’s insurance business is regulated by the Insurance Act 1982, ss 28-31A of which provide for a strict separation of assets in respect of “long term business” and “general business” so as to protect the funds available for long term policyholders. A discrete long term fund must be established, and assets representing a company’s long term business are to be applicable only for the purpose of that business and may not be transferred so as to be available for other purposes.

In accordance with the requirements of the 1982 Act, all with-profits business written by L&G is accounted for within the Life and Annuity section of its Long Term Fund. The assets which notionally underlie an individual AF80 policy, therefore, form one small part of a mass of AF8O assets which lie in the with-profits segment of the Life and Annuity section within the Long Term Fund. There are other with-profits products besides AF80s in that segment. The assets ascribed to the totality of with-profits contracts form part of a far larger mass of assets, the remainder of which have no direct influence upon the AF8O contracts or other with-profits products, but which also lie within the Long Term Fund. The assets underlying the CCA AF80, for example, were in the region of £3m. However the value of the assets attributable just to the totality of AF80 contracts written by L&G was in the region of £1 Billion.

Because AF80s are with-profits contracts, the grantees are entitled to participate in L&G’s profits. This is achieved by accruals by way of Contractual Annual Increments, (“CAIs”) and growth bonuses, income bonuses and (at times) special bonuses which are applied annually to each AF8O. CAIs are measured by the yield on High Coupon 15 and 20 year British Government Stocks, while bonuses are calculated by reference to the performance across the life and annuity section of the Long Term Fund. The CAIs are one aspect of the guarantees applying to the AF8O policy: the increment must be credited each year, and this means that the range of investments used by the Society to back their ongoing liability under the AF8O product must take this guarantee into account. This has typically meant that a 50% fixed interest content has been ascribed to the AF8O product to service this guarantee. Bonuses are discretionary, in the sense to be explained below, but they have invariably been credited every year during the life of the AF8O product.

The remaining 50% of the ascribed contents of the funds underlying AF80 policies consists of equities and commercial properties. Over the long term they are expected to produce greater returns than the fixed interest investments but they are more volatile and therefore more risky at least in the short term. As is the case with all with-profits policies, a key approach of L&G in declaring bonuses on AF8O policies is to maintain a “smoothed” investment performance, insulated from the vagaries of the market in which the assets underlying the policies are invested. This has been particularly valuable for the trustees and employers of pension schemes using the insured route as their investment strategy. This smoothing of the investment performance means that for normal extractions, that is the month-by-month withdrawals which are made in the ordinary course of business to meet benefits payable under an AF80 contract, the grantees and the principal employer benefit from a stable contribution rate, and a relatively consistent level of growth on premiums invested, irrespective of market conditions and the fluctuating value of the assets underlying the nominal value. This can represent particularly good value when such extractions are made during a period when the underlying value of the assets is low relative to the nominal value of the policy. For example I was told that in 1990, when significant negative investment returns were achieved by L&G’s Long Term Fund (in common with most such funds), a total yield of +12.7% was credited to all its AF80 policies, including CCA AF80.

It will be appreciated that, as a result of this smoothing procedure, a gap will frequently exist between the nominal value of a policy, as expressed in the form of the Total Cash Pool (which will include all the declared bonuses referred to above), and the underlying value of the policy which relates to the value of the assets corresponding to the nominal value. At times the underlying value will exceed the nominal value and at other times it will be the other way round.

It is adjustment of the bonuses which allows L&G to smooth the performance of AF80 contracts. In deciding what the bonuses will be for any particular year, L&G must take into account many factors including the investment gains or losses attributed to the mix of assets ascribed to the family of AF80 policies, certain costs, tax liabilities and so on. Furthermore it will have to take into account its assessment of how the values of the ascribed assets are likely to fluctuate in the future. Therefore the setting of the bonuses is not an exact science. In this sense it is within the discretion of L&G.

One of the principle objectives of the AF8O product is the protection it affords, during the currency of the policy, against volatility which could affect both investment and contribution rates. For example L&G, like its competitors, offers other products which are more volatile and which can, depending on the way relevant markets move, give greater returns. But these are normally at the cost of greater risk. For example Managed Funds will contain a much greater equity content. If the markets in which a Managed Fund is invested perform well, so will the Fund. On the other hand if, in the recent past, a Managed Fund had invested heavily in internet based companies, it will have performed very poorly indeed.

Thus far, I have considered, in general terms, how AF80 contracts work during the period when premiums are being paid and normal extractions are being made. The policy is intended to be, and is operated as, a long term investment. However the terms of an AF80 contract allow for discontinuance. The contract does not include a guaranteed surrender value but it contains a number of discontinuance options any one of which can be opted for by the policy holder. If the policy holder decides that none of the options are suitable, he can decide not to discontinue and can revisit that possibility at a later date when, for example, market conditions have changed and one or other of the discontinuance options appears to be more attractive.

For present purposes, it is only necessary to refer to one of the discontinuance options which is referred to as “Option (ii)”. This is a “Cash on Discontinuance” (“COD”) option. The relevant provision in the contract reads, so far as material, as follows:

“[In specified circumstances] the Grantees shall be entitled upon giving written notice to the Society to exercise one of the following options:-

(ii)

to require that the Society shall immediately reduce the Total Cash Pool by the whole of the amount thereof, and shall pay to the Grantees either (a) a single cash sum calculated on the basis currently in use by the Society for this purpose or (b) at the option of the Society a series of payments made over a period not exceeding 24 months which are in the opinion of the Society equal in value to the said single cash sum …”

The grantees are free to ask for this form of discontinuance if they wish. The choice is theirs. If they do so, L&G must calculate the due sum “on the basis currently in use by” L&G. This is a sum which principally (but not only) reflects the current market value of the assets underlying the Total Cash Pool. As mentioned above, as a result of the smoothing procedure applied by L&G, a gap will frequently exist between the nominal value and underlying value of the funds ascribed to an AF80 policy. Sometimes the COD option will result in a payment to the grantees of a sum which is larger than the nominal value of the cash pool. At other times, the COD will be smaller than the nominal value. This is particularly likely to be the case where the COD option is being contemplated at a time when markets have been performing very badly. The smoothing operation will have given the policy a notional value (on the basis of which normal extractions will be calculated and paid) which exceeds the momentary value of the underlying assets ascribed to the policy.

In the case of the CCA AF80 contract, the Trustees of the Scheme chose to discontinue and selected Option (ii) in April 1992. This was a particularly unfavourable choice. The nominal value of the Total Cash Pool in CCA AF80 at that time was in the region of £3.7m whereas they were advised that the COD was about £3m. Although the Trustees could have chosen one of the other options or to continue with the contract, they decided not to do so. It is the gap between the nominal value and the COD which gave rise to the complaints at issue here. As Mr Inglis-Jones put it, at its heart, CCA’s complaint was that the surrender value paid out by L&G was too low and unfair.

