Dear Trustee,

I am going to call you a trustee although I know there is a technical difference between a trustee and a director of the trust - the procedures allow me some informality, I believe.

I appreciate that there is something Quixotic in asking the trustees to conclude that they and previous trustees have not directed themselves properly. However, the Internal Disputes Resolution Process requires it, to process my complaints.

I have two complaints that relate to the Trust. The response to the Stage 1 IDRP by your appointed first-stage decision-maker (FSDM) said that he was confused about one of them. The other he did not address. The OPRA guide notes for Stage 1 call for an investigation: "The person who investigates a complaint at the first stage..." and say that "you [the trustees via the FSDM] should explain clearly what details you are going to need to be able to investigate the matter fully". I heard nothing from the FSDM in two months, not even an acknowledgement of receipt of my complaint.

There was so much wrong with Stage 1 that I have difficulty in describing it all. The approach I have taken has several parts. I have provided some preamble. These are probably things the FSDM knew but the Stage 1 output does not make it possible to tell whether they were context to his decision or not. They should have been.

Note - where I say something along the lines of "the FSDM should have considered", I intend it to be taken as both something I think you, the trustees, should consider in Stage 2 and something that should have been considered in Stage 1. The response to Stage 1 makes it impossible to tell what was considered in Stage 1 with respect to my complaints, or if anything was. Note that this is emphasis. Everything mentioned in my complaint deserves consideration, even if it is not emphasised.

I have retained my complaints in the same words as in Stage 1. I could find nothing in the Stage 1 response to cause a change to these complaints.

I have changed from a prose style to an item-by-item style in describing the behaviour to be considered. I cannot be sure that is an improvement over the text that the FSDM could not follow, but it seems to be. Things I expected the FSDM to consider in his investigation, and which he should have considered, are made explicit.

I expected the FSDM to investigate the stresses leading to the behaviour. I have spelled out what he should have looked at.

I have retained a complaint against IBM in this text, as I had one in the original, although you are not expected to respond to it. As before, I am copying it to them for any comment they may see fit to make.

Yours faithfully,

Dr B L Marks

Here is a preamble on Limited Price Indexation to give some context to this complaint:

A 1989 report by the Occupational Pensions Board entitled "Protecting pensions: safeguarding benefits in a changing environment" recommended that:

"Employers and Trustees should be urged as a matter of good practice to guarantee, on pensions in payment other than Guaranteed Minimum Pensions, increases at least equal to price-indexing up to 5% a year."

Through the decade, companies have adopted LPI as good practice. Most members of occupational schemes are protected by a guarantee of LPI or a guarantee at least as good, eg a fixed rate of 5% or full Retail Price Indexation. According to the surveys made by the periodical "Occupational Pensions" the percentage of members with this protection was in the mid-nineties throughout the decade.

Of the members who were not protected by a guarantee, about half received increases of LPI or more. Hence some 98% of members received LPI or more, throughout the decade.

IBM Defined Benefit plan members have no guaranteed protection of the value of the majority of their pensions.

Again calculating with figures from "Occupational Pensions", the values (in 1990 pounds) for members in different surveyed schemes, having the same £10,000 pension at the end of 1990, are:

I do NOT complain about these numbers in themselves. They are context for the complaints that I do make.

Here is a preamble on the trustee system for managing pensions.

In everyday commerce it is never questioned that when a company hires and fires the majority of the directors of another company then it controls that company. However, when the subject directors are called trustees, and given some training in what that means, the control is nominally lost. When the directors are operating for the trustee company, they are required to make judgements in the interests of the members of the trust, and not be under the control of the company that hires and fires them.

This skill in preventing thoughts from one part of the mind inter-acting with thoughts from another part when making decisions is not a natural skill. Judges train for decades to do it and then not everybody is confident they do it. You will recall the judge in the Pinochet case who was deemed (after he had judged) to be unsuitable because of possible public doubts that he could distinguish his support for Amnesty International from his role in judging Pinochet.

However, trustees steeped in decades of single-minded profit making for the company, and who operate in this mode all the time outside of trustee meetings, are assumed to acquire this detachment skill when they become trustees.

Nobody argues, I hope, that the trustee system is infallible; it is just the best choice of system available for running a trust. It will fail the members when the influences for company profit overwhelm the influences benefiting the members to such an extent that trustees (consciously or unconsciously) make decisions that do not properly balance the various interests.

I do NOT complain about the trustee system. It can work well when influences are reasonably balanced and a British sense of "fair play" is brought to bear.

Here is a preamble about the quality of evidence in this complaint.