At the heart of the issues I have to decide is the question of how L&G calculate the surrender value of an AF80 contract on discontinuance. I have already described how the Total Cash Pool is calculated and how the smoothing operation is likely to create a gap between the latter and the “real” value of the underlying assets. When a contract is terminated and the grantee opts for the COD, a figure for the “real” value of the underlying assets has to be arrived at. This is not a simple matter of identifying the assets purchased with the premiums paid under the contract and then assessing their current value. An AF80 contract does not, and can not, work like that. When, for example, a premium is paid under CCA AF80, it is not used to purchase a discrete and identifiable asset, still less is it split so that 50% is invested in fixed interest assets and the remaining 50% used to purchase equities and commercial property. The premiums (and, for example, dividends from equities) all are paid into the Long Term Fund of L&G and are used there to purchase assets for the whole of that Fund. For the same reasons, it is not possible to identify individual costs which relate to the purchase of specific assets for a specific AF80 contract; nor will it be possible to identify individual costs which relate to the sale of specific assets when normal extractions are made. These policies are part of a web of mutually supporting and sustaining assets. The AF8O policyholder will benefit from participating in a large fund with all the advantages of investment which enures to such a fund. The beneficiaries under individual, comparatively small, AF80 contracts, will obtain the benefits of sharing in the buying power which L&G wields through operation of its Long Term Fund. As Mr Inglis-Jones put it, each AF80 contract benefits from the mutual support provided by being within a large fund. This is referred to as “synergy”.

It will be appreciated that when a grantee decides to discontinue and to take the surrender value of the policy, it is necessary not only to try to attribute to the policy a value of that proportion of the assets within the Long Term Fund which is to be attributed to the policy but to take into account notional sales costs and the negative impact of loss of synergy. L&G’s obligations in relation to its Long Term Fund are not simply to those discontinuing an AF80 contract, but to all those other investors who are continuing their long term investment. Discontinuance performed under Option (ii) requires more than simple disinvestment by the sale of assets; it involves the unravelling of the policy in question from the complex mechanism of assumptions, procedures and guarantees which provides mutual support for all AF8Os (and other products, the assets generated by which are invested in the Long Term Fund), in order to proceed towards the net asset share of the policy concerned. It may require the sale of a part of the assets of the Long Term Fund and thus notional sale costs have to be attributed on each discontinuance, together with the expenses incurred by that discontinuance. To pay out any more than an amount that reflects the net asset share on discontinuance would be in derogation of L&G’s duty to other policyholders participating in the Long Term Fund, and to its shareholders. It must always go behind the Total Cash Pool which, for reasons explained above, is an ‘artificial’ figure arrived at by the smoothing process.

The way in which the COD is arrived at is as follows. An AF80 contract involves payment of premiums. Those premiums are attributed each year to a “cash pool”. Thus an AF80 commenced in 1988 will have a 1988 pool, a 1989 pool and so on. It is into each of these pools that CAIs and bonuses are notionally paid. The Total Cash Pool, which is the nominal (smoothed) value of the investment out of which normal extractions are made, is the sum of these annual cash pools. However, L&G, like its competitors, may be faced at any time and with little notice with a request for immediate surrender of a policy. It therefore has to have available a means for attributing a value to all of the policies within the Long Term Fund. To calculate the net asset share figure at any given time, the Society employs a series of monthly value adjustment factors, (“MVAFs “). As the name suggests, they are reviewed and adjusted monthly. These relate to specific annual pools and, given the pooled nature of with-profits insurance contracts, are calculated on a portfolio wide basis rather than a policy wide basis, thus reflecting the fortunes of the overall investment of the Long Term Fund. In other words, at any given time (say, for example, April 1992) there will be a MVAF for each year of the AF8O pools within the entire AF8O portfolio. Each MVAF will have been adapted from the March 1992 MVAF to take account of market conditions in the intervening month, as the March 1992 MVAFs were from the February 1992 MVAFs and so on. It should be pointed out that the MVAFs apply not just to the totality of contracts within the AF80 portfolio but also to funds underlying L&G with-profits unit trust business as well. As I have noted above, the total value of assets within the former is in the region of £1 Billion. Mr Inglis-Jones was not in a position to put a figure on the unit trust assets, but they were, by comparison to the AF80 assets, substantial. It must be emphasised that the MVAFs are not produced on a policy by policy basis.

As explained above, the calculation of MVAFs is an iterative process, the April MVAFs being derived from the March MVAFs and so on. Mr Inglis-Jones told me that from time to time the MVAFs are calculated afresh by reference directly to the value of the underlying assets, but this is only done infrequently and to correct any tendency of the iterative process described above to gradually build up inaccuracies. In other words, by this means the MVAFs are, from time to time, ‘zeroed’ to ensure accuracy.

It will be appreciated that MVAFs are a Long Term Fund-wide device for transforming the nominal value of individual pools to a figure which more nearly represents the value of their net assets. They are adjusted from month to month to ensure their accuracy and they become for each month the figure by which the nominal value of the Annual Cash Pools must be multiplied in order to produce the net asset share. MVAFs may be either greater than or less than 1, so that the net asset share is, almost always, either more than or less than the nominal value. Mr Inglis-Jones told me that a substantial number of AF8Os administered by L&G in the later part of the 1990s had COD values in excess of their nominal value, and that a number of such policies have been discontinued, in circumstances in which the grantees have chosen Option (ii), and have received a cash surrender value which exceeds the nominal value of their policy. This illustrates the gap between nominal and COD values, and thus the importance of the time at which grantees choose to discontinue.

The assessment of the MVAF for any particular month is not a matter of science. It involves taking into account a variety of factors. Just as the size and longevity of typical with-profit funds facilitates the smoothing approach which is a characteristic of this type of contract, so the abrupt discontinuance performed under Option (ii) requires more than simple disinvestment by the sale of assets: it involves the unravelling of the policy in question from the mutual support between it and all other AF8Os (and other products, the assets generated by which are invested in the Long Term Fund), in order to proceed towards the net asset share of the policy concerned. It may require the sale of a part of the assets of the Long Term Fund and thus notional sale costs have to be attributed on each discontinuance, together with the expenses incurred by that discontinuance. It will include, for example, the cost to the individual AF80 contract and to the remainder of the funds in the AF80/unit trust pool of the unravelling of the synergy referred to above.

Thus it can be seen that the assessment of the surrender value involves two distinct inputs. First there is an evaluation, involving balancing many factors, of the MVAFs which apply to the whole of the Long Term Fund which contains the assets attributable to AF80 contracts and unit trust business. This involves judgment based on reconciling many factors. This exercise is carried out without regard to the particular features of individual AF80 contracts. Thus the same process of assessment and judgment would have to be carried out whether or not a particular AF80 policy, such as CCA AF80, existed and whether or not it was being discontinued. The second part is the assessment of the surrender value which is a distinct stage. This consists of applying the MVAFs to the annual cash pools for an individual AF80 contract. This is a mathematical calculation. Someone who is mathematically competent can check whether the multiplication has been carried out successfully.

Referring back to the relevant terms of CCA AF80 set out in paragraph 11 above, a grantee who decides to discontinue the contract and chooses COD under Option (ii) is entitled to a cash sum calculated “on the basis currently in use” by L&G. As explained above that basis involves the multiplication of the sums in the annual cash pools by the relevant MVAFs.

There is one additional matter which is of some importance to this appeal which should be referred to at this stage. The AF80 included more than just an investment vehicle for use in pension schemes. It was a package comprising investment, administration, funding and documentary services. Thus L&G also became the administrator of the Scheme in which CCA AF80 was used as the investment vehicle. The question of when L&G ceased to act as administrators of the Scheme is a matter which will have to be considered later.

Having said something about the nature of the CCA AF80 contract, it will be helpful to give an outline of the history which led to this appeal.

The CCA AF80 contract had been running for some 12 years when, in early February 1992, L&G were notified that CCA had been acquired by a Mr Scholes. On 26 February 1992, PJC Pension Services (“PJC”) wrote to a letter to L&G enclosing a letter dated 24 February from Mr Scholes to L&G. The enclosed letter was in the following terms:

“Dear Sirs

CCA STATIONERY LIMITED PENSION & INSURANCE SCHEME

We hereby authorise you to release any information that may be necessary to P.J.C Pension Services in connection with the above scheme. This authorisation should be taken to include all pension arrangements with Legal & General …

In addition we hereby inform you that with immediate effect P.J.C. Pension Services have been appointed as Administrators of the above scheme with authority to negotiate with you as to the discontinuance terms on the Scheme’s Pension contracts with yourselves.

You are also instructed to negotiate terms with P.J.C. Pension Services as to the continuation or otherwise of the scheme investments currently held by the Scheme with yourselves.”

The covering letter from PJC included the following:

“You will note that we are authorised to contact you and to take over with immediate effect all administration services on the above scheme.

More immediately, I should be grateful if you could please let me have your discontinuance terms for a) moving the administration from Legal & General, b) moving the administration but leaving the investment in Legal & General’s managed fund and c) a total withdrawal of all moneys and administration from Legal & General and the cessation of the pension contract.

As we are to take over administration with immediate effect could you please let me have copies of all historical data on all members of the pension scheme including salary histories and contracted-out earnings histories. We also need to know the up-to-date value of the AF80 contract.”

In the following correspondence, L&G pointed out on a number of occasions that there was a difference between the nominal value of the funds in the CCA AF80 contract and the value of the underlying assets and that it was the latter which forms the basis for the surrender value on discontinuance. L&G supplied reports which set out an approximation of the sums which would be realised were PJC to take Option (ii). Notwithstanding this, PJC chose the latter option. It protested that L&G was trying to impose a penalty for early closure of the contract – an allegation strongly denied by L&G. I do not understand it now to be suggested that a penalty was imposed. In any event, on 7 August 1992, PJC wrote to L&G stating, amongst other things, that the CCA AF80 contract was to be treated as discontinued as and from 24 February, as was L&G’s engagement as administrator of the scheme. On 30 March 1993, L&G informed PJC that the cash surrender value of the policy was slightly more than £2M. That surrender value was calculated as of April 1992, the date of discontinuance which PJC had requested.

CCA (and presumably PJC) were not content with the cash surrender value. It complained to the Pensions Ombudsman on 10 March 1998. At about the same time a complaint was made by a former employee of CCA. The Pensions Ombudsman started to deal with the complaints. This resulted in two sets of proceedings being commenced by L&G. The first was an appeal to the High Court. In it, amongst other things, L&G challenged a preliminary ruling by the Pensions Ombudsman that the fact, if such it be, that L&G was no longer responsible for the management of the Scheme at the relevant time does not affect the Ombudsman’s jurisdiction over former managers. The second were proceedings for Judicial Review. The main issue in the latter was whether it was within the remit of the Pensions Ombudsman to investigate that part of the complaint which relates to the fairness or otherwise of L&G’s surrender terms.

Both matters came before Lightman J. His judgment is reported (Legal & General Assurance Society Ltd v Pensions Ombudsman [2000] 2 All ER 7). Neither CCA nor the other complainant were represented. The Pensions Ombudsman was. Among other things, Lightman J held as follows:

“23.

The remit of the ombudsman is to investigate maladministration by L&G acting in the capacity of a manager. The concept of ‘maladministration’ is broad and includes bias, inattention, delay, incompetence, ineptitude, perversity, turpitude and arbitrariness: see R v Local Commissioner for Administration for the North and East Area of England ex p Bradford Metropolitan City Council [1979] 2 All ER 881 at 898, [1979] QB 287, at 311-312. It is concerned with the decision-making process rather than the merits of a decision. I agree with Mr Inglis-Jones QC (for L&G) that the jurisdiction to investigate maladministration does not enable the ombudsman to reopen and consider the merits or fairness of the terms of the 1980 contract, which is a commercial contract between the trustees and L&G. The terms (and in particular the terms for discontinuance and the formula) may turn out to be advantageous or disadvantageous for the scheme members. But neither the entry into the 1980 contract nor its continuance in force constitutes an act of management, let alone within the purview of the ombudsman: nor in the ordinary course can the question of the perceived fairness or unfairness of the exercise by L&G of contractual rights granted by the 1980 contract to L&G for its own benefit fall within the remit of the ombudsman. …

The ombudsman concedes that he can afford no relief to CCA if he finds unfairness, that he cannot alter the terms of the 1980 contract or vary the sum payable thereunder on discontinuance. It is plain that he has no greater power in this regard than the court: see Edge v Pensions Ombudsman [1999] 4 All ER 546. None the less he maintains that he is free to investigate this issue. In my judgment, the fairness of the terms is outside his remit and L&G ought not be subjected to an investigation on this topic.”

The reference to “the formula” in this passage is a reference to the MVAFs applied to the annual cash pools. Lightman J then turned to the issue of management:

“24.

I turn back to the third issue on the appeal. The short question raised is whether the process of the calculation of the sum payable on discontinuance made on 13 March 1992 is capable of constituting maladministration by L&G. In my view the refusal on the part of L&G to disclose the formula was in breach of the terms of the 1980 contract. The principle must be clear that (in absence of some contractual provision to the contrary) where a contract provides for a payment to be made calculated in accordance with a formula known to one party alone, that party must disclose the formula to the other party: one party cannot require the other to accept his calculation made in accordance with a formula without such disclosure and to accept his figure in blind faith that the calculation is correct. There is no provision in the 1980 contract requiring the trustees to accept the calculation by L&G without disclosure of the formula and an opportunity to check the correctness of the calculation. If the calculation of the sum payable under the 1980 contract is an act of management by L&G, then it does seem to me open to the ombudsman to investigate the conduct of L&G as possible administration.”

This paragraph is of significance. Lightman J distinguished between the contractual obligations on L&G under the CCA AF80 and the consequence of breach. As far as the former is concerned, he held that where a payment is to be “calculated in accordance with a formula” there must be a term in the contract that the payee is entitled to see the formula to ensure that “the calculation is correct”. However the existence of such a, presumably implied, term does not mean that its breach constitutes maladministration within the Pensions Ombudsman’s remit. That is only the case if the calculation is an act of management.

On the latter point, Lightman J said the following:

“25.

It is apparent that the full facts regarding the cesser of management functions by L&G have yet to be explored and it is not possible to say what stage had been reached on 23 March 1992. I am not prepared to hold that the calculation was not an act of management by L&G and that it is not an area within the remit of the ombudsman. Indeed there is much to be said for holding that the calculation and payment of the sums due under the 1980 contract were the concluding acts of management by L&G and that so long as L&G withheld the formula, this was an area which called for the attention of the ombudsman in order to protect the interests of members of the CCA. I should however mention that (under pressure from me), at a very late stage L&G agreed to provide the ombudsman and CCA in confidence with the formula and the ombudsman agreed to receive this information in confidence and only to pass it on to CCA in confidence. This action on the part of L&G should enable any doubts or anxieties about the correctness of the calculation by L&G (albeit belatedly) to be resolved without delay.”