The quality is less than ideal. The FSDM chose not to talk to me. The trust is not required to tell me the most rudimentary facts about a meeting, such as whom attended. Even where the trust administration will provide answers, the time taken for me to get them is an order of magnitude longer than it would have taken the FSDM. I judged it not reasonable to pay for professional actuarial calculations. Back issues of periodicals produced for the pensions profession are not held by any public reference library. The data I can gather, e.g. on the UK Pensions scene, is limited by my sources and research time. Any legal bits I refer to are the result of sparse website browsing of what is available for free. Perhaps there is material in this complaint that is extraneous. Without the benefit of any exchanges with the FSDM, I thought inclusion was safest. I have not put sources for everything that is in here; it would be pedantic to do so. Where appropriate, if something is contradicted, I will supply the reference on request.

The table in the section on my personal involvement is flawed, as a result of a mistake that IBM and the Trust are jointly responsible for. I have left the flaw in the table in order to show the table on which I made my personal retirement decisions.

My complaints are serious. I hope you will see any limitations in this evidence as a reason for acquiring more understanding yourself, rather than as a reason to rush to judgement.

The Complaints

This is a complaint about the pattern of behaviour of IBM UK and the Trust over the period since 1991, especially the relation of that behaviour to the behaviour prior to 1991. The behaviour since 1991 shows that the Trust has set out solely to maximise benefits to the IBM balance sheet. This was a misdirection and I complain about that. IBM UK contributed by their choice of directors appointed to the Trust board and I complain about that. The trust administration contributed by their unwillingness to work with the elected members of the board and I complain about that.

These are the only three complaints.

What the FSDM should have found.

Insofar as I could, I have done the investigation that the FSDM should have done. I have found that there is a pattern of the Trust considering, and implementing, changes that would benefit IBM, with no balancing considering and implementing of changes that would benefit the members.

This pattern establishes that the trustees were not applying the proper principles. From what the Trust has said, it can be deduced that the wrong principle the trustees were applying was that what IBM wanted should be agreed to if the trustees considered IBM could conceivably do something, legally defensible, which would damage the members. This was a wrong principle to adopt, even though there were in practice no realistic threats from IBM.

The trust system was uniquely stressed by a combination of factors. The change of management culture by IBM US early in 1990s has produced powerful forces for short term profit maximising without any balancing forces through union recognition. There is no force outside the Trust acting for the benefit of retirees.

The trust system was stressed by the actual choice of IBM appointed trustees. The choice of people under extreme pressure to deliver IBM advantages made it excessively difficult for them to put aside those pressures in their trustee roles.

Some trustees wished the potential benefit of the reserves to the members to be considered. They were hampered by the trust administration's unwillingness in marshalling the data and arguments entailed in that. They were thwarted by the trust not allowing the subject to be discussed, acting on the principle of doing what IBM wanted, when IBM did not want some of the potential benefits to members to be considered.

In support of the statement:

" I have found that there is a pattern of the Trust considering, and implementing, changes that would benefit IBM, with no balancing considering and implementing of changes that would benefit the members."

An analogy may help to show why behaviour over a period is important. Suppose a footballer shoots at goal and the shot is saved. Do we conclude that he is not up to the job at that level of football? The answer is that it depends - if this was one miss in the middle of a series of goals scored there can be a different answer than if this miss is the latest in a line of misses. Similarly, what is at issue with misdirection is a series of decisions favouring the IBM balance sheet over the interests of members. My complaint will not have been considered until, at least, all the decisions referenced here have been considered and their collective significance reviewed in addition to their individual significance.

The decision not to review increases of pensions in payment on an annual basis.

The decision to allow Defined Benefit plan funds to be used to fund Defined Contribution plan liabilities.

The decision not to tell the Members about the Defined Contribution plan funding plan.

The decision not to discuss proposing to IBM a 100% RPI pension increase in 1999.

The decision to approve the pension aspects of IBM outsourcing contracts.

The decisions in the directions to the auditors about Statutory Surplus calculations.

Other decisions.

I do not claim that the Trust should never make a decision which is beneficial to IBM and detrimental to the members, any more than I claim that a successful striker must score with every shot. I do claim that the pattern is sufficiently marked as to be relevant to the justice of what has occurred.

I now repeat this list of decisions with material relating to each of them.

The decision not to review increases in pensions on an annual basis:

The trustees were, or should have been, aware that almost all UK occupational pensions are reviewed annually. The actuarial calculations solicited by the trustees were based on annual review ("Prefunded" annual increases). Review periods longer than a year disadvantage members because they lose out in the period after the elapsed year and before an increase is considered. Review periods longer than a year benefit IBM by storing up more reserves that can potentially be used to improve the IBM balance sheet. This decision favoured the IBM balance sheet and was detrimental to members.