A number of points are worth making about that paragraph. First, Lightman J was not saying that calculation of the COD was an act of administration. He only expressed a view that it might be and that it was therefore not appropriate to prohibit the Pensions Ombudsman from investigating this matter. Second, in coming to this conclusion and expressing his views that there was much to be said for holding that this was an act of administration, he appears to have been under the impression that the calculation had been done by L&G in March 1992. Third he records, accurately, that L&G agreed to hand over the “formula”, that is to say the relevant MVAFs, under pressure. They did not accept then, nor do they now, that there was any obligation to do so, at least not to the Pensions Ombudsman.

Following Lightman J’s judgment, L&G disclosed the relevant MVAFs to CCA’s lawyers. There is no doubt that those MVAFs, when applied to the annual cash pools, produced the COD which L&G had paid on cessation of the CCA AF80 contract. However, CCA’s solicitors complained that they had not been supplied with the underlying material which would enable them to determine whether the MVAFs themselves were “accurate”.

It appears that each month the process of assessment of factors which have to be taken into account to generate the MVAFs culminates in the generation of a two page manuscript document. This refers to the various factors and includes mathematical calculations which lead to the MVAFs.

L&G objected to the disclosure of any more material. It said, and repeats on this appeal, that the disclosure of earlier documents which explain how it arrives at the judgment which is reflected in the MVAFs would be highly damaging to its business. In essence it would allow outsiders to understand how it balances various factors and would allow independent financial advisers (“IFAs”) to select against it. It says that this is not an unjustified fear. Similar information relating to one of its competitors was released and resulted in significant financial damage being inflicted on it by IFAs selecting against it. Not only did L&G object to going any further back than the MVAFs but on 3 October 2000 it also reported to the Pensions Ombudsman that it no longer had the two page manuscript document which immediately preceded the April 1992 MVAFs. It reported that there was a strong likelihood that the file containing this document had been destroyed.

The Pensions Ombudsman responded on 5 October 2000. He said, amongst other things and before hearing any arguments from L&G, that the loss of this document constituted maladministration.

On 24 April 2002, the Pensions Ombudsman sent the solicitors to L&G and CCA copies of his Preliminary Conclusions – effectively a draft of the Determination. Besides containing intimations that he intended to make certain findings of maladministration, the draft contained a direction in the following terms:

“89.

L&G shall appoint an independent actuary by agreement with CCA or, failing agreement, within 21 days of the date of this Determination shall request the President for the time being of the Institute of Actuaries to nominate an independent actuary to identify the MVAFs likely to have been used in April 1992 bearing in mind the figures used in months for which calculations are known. Any decision of the independent actuary shall be final and binding on all concerned and shall be given by him or her as an expert and not as an arbitrator. Any sum payable by L&G to CCA (on behalf of the Trustees), following the independent actuary’s determination, shall be paid by L&G within 14 days of the independent actuary’s determination, together with simple interest at the base rate quoted for the time being by the reference banks from the date that the discontinuance sum first became payable under the AF80 Policy. L&G shall provide all information within its possession or control to the independent actuary as he or she may request in order to reach his or her decision of the calculations. The costs of the independent actuary shall be borne by L&G.”

The receipt of the Preliminary Conclusions resulted in a detailed repost on behalf of L&G. Among other things, it argued that the making of a direction like that set out above was ultra vires. Perhaps of greater significance is the response from Messrs Simmons & Simmons, CCA’s solicitors. They wrote raising four objections to the proposed direction. First they said that to order the parties to agree anything was, on the basis of authority, an invalid direction. Second they pointed out that the direction appeared to be directed primarily at the President of the Institute of Actuaries, who was not a person within the classes of person on whom the Ombudsman’s determinations are binding. Third they said that the direction could be categorised as an abdication of the Ombudsman’s power to make determinations. He could not refer a determination to another person. Fourth, they pointed out that a court could not make an order that the parties to a dispute brought before it should be bound by an expert’s determination. They referred to Hillsdown Holdings plc v Pensions Ombudsman & others [1997] 1 All E R 862. The Ombudsman could not make an order which the court would not make. On the basis of that, they said that the Pensions Ombudsman had no power to make the direction. In other words, they agreed with L&G that the proposed direction was ultra vires.

On 9 December 2002 the Pensions Ombudsman issued the Determination. He held that three of CCA’s four complaints were not made out. However he made a number of findings which are the subject of this appeal. First he held that L&G had failed to provide “full details” of how the surrender value of the CCA AF80 policy was calculated. This, he said, was breach of contract or maladministration. Second, he said that the loss of the two page document which sets out how the April 1992 MVAFs were calculated is maladministration. Third he made a direction (“the Direction”) in the following terms:

“103.

L&G shall within 21 days of the date of this Determination request the President for the time being of the Institute of Actuaries to nominate an independent actuary (whom L&G shall immediately instruct) to produce, calculate and identify the basis of (ie breakdown) and resulting MVAFs likely to have been used in April 1992 bearing in mind the figures for which calculations are known and adopting where necessary other actuarial guidance, principles and assumptions that are within the range of reasonable responses. Any decision of the independent actuary shall be given by him or her as an expert and not as an arbitrator. L&G shall provide all information within its possession or control to the independent actuary as he or she may request in order to reach his or her decision of the calculations. The costs of the independent actuary shall be borne by L&G.”

It will be seen that, save in respect of Simmons & Simmons’ first objection, the Pensions Ombudsman has stuck to his guns. In particular, he ignored the suggestion from both sides that a direction of the sort made is ultra vires. This is not an accident. At paragraph 85 of the Determination, the Pensions Ombudsman expressly addresses the issue of whether he had powers greater than the court. He said:

“Maladministration is not something for which a court provides a remedy. The fact that the court may not be able to direct the steps I am directing should be taken to remedy the maladministration does not in my view mean, as L&G assert, that I cannot make such a direction.”

It will be appreciated that these three points are interlinked. It is because the Pensions Ombudsman decided that L&G was obliged, in performing its duties to administer the Scheme, to disclose all the workings which led to the calculation of the surrender value of the policy that failure to do that and the failure to disclose the documents on which some of the workings of the MVAFs were recorded was maladministration. The purpose of the direction is presumably to recreate the workings which led to the April MVAFs and to confirm that they were “correct”.

L&G’s appeal from the Determination came before Etherton J on 11 June of this year. Mr Inglis-Jones and Mr Hitchcock appeared on behalf of L&G. CCA were not represented, nor was the Pensions Ombudsman. It was apparent to the judge from a reading of the appellant’s skeleton argument that the issues raised on this appeal, including, but not limited to, the question of the scope of the Pensions Ombudsman’s powers, were of considerable general importance. Indeed he could hardly have been left in any doubt because he had before him a letter dated 6 February 2003 which the Pensions Ombudsman had written to the Court in which he stated:

“The purpose of this letter is to clarify an issue in relation to this appeal. In doing so the Pensions Ombudsman does not seek to persuade the Court or to act as an advocate in his own cause, but merely to explain a jurisdictional issue that is raised by the appeal. It is not the Pensions Ombudsman’s intention to participate in this appeal, unless the Court considers it appropriate that he do so, and subject to an assurance on costs.

At page 12, paragraphs 10(b) and 10(d) of the Notice of Appeal it is suggested that I cannot make the direction made at paragraph 103 of the Determination because it is not an Order that a Court could have made. … It is my view that my jurisdiction is wider than the courts in that I can investigate instances of maladministration, which is not a cause of action before the courts and that accordingly my power to direct should not be so narrowly construed. The power to direct, conferred on me by statute, provides (at section 151(2) of the Pension Schemes Act 1993): ‘Where the Pensions Ombudsman makes a determination under this Part or under any corresponding legislation having effect in Northern Ireland, he may direct any person responsible for the management of the Scheme to which the complaint or reference relates to take, or refrain from taking, such steps as he may specify’. This section is clearly widely drawn and there is nothing within it constraining it in any way to the limits of the courts powers.”

Etherton J was not happy that these important points should be considered without argument from the other side. Since CCA’s lawyers appeared to agree with L&G that the Pensions Ombudsman was acting ultra vires in making the Direction, in practice the only person who could be expected to offer arguments in support of the Direction was the Pensions Ombudsman himself. In the result, and in the face of opposition from L&G who were concerned at the growing costs involved in this case, he adjourned the hearing of the appeal. He gave a short judgment explaining his reason for doing so. He explained that all the points raised on the appeal appeared to have significance beyond the facts of this case and, in relation to the question of vires, he said:

“32.

The question of the power of the Pensions Ombudsman to make the kind of direction which he did, in para. 103 of the Determination, is of significance well beyond the specific facts of this case. The Pensions Ombudsman himself plainly thinks that the point is not only arguable in his favour but is of wide significance, since that is the reason that he wrote the letter to the Court of 6th February 2003. Further, it seems to me that, if Mr. Inglis Jones is right, and that it is and was perfectly plain that the Pensions Ombudsman had no power to make the order in para. 103, so that in effect his decision to do so was demonstrably perverse, then that is a further reason why the Pensions Ombudsman should explain to the court, by way of participation in the appeal itself, why it is that he considers that the direction was properly given, and that he has the powers which he sets out in his letter of 6th February 2003.”

Notwithstanding L&G’s concerns, he decided to request the Pensions Ombudsman to participate in the appeal in order to provide a balanced set of arguments before the court for the determination of the issues which it raises. He acknowledged that he could not compel the Pensions Ombudsman to attend but he noted that the Pensions Ombudsman was a public officer and that there was a public interest in resolving the points raised on the appeal. He said:

“My expectation is that, as such a public officer, the Pensions Ombudsman would respond affirmatively to any such request of the court.”

He noted that it was neither possible nor appropriate for him to give any assurance as to the Pensions Ombudsman’s costs.

On 23 June, a copy of Etherton J’s judgment was sent to the Pensions Ombudsman under cover of a letter from the judge’s clerk which included the following:

“I enclose a copy of an approved transcript of the judgment of Mr Justice Etherton which he gave on that day. As you see from the transcript, the Judge adjourned the hearing of the appeal to enable the Court to request you to participate in the appeal so that there can be balanced argument on the points raised on the appeal. The significance of those points is not, or is unlikely to be, restricted to the facts of this particular case. It is, therefore, in the public interest that the Court should hear both sides of the argument on the appeal, including the important issue of your jurisdiction which you address in your letter to the Court of 6 February 2003.

I have, therefore, been asked by Mr Justice Etherton to communicate to you his formal request that you participate in the hearing of the appeal. It is appreciated that your letter to the Court of 6 February 2003 said that you would be willing to do so subject to an assurance on costs, but, as I am sure you understand, it would be wholly inappropriate for the Court to make any order as to your costs at this stage. All issues of costs can only properly be dealt with by the Judge at the conclusion of the appeal.

In view of the considerable time CCA’s complaint has been under consideration, it is plainly important that the adjourned hearing of the appeal be heard as speedily as practicable. The Judge therefore hopes that you will be able to give a positive indication to the Court and to L&G within 21 days that you will participate in the appeal.”

Unfortunately Etherton J’s hopes were not realised. By letter dated 14 July, the Pensions Ombudsman responded. He said that he understood Etherton J’s views that some of the issues raised in the appeal are of general importance. On the other hand he had to bear in mind the “risk to public funds” which would be involved in his attending. He suggested the appointment of an amicus curiae (presumably still out of public funds, but not from his budget) or a preliminary hearing limited to some matters. Absent that, or an assurance as to costs, he declined to attend.

Since Etherton J’s view was that all the points raised on the appeal were likely to be of significance beyond the facts of this case, the suggestion of a preliminary hearing limited to some of them only rather misses the point. In addition it is a suggestion which, if acceded to, was likely to prolong and increase the cost of these proceedings. Furthermore, it is unfortunate that the Pensions Ombudsman was only prepared to attend the whole of the hearing on condition that L&G agreed in advance to pay his costs. If, as L&G and Simmons & Simmons had said, the Direction was ultra vires, this appeal will have been caused in large part by the Pensions Ombudsman having pursued a course which both parties warned him against. Why, in those circumstances, it would have been equitable for L&G to pay his costs in advance is not clear to me. The Pensions Ombudsman has offered no reason.

In any event, the Pensions Ombudsman has chosen not to attend on this appeal. The result is that an additional burden has fallen on Mr Inglis-Jones to present his client’s case in as balanced a manner as possible.

It is self evident that L&G can only be guilty of maladministration in relation to acts of administration. As Lightman J said in Legal & General Assurance Society Ltd v Pensions Ombudsman (see paragraph 28 above), in this case the remit of the Pensions Ombudsman is to investigate maladministration by L&G “acting in the capacity of a manager”. The relationship between L&G and the Scheme has more than one facet. In part it is an arm’s length commercial one with L&G supplying its AF80 product to the trustees. In other areas L&G has undertaken the task of managing the Scheme. It is necessary to ensure that the activities or inactivities complained of fall within the management function of the party being investigated. The difference between the commercial and managerial capacities in which L&G acted is inherent in Lightman J’s decision that the Pensions Ombudsman could not consider the merits or fairness of CCA AF80 because neither “the entry into the 1980 contract nor its continuance in force constitute an act of management.”

The scope of the management duties will be defined by the nature of the Scheme and what functions the manager has been asked to and has agreed to undertake. For so long as he continues in post, the way in which he discharges his duties is subject to review by the Pensions Ombudsman. But the latter’s remit does not cover activities which are not acts of administration, either because they are not the types of activity which constitute administration or, even if they are capable of being so, are being carried out by a party who has no managerial function, either because he has not been appointed a manager or has ceased to act in that capacity.

The first finding of maladministration made by the Pensions Ombudsman relates to the calculation of the MVAFs. First he concluded that the act of calculating the surrender value of the policy was an act of management:

“65.

That L&G was contractually bound to calculate and pay the discontinuance sum (whether as part of the AF80 Policy or otherwise) did not alter its status as manager. It contractually undertook to provide a package of services comprising the management of the Scheme. The termination of the policy, calculation of the discontinuance sum payable and its payment were facets of that package.”

The Pensions Ombudsman accepted that the MVAFs disclosed by L&G pursuant to Lightman J’s suggestion were the ones used to calculate the surrender value (see Determination para 68). However he went on to hold:

“71.

Simple disclosure of MVAFs as percentages does not enable the checking of the correctness of the calculation of the discontinuance sum. To my mind what is required is disclosure of the entire process of the calculation of the difference between the notional value and the discontinuance sum - this being the essence of the basis currently in use/the Formula. It is not disputed between the parties that the MVAFs are a key component responsible for the difference between the notional value and the discontinuance sum. It seems therefore that the process/basis of any significant component of the calculation should be disclosed. The basis currently in use is defined by L&G as the MVAFs. It is that basis which requires verification of accuracy when applied to calculate the discontinuance sum.”