There were dozens of trust meetings in the decade and annual reviews could have been introduced at any of them. The decision at any one of them represents the decisions at all of them, not to go to an annual basis.

The decision to allow Defined Benefit plan funds to be used to fund Defined Contribution plan liabilities.

IBM proposed a mechanism to transfer funds away from the Defined Benefit plan and benefit the IBM balance sheet by the same amount. Legality of the mechanism is not an issue in my complaint, except insofar as it would undoubtedly have been legal to not use the mechanism.

The decision to allow the mechanism to be used was to detrimental to members because, in comparison with the alternative of allowing the M-plan and not allowing the transfers of funds, it reduced the reserves of the Defined Benefit plan (or of the encompassing fund, if it is decided that there are not different funds). If the transfers were not allowed, IBM UK would have funded the Defined Contribution plan with a resulting increase in the funds.

Increased reserves are of benefit to members as security. Increased reserves are of benefit to members because they increase the prospects of the fund becoming over-funded and the paragraph 2(3) of schedule 22 to the Income and Corporation Taxes Act 1988 coming into play. Increased reserves are of benefit to members in that they reduce the influence of contribution holiday considerations in judging the appropriate balance between actions that benefit the company and those that benefit the members. Increased reserves are of benefit to members because they increase the likelihood of fulfilling the expectations that members were given, by IBM and the Trust, that erosion in the value of their pensions would be made less if economic conditions were favourable.

The decision to allow transfers benefited IBM because otherwise IBM would have had to fund the Defined Contribution plan. The decision made benefited IBM, did not benefit any member, and was detrimental to some members.

The decision not to tell the Members about the Defined Contribution plan funding plan.

The addition of the word "Section" to the Members' Report cannot be regarded as telling the members of a plan to transfer approximately a billion dollars from the Defined Benefit funds. The appearance of the first of the transfers in the Members Report is similarly inconspicuous and inadequate. The transfers were detrimental to members for reasons given elsewhere. The members' reactions when they discovered the plan are a measure of how important it was that they be told.

The Trust chose to tell the retirees the minimum necessary to make the Members' Report accurate in fact, rather than to appraise the retirees of the change in their prospects.

This choice minimised the threat to IBM's balance sheet from Member's complaints.

The decision not to discuss proposing to IBM a 100% RPI pension increase in 1999.

At this time it was apparent how far IBM's pensions-in-payment policy had fallen behind the policy of companies following what the Occupational Pensions Board regarded as best practice. The reserves allowed for such an increase without actuarial risk to contribution holidays.

The proposal to discuss such an increase was made by Barry Morley at January 1999 meeting . Barry Morley was very experienced and knowledgeable in this area, as he had been a trustee for more than ten years and a pensions professional before that. He was Chairman of the Benefits Allocation Committee within the Trust.

The decision not to discuss was detrimental to members because it might well have led to a proposal to IBM. We cannot know for certain whether discussion would have led to a proposal, or whether IBM would have accepted the proposal.

The Trust management has said, with respect to increases that would ameliorate the degradation in members' pensions, that:

"In practice there have not been two sets of figures, one for what was proposed and one for what was approved because these numbers have been the same."

It follows that either:

The decisions about what IBM would approve were made outside the trust meetings and carried into the meetings.

or,

The trust decided it was not in the interests of the members to discover what IBM would or would not approve.

The decision not to discuss was potentially beneficial IBM and potentially detrimental for members because of its effects on the reserves and their use.

J D Serkes and T F Cadigan were eligible to vote at the specified meeting.

The decision to approve the pension aspects of IBM outsourcing contracts.

On 15th December 1997, when 514 people from Data Sciences transferred into the Defined Benefit plan, the "surplus" in the plan, per member, was reduced. Reduced reserves per member are detrimental to members' interests in the same way as reduced reserves are.

Whether IBM gained financially from such deals we can only guess; obviously they intended to.

In competing for such deals, IBM had an advantage over companies which were not able to use an existing pension scheme in this way, if there were any such competitors.

This decision was beneficial to IBM and detrimental to the members.

I do not have the details of other outsourcing contracts, but there have been some.

The decisions in the directions to the auditors about Statutory Surplus calculations.

I have explained elsewhere why retirees have an interest in whether there is a statutory surplus greater than 105%.

The Surplus Certificate is a tax return. The document, Statutory Instrument 1987 No. 412, "The Pension Scheme Surpluses (Valuation) Regulations 1987", which specifies how the calculation is to be made is not a series of hints to actuaries, it is the description of an algorithm. It specifies that the "Projected Accrued Benefit Method" shall be used and hence that no other should be used. It specifies what the actuarial assumptions and requirements for the purposes of the valuation by that method are to be, and hence that no others are to be used.