Mr Inglis-Jones attacks the conclusions in both of these paragraphs. It is convenient to consider the second one first.

The initial step is to understand the extent of the obligation of disclosure on L&G if the Pensions Ombudsman is correct. As will be seen from paragraph 71 of the Determination set out above, what was required was disclosure of the “entire process” of calculation of the difference between the notional value and the surrender value. As I have explained, the setting of MVAFs is an iterative process. The April 1992 MVAFs were set by modifying the March 1992 MVAFs and so on back to the point at which new MVAFs were calculated directly from the valuation of underlying assets. It seems to me that if L&G are obliged to disclose all that is necessary to verify the April MVAFs, this must include disclosure of the major ingredient, namely the March MVAFs, as well as material relating to all the factors taken into consideration in moving from the latter to the former. But no verification would be possible unless the accuracy of the March MVAFs could also be investigated and so on backwards. I think one is driven to the conclusion that the disclosure obligation imposed by the Pensions Ombudsman could involve the disclosure of material generated over a period of months or maybe longer, including material generated before Mr Scholes had taken over CCA and when, presumably, there was no interest in terminating CCA AF80. Certainly the independent actuary who the Pensions Ombudsman wants appointed may take that view.

Mr Inglis-Jones argued that it is possible, but unclear, that the Pensions Ombudsman construed CCA AF80 as including a contractual obligation on L&G to disclose all this material. I do not think he did so. He did not point to any clauses in the contract which could be construed as creating such an obligation, nor did he point to any factors which would have justified implying any such term. As Mr Inglis-Jones pointed out, the implication of any such term might cause grave difficulties to L&G. If there were an implied contractual obligation to disclose all the supporting material used to set the MVAFs, logically there ought to be equivalent implied obligations to disclose all the workings behind L&G’s setting of the premiums to be received and payments to be made under the contract. Such terms would require L&G to disclose a large amount of internal and highly confidential documentation. There is no basis upon which it could be said that such terms represent the presumed intention of the parties, and certainly not the presumed intent of L&G, or would be necessary to give business efficacy to CCA AF80. In any event, as Mr Inglis-Jones argued, whether or not there was a contractual obligation on L&G to give this sort of disclosure is more or less irrelevant. What matters is whether, whatever the contractual terms, what L&G did or did not disclose was an act of maladministration of the Scheme. This is the core question.

Once the fund-wide MVAFs are set, they are then available to be applied to the individual Annual Cash Pools within each with-profits contract, including all the AF80 contracts operated by L&G. In the case of CCA AF80, in order to determine the surrender value, it was necessary to carry out a simple mathematical process which involved multiplying the relevant Annual Cash Pools by the relevant MVAFs. The figures so produced were specific to CCA AF80. Lightman J, without determining the point, was prepared to contemplate that the performance of this calculation was an act of administration of the Scheme and could therefore be investigated by the Pensions Ombudsman. It is apparent from his submissions that Mr Inglis-Jones did not accept that this preliminary view of Lightman J was correct.

However Mr Inglis-Jones argued that, whatever the status of that calculation and in whatever capacity L&G performed it, it is apparent that working out the individual MVAFs is not an act of administration of CCA AF80, any more than a bank’s calculation of its base rate will be an act of administration of the individual accounts at the bank, including – but not limited to – those which, because overdrawn, will be charged interest calculated by reference to that base rate.

No doubt, when the MVAFs are set, L&G does not have in mind the individual policies and funds within its AF80 business. For many years L&G assessed MVAFs while it was also administering CCA AF80. During those years there was no suggestion of termination of the contract. So long as that continued to be the case, the MVAFs were not used to calculate the surrender value of CCA AF80, yet the MVAFs continued to be assessed none the less. This is because the assessment of the MVAFs is not part of the administration of individual AF80 contracts, but is an act of administration of the whole of the Long Term Fund. In much the same way, decisions of what assets to buy and sell within that Fund and the prices to be paid all affect the performance of that Fund and are part of its management. Those decisions also have an impact on the performance of the individual policies within the Fund. But that does not mean that in making those decisions L&G is acting in its capacity as manager of the Scheme in which the policy is employed.

In my view, these submissions are correct. The process by which L&G takes into account factors which have a bearing on the current value of the Long Term Fund, so as to arrive at MVAFs, is not a function it performs in its capacity as administrator of CCA AF80. No doubt, the same process of assessment would be carried out by the same L&G employees and probably with the same result even if CCA AF80 did not exist.

It follows that L&G’s refusal to disclose its method of setting MVAFs is not within the remit of the Pensions Ombudsman. Even if done incompetently or in an arbitrary way, it does not amount to maladministration of CCA AF80. Mr Inglis-Jones pointed out that this does not mean that L&G or its competitors have a free hand to act incompetently or worse while running their respective Long Term Funds. It is the identity of the supervisor which changes. In the case of pension schemes, supervision is in the hands of the Pensions Ombudsman. When it comes to the operation of investment business like that involving L&G’s Long Term Fund, it is the Financial Services Authority which supervises. Each has its own expertise which is best suited to its supervisory role. The Pensions Ombudsman probably does not have the knowledge and expertise to assess whether L&G was running its Long Term Fund properly with regard to the interests of its customers and shareholders and, in particular, whether it took into account all the relevant factors and gave them proper weight in deciding on each MVAF. It is, no doubt, for that reason that the Pensions Ombudsman here gave the Direction he did which, in effect, purports to delegate to an independent actuary the task of determining whether the MVAFs used by L&G were appropriate.

It follows that I accept Mr Inglis-Jones’s submission that the Pensions Ombudsman had no jurisdiction to investigate how L&G set the MVAFs.

In the light of this it is not strictly necessary to deal with the issue raised in paragraph 65 of the Determination set out above, namely the finding that “termination of the policy, calculation of the discontinuance sum payable and its payment” were facets of the administration package which L&G had undertaken. However, because of the impact this case may have on others, I shall deal with this.

When L&G set the MVAFs which were subsequently used to calculate the COD, it was during the time frame within which it was still exercising its functions as an administrator of the Scheme. However, the calculation of the surrender value, that is to say the multiplication of the relevant Annual Cash Pools by the appropriate MVAFs took place about a year after L&G had been told, in non-negotiable terms, that it was no longer the administrator of the Scheme and that its functions had been taken over by PJC. The fact that this calculation took place so much later lends support to Mr Inglis-Jones’s suggestion that it cannot be properly categorised as an act of administration. First, the passage of time emphasises that the calculation of the surrender value is quite distinct from the setting of the MVAFs. They are different activities carried out at different times. Even where the act of calculating the surrender value is an act of administration of the Scheme, it has nothing to do with setting the MVAFs. It does not bring the latter activity within the remit of the Pensions Ombudsman.

Furthermore, I do not agree with the Pensions Ombudsman’s finding that the termination of the policy, calculation of the discontinuance sum payable and its payment fell within L&G’s functions as administrator of the Scheme. If, as Lightman J held, neither the entry into the CCA AF80 nor its continuance in force constitutes an act of management (see paragraph 28 above), I cannot see how termination of the contract can be. If termination of the contract is not an act of administration, it is also difficult to see how calculation of the sums due on termination can be an act of administration of the Scheme either. It is true that Lightman J was prepared to contemplate that the performance of the calculation was an act of administration which could be investigated by the Pensions Ombudsman, but he was not called uponto decide, and did not decide, the point. L&G’s consent to his suggestion to disclose the April 1992 MVAFs made it unnecessary to do so. In coming to that tentative view, it appears that he was not aware just how long after L&G’s dismissal as the administrator the calculation took place.

In my view, everything points to the carrying out of the calculations being a purely commercial transaction by L&G and not an act of management of the Scheme. Indeed, it seems that that must have been CCA's view as well. It will be recalled (see paragraph 24 above) that in February 1992, Mr Scholes of CCA wrote a letter to L&G which said that, with immediate effect, PJC had been appointed as the administrators of the Scheme with authority to negotiate with L&G as to the discontinuance terms under CCA AF80. The covering letter from PJC (see paragraph 25 above) also confirmed that it was to take over administration with immediate effect. Consistent with that, during the period that the discontinuance terms were being negotiated and surrender value was being calculated, it was PJC which was the administrator of the Scheme, not L&G. Although there was a discussion in correspondence as to when L&G ceased to be administrators, it appears that any doubts had been resolved by August 1992, well before the surrender value was calculated. The Pensions Ombudsman expressed the opinion (Determination para 62 and 63) that discontinuance of the CCA AF80 was part and parcel of the management of the Scheme. I can see no basis for that.

To the extent that the Pensions Ombudsman relies upon the fact that the MVAFs were used to calculate the surrender value as a justification for saying that it was maladministration of the Scheme not to disclose the documents relating to the manner in which those MVAFs were set, he was relying on an activity which itself was not an act of administration.

For the above reasons, I have come to the conclusion that the Pensions Ombudsman was in error in this finding of maladministration.

The second finding of maladministration concerns L&G’s inability to locate the two page manuscript document on which the factors used to set the April 1992 MVAFs on the basis of the March MVAFs are set out. The finding is expressed as follows in the Determination:

“My conclusion is that the failure to provide full details of the Formula is maladministration as, indeed is the subsequent loss of the calculations. That loss prevents an easy way of now putting the matter right but I make a direction designed to overcome the difficulty. I wish so far as I can to enable CCA to check the correctness of the calculation.”

It must follow from my conclusion in relation to the first head of alleged maladministration that the production and retention of this document, just as much as the process in the course of which it was produced, is not an act of administration of the Scheme. Even if misplacing this document was incompetent or worse it is not within the Pensions Ombudsman’s remit.

It seems to me that there are additional reasons why the finding of maladministration was not justified. As I think Mr Inglis-Jones accepted, proper administration of a pension scheme carries with it a duty to make and retain relevant records. If setting the MVAFs had fallen within the scope of L&G’s duties as administrator of the Scheme, retention of this document might have been a consequential obligation. However an obligation to retain records is not open ended. It must only be to retain them for a reasonable number of years. In this case, the first time it was suggested that this document was relevant or should have been retained was over seven years after L&G had been told it was no longer administrator of the Scheme. The Pensions Ombudsman does not address the issue of how long the need to retain lasts. Mr Inglis-Jones complained that this finding was made without any investigation having been undertaken, or questions having been asked, by the Pensions Ombudsman as to when, or in what circumstances, these calculations had no longer been retained, whether L&G had, or should have had, a standard system for retaining records of such internal workings and, if so, what that system should have been and whether it had been adhered to in this case. He also pointed out that the finding was unsupported by any reasons. These are substantial objections. I am not persuaded that there is a general obligation on a former administrator to keep all documents for over seven years after ceasing to act in that capacity. The finding of maladministration is unjustified.

This takes me to the Direction. A major part of Mr Inglis-Jones’s submission was that, whether there had been maladministration or not, the Direction made was inappropriate and that it is important, not just for this case, to clarify what directions the Pensions Ombudsman may make.

There is no dispute that the Direction is not one which the court would make but, as explained above, the Pensions Ombudsman does not suggest otherwise. His argument, as expressed in his letter of 6 February 2003 is that the Act give him wider powers to make directions than the court has and that the language of the Pensions Schemes Act 1993 is unlimited in scope and his powers are unlimited also.

The Pensions Ombudsman’s position is somewhat reminiscent of one of the arguments advanced in Gouriet v Union of Post Office Workers in relation to the court’s power to grant interlocutory injunctions. S 45(1) of the Supreme Court of Judicature (Consolidation) Act 1925 gave the court the power to grant such injunctions “in all cases in which it appears to the Court to be just and convenient so to do.” Based on this, the majority of the Court of Appeal held that Mr Gouriet could seek interlocutory relief to prevent the Union from carrying out its threat to boycott all mail to South Africa because of the latter country’s then apartheid policies. They were prepared to countenance such relief because they thought it was just and convenient to do so notwithstanding the fact that Mr Gouriet had no cause of action against the Union and, for that reason, would not be able to obtain final injunctive relief. The House of Lords unanimously held ([1978] AC 435) that the wide wording of the statute did not mean that the courts had, in effect, limitless powers to grant interlocutory injunctions whenever they thought it convenient to do so. So here, the fact that the legislation is in very wide terms does not necessarily mean that the Pensions Ombudsman can make whatever directions he likes.

The scope of the Pensions Ombudsman’s powers has been considered in a number of cases. A convenient starting point is Hillsdown Holdings plc v Pensions Ombudsman [1997] 1 All ER 862 in which Knox J had to determine whether the Pensions Ombudsman could make orders which the court could not. This is the authority to which Simmons & Simmons referred. He started by pointing out that there was some difference of judicial opinion on this issue at first instance as reflected in Miller v Stapleton [1996] 2 All ER 449, Westminster City Council v Haywood [1996] 2 All ER 467 and Wild v Pensions Ombudsman (1996) Times 17 April. The judge came to a firm conclusion that the powers did not extend beyond those of the court. His reasoning was expressed as follows:

“My main reasons for holding the view that I have expressed regarding the Pensions Ombudsman’s inability to direct an employer or trustee to replace funds improperly removed from a pension scheme unless the court itself could so order are twofold. First, it seems to me that there is a real distinction between ordering compensation for inconvenience and distress caused by maladministration as an adjunct to the power to remedy injustice caused by maladministration which, in line with Robert Walker J’s decision, I take to be permissible, and requiring the repayment of what might well be, and in this case were, very substantial sums by way of payment out from a pension fund, on the other hand. It is trite law that pension funds must operate within the law and it does not seem to me right that there should be a different answer to the question ‘are you legally liable to repay this sum’ according to the tribunal to which resort is had so that the answer is: ‘If I am sued in court, No, but if a complaint is made to the Pensions Ombudsman, Yes.’ The injustice through maladministration must in this case consist of the detriment suffered by the payment out itself and is in no sense ancillary as are claims to compensation for inconvenience and distress. My second reason is tied up with the first and is that s 146(6)(a) of the 1993 Act prevents the Pensions Ombudsman from investigating a complaint if before the complaint is made proceedings have been begun in court in respect of the matters which would be the subject of the investigation. That suggests that the two are intended to be mutually exclusive alternatives and it would be strange if it was contemplated that the alternatives would or might produce different results as to the substance of the dispute. I can well imagine that the two tribunals would be contemplated as having radically different procedures and it may be types of relief but I would not expect differences on such fundamental matters as whether there was a liability to repay capital sums. Also there would be a possibility of abuse if it were possible to avoid an impending complaint to the Pensions Ombudsman by a well-timed application for the determination of a dispute of fact or law.” (p 899)

This decision, and Knox J’s reasoning, was followed by Sir Richard Scott V-C in Edge v Pensions Ombudsman [1998] Ch 512. He said:.