The document, in 5 (2)(c), says

(i) the valuation rate of interest shall be reduced to take account of pension increases at the rate provided for under the scheme, provided the net investment yield is not so reduced to an amount less than 3 per cent. per annum;

(ii) where it is provided under the scheme that there shall be regular reviews of pensions currently payable and for pension increases (not exceeding any relevant increase in the retail prices index) to be given at the discretion of the trustees or employer (subject to availability of funds), paragraph (i) shall apply as if references to increases included references to such discretionary increases;

Paragraph (i) does not apply to discretionary increases. Paragraph (ii) does not apply to the Trust as there are no regular reviews provided under the C-plan.

The Trust directed or allowed the auditors to make an allowance for pension increases, which the algorithm does not allow in the circumstances.

When asked what paragraph validated assuming discretionary increases in the 1997 audit, the auditors did not answer that question, but they did say:

"When carrying out the surplus test calculations for the last formal valuation in 1997, the actuary took into account both guaranteed increases to pensions in payment and an allowance for discretionary increases. The discretionary increases were taken into account having regard to clause 9 of schedule D of the Trust Deed and the Plan's history of granting and funding for discretionary pension increases. The allowance included for discretionary pension increases was determined by reference to the actual discretionary increases granted in the three years to the valuation date."

The referenced section of the deeds refers to events happening "from time to time". This very different from "regular", which the regulations are using in the sense "calculable in time".

I do not have a tax expert available to me. The trust has, and should examine the conflict. The IBM trust is an unusual case because almost all UK pension schemes review pensions annually.

Whether or not the regulations permit the increases to be assumed, they should not have been assumed. IBM has explicitly refused to guarantee increases. (If that was not known by the end of 1997, it should have been tested earlier.) Failure to guarantee would be of no value to IBM unless they envisaged not giving the increases. The UK Human Resources director does not recognise any interest of the retirees beyond their "entitlement" to pensions with zero increases. The IBM UK Finance executive J S Lamb, writing from the Office of the Chief Executive of IBM, takes the same position. In the light of these factors, the long term is not properly predicted by recent history. The Surplus Certificate is intended to be a reproducible calculation from the facts of the scheme, not an expression of the trustee's hopes.

The choice made by the trust benefits IBM because it reduces the prospects of "over-funding". It is to the detriment of retirees because it reduced the prospects for an enforced plan that distributed extra benefits. It follows the pattern of doing what IBM wants.

Other decisions

Barry Morley found that the way the trust was run prevented him from doing his duty as a trustee. Rather than continue not doing his legal duty he resigned. Barry has listed 12 items in support of his view that "IBM was not exercising an adequate duty of care which brings with it an obligation of good faith to employees". Barry has listed 10 items in support of the view that "The trustee is not acting in the best interest of the members and not equitably between active members and pensioners"

I am not privy to enough information to fully evaluate these items, but the trust is. The pattern of behaviour will not have been properly studied until these items are taken into account.

In support of the statements:

"This pattern establishes that the trustees were not applying the proper principles. From what the trust has said, it can be deduced that the wrong principle the trustees were applying was that what IBM wanted should be agreed to if the trustees considered IBM could conceivably do something, legally defensible, which would damage the members. This was a wrong principle to adopt, even though there were in practice no realistic threats from IBM."

The football striker who consistently misses is demoted. The roulette wheel that always yields the same number is faulty. Consistent failure to exhibit balance establishes that the Trust did not have a principle of balancing interests.

The trust is vague and selective in what it says about its reasons. In relation to the transfers to the M-plan "The Trustee consented to the introduction of the M plan and the employer contribution holiday in the context that IBM was able in any event to enjoy a protracted contribution holiday and could have run down the surplus by such means, terminated the Plan and introduced a new Plan with benefits equivalent to those under the M Plan. The Trustee was also aware that IBM had previously considered the possibility that the additional costs of providing statutory LPI increases, as required by the Pensions Act 1995, might be covered by an increase in the member contribution rate or, in the alternative, that the future accrual rate of benefits might be reduced."

Despite the vagueness, these statements are intended to explain reasoning: it writes "The Trustee agreed to this proposal for the reasons set out above..." It should be investigated whether these "reasons" match those recorded in minutes at the time, or recollected by individual trustees.

Notice the phrase "IBM was able in any event to enjoy a protracted contribution holiday". The term "in any event" shows that the Trust felt that IBM had unbridled power. It did not consider that its own powers, the results of actuarial calculations, any retiree opposition, or the employer's duty of good faith could affect the matter.