“I respectfully agree with this approach. In a case in which the maladministration complained of consists of an alleged breach of trust, the Pensions Ombudsman has no power, in my judgment, to direct remedial steps to be taken that are not steps that a court of law could properly have directed to be taken. … In these circumstances, and having regard to the respective positions of the employee members and the employers, a court could not, in my judgment, have ordered the deed to be set aside. A court could not have directed the trustees to take steps that could only be justified on the footing that the deed had been set aside. Nor, in my judgment, could the Pensions Ombudsman do so.” (p 520)

Although that case was concerned with breach of trust, there is nothing to suggest that the limitation on the Pensions Ombudsman’s powers was limited to such cases. The reasons for limiting the powers so that they do not exceed those of the courts are set out in Hillsdown. They apply generally.

The same issue also arose for consideration in Wakelin v Read [2000] PLR 319. It is dealt with in the greatest detail in the judgment of Mummery LJ. He said:

“73.

A point has been raised by the Trustees on the width of the Ombudsman’s discretion to give directions under s 151(2) of the 1993 Act. Mr Simmonds QC sought on behalf of the Trustees to uphold the decision of the Ombudsman to decline to direct the Trustees to pay the benefits on the ground that Mr Read did not come with clean hands. He submitted that the sub-section confers the widest possible discretion both as to the nature of the remedy granted and as to whether he should grant any remedy at all. He was not limited to those courses of action which would be open to a court in litigation.

I am unable to accept this submission. …”

Then, having quoted with approval from Knox J’s judgment in Hillsdown he continued:

“75.

[Knox J] added that the limitations on [the Pensions Ombudsman’s] power are to be found in the definition of his function, in particular the determination of disputes of fact or law. At p 899d he repeated that it was trite law that ‘pension funds must operate within the law’ with the result that, on a question of legal liability to refund payments improperly made from a scheme, it was not right that there should be a different answer as to the substance of the dispute according to whether the dispute was decided by a court or by the Ombudsman.”

Although in these cases the Pensions Ombudsman purported to exercise his powers in a way which differed from his exercise of powers here, the underlying issue is the same. In my view, these authorities, and in particular the analysis in Hillsdown, are compelling. The Pensions Ombudsman cannot make an order which the court could not make. Since the Direction could not have been made by a court, he should not have made it either.

In my view this is not just an arid question of the width of the powers given by statute. There are good reasons why a court would not have made an order in terms similar to those in the Direction. Those reasons are just as valid for the Pensions Ombudsman.

First, it is unclear what the effect of the Direction is. What happens, for example, if L&G comply with the Direction but the President for the time being of the Institute of Actuaries (“the President”) refuses to appoint? When this point was raised with the Pensions Ombudsman, he said that this was a theoretical rather than a practical concern. It seems to me that this response misses the point. Once L&G has complied with the Direction, it has done what was required of it. It is unclear on what basis the Pensions Ombudsman could take any further step. Further, even if the President accepts the invitation to appoint an actuary, it is unclear who, if anyone, has any control over him. For example, what happens if the appointee is incompetent or unreasonably slow? Since he was not appointed by the Pensions Ombudsman and is not under his control, it is unclear what steps could be taken to encourage him to perform better. It is also unclear how the Pensions Ombudsman could remove him.

The central purpose of the Direction appears to be to delegate the Ombudsman’s investigating power to a third party, an actuary, appointed by someone else, namely the President. However the Ombudsman has no power to delegate like this and, even if he did, it is difficult to see how this type of delegation could be justified. It may well be that behind this lies a recognition that the Pensions Ombudsman does not have the necessary expertise to decide whether the MVAFs had been set fairly. As I have mentioned above, this is consistent with the exercise of setting MVAFs being one which does not relate to the administration of the Scheme (where the Pensions Ombudsman does have expertise), but whether that is so or not, the obligation to make relevant findings of fact rests with the Pensions Ombudsman and it cannot be proper for him to pass that task to someone else, particularly someone who is not chosen by him and over whom, apparently, there would be no control once appointed.

The handing over of the task of fact finding to the independent actuary results in strange consequences. L&G is required to “provide all information within its possession or control to the independent actuary as he or she may request in order to reach his or her decision of the calculations.” Presumably the actuary can decide, without reference to the Pensions Ombudsman, what information he wants. He can ask for as much or as little as he wants. It is not clear what would happen if, for example, L&G refused to hand over documents because it was of the reasonable view that they were neither necessary to the actuary’s task and were highly confidential. Who resolves the dispute? What enforcement is possible?

In addition to these points, I can see no justification for deciding in advance that the costs of the independent actuary are to be borne by L&G. What happens if the actuary takes far too long and wastes time on pursuing irrelevant lines of inquiry? What happens if the actuary comes to the conclusion that not only were L&G’s MVAFs reasonable, they were actually more generous than they need be? Why should L&G have to pay for such an exercise? It appears to me that these are the sort of considerations which are likely to have caused Etherton J to say, through his clerk, that it would be wholly inappropriate for the Court to make any order as to the Pensions Ombudsman’s costs in advance and that the issue of costs could only properly be dealt with at the conclusion of the appeal.

In addition to these criticisms, the purpose of the Direction is unclear. As written, the independent actuary’s task is to calculate the “MVAFs likely to have been used in April 1992.” This cannot be the true intention because L&G have disclosed the MVAFs which were actually used. It is to be assumed, therefore, that this is really an exercise designed to test whether the MVAFs used by L&G were reasonable in all the circumstances. This may account for why the terms of reference are so widely drawn, namely to bear in mind “the figures for which calculations are known and adopting where necessary other actuarial guidance, principles and assumptions that are within the range of reasonable responses.” However the Direction leaves L&G in the air as to what will be the consequence of the actuary’s determination. It is to be assumed that if the actuary produces figures substantially the same as L&G’s, that will be an end of the matter. But what if he produces MVAFs which would have been much more generous to those with an interest in the Scheme? This means that they got a bad deal, or a worse deal than the actuary believes was appropriate. It is difficult to see what further steps could be taken. As Lightman J said in L&G, the question of the perceived fairness or unfairness of the exercise by L&G of contractual rights granted by the 1980 contract to L&G for its own benefit do not fall within the remit of the ombudsman. It seems to me that the outcome of the actuary’s exercise has been acknowledged by the Pensions Ombudsman to have no relevance to the complaints he was investigating. In a letter to Simmons & Simmons he explained:

“To the extent that the actuary’s figures are materially different from the existing figures the Ombudsman considers that this is not relevant to the complaint before him but could be the subject matter of a separate complaint – normal considerations applying. ”

In the light of the above conclusions, it is not necessary to consider Mr Inglis-Jones’ additional argument that CCA was not in a position to complain of any supposed failure of L&G to honour its obligations under CCA AF80 since it was not a party to that contract.

For the reasons set out above, I will allow this appeal.