I have no idea if this bit of law still stands but:

In Imperial Group Pension Trusts Limited -v- Imperial Tobacco Limited [1991] 2 All ER 597, Browne-Wilkinson VC decided the employer owed its employees a duty of good faith in the exercise of all of its powers under a pension scheme. In exercising a pension scheme power, an employer is not acting as a fiduciary for the scheme members. Nor, however, can it act solely with regard to its own interests. The company can have regard to its own financial interests but only to the extent that, in so doing, it does not breach the obligation of good faith to its employees. The duty is one of "trust and confidence" - that the employer should not abuse its powers under the scheme. Such an abuse of "trust" will depend on the particular circumstances complained against. But this duty creates the possibility that actions which are quite normal amongst employers could still, in the right circumstances, constitute a breach of the duty. Thus, in Stannard -v- Fisons Pensions [1992] IRLR 27 at 32, Staughton LJ could envisage circumstances in which a contribution holiday would be a breach of the duty.

The Trustee does not say that IBM made any threats. The scenario where they did, and there was a battle of wills between IBM and the Trust is interesting to speculate about. The clash of the Trust powers, for instance to alter deeds, and IBM's powers, limited by its duty of good faith, would be titanic. No doubt IBM's powers to appoint trustees would triumph. But we know this scenario is totally unrealistic. This particular Trust never would oppose IBM plans. Even in the absence of a threat it chose to imagine one.

The suggestion that there was an IBM threat, or imagined threat, to make the active employees pay more if the Trust did not do what IBM wanted is an interesting one. IBM could have made employees pay more, whether or not the Trust agreed to transfers. However, it is unlikely that would have had much effect since employees, unlike retirees, can take their services elsewhere. Commercial logic requires that if employees were to be made to pay more in contributions they would require more pay, if IBM was to recruit and retain the same calibre of staff.

We do not know IBM's reason for wanting an M-plan but it can reasonably be assumed to be the same as the reason other companies have introduced such plans - the extra predictability. Nobody has suggested that the only reason IBM wanted an M-plan was to exploit the C-plan reserves.

Mike Cawley has suggested that there was a (perhaps unrecorded) agreement on a three part bundling where the M-plan, the transfers, and improved pensions in payment were all to be introduced. Any trustee who had agreed to such a bundling would have been acting in accord with their duty when agreeing to just part of that bundling, but susceptible to being "Stitched up by Gerstner's moneymen".

In support of the statements:

"The trust system was uniquely stressed by a combination of factors. The change of management culture by IBM US early in 1990s has produced powerful forces for short term profit maximising without any balancing forces through union recognition. There is no force outside the Trust acting for the benefit of retirees."

IBM US follows an aggressive policy in seeking profits and a high share price. In today's corporate climate that is to be expected. It is to be expected that IBM US will impose the same policy on its subsidiary IBM UK. It is to be expected that the senior executives in both companies will think in accord with this policy. The policy is characterised by operating near the margins of the law, reducing pension benefits, and motivation through stock options.

In respect of active members of the IBM UK pensions scheme this profit seeking will be balanced by competition - the ability of active members to take their skills to another company.

For retirees there is no balancing force. IBM UK does not negotiate with unions and in any case few retirees are union members - the result of previous decades when there was an implicit contract between company and employees that the company would be one of the best and unions were not necessary.

It was the custom for the Chairman of the Trust to meet with trustees from the US when they flew in for a trustee meeting. It would have been proper for these gatherings to discuss company profit potential because these gatherings were not in the roles of trustees. It was the custom for the company appointed trustees to meet immediately prior to the trustee meeting. It would have been proper for these gatherings to discuss company profit potential because these gatherings were not in the roles of trustees.

In a subsequent trustee meeting, they would have been fully aware of the company profit potential of their actions. Without a balancing influence benefiting the members, this made it more difficult for them to separate their thoughts in the meeting from plans formed outside.

The majority of trustees at trust board meetings were paid to be there. The Occupational Pensions Regulatory Authority has acknowledged the possibility of payments affecting actions, as in the case of the sole trustee who took 20 years in winding up a trust while being paid to do it. While paying trustees is common practice, and not something to be discouraged, it is an incentive to remain in favour with IBM.

Some trustees held IBM stock options. These produce an incentive to remain in favour with IBM because IBM can prevent the options being exercised. They also produce an incentive to actions that will raise the share price.

Some of the trustees were IBM employees. IBM's hire-or-fire control over their livelihoods was a powerful incentive to remain in favour with IBM.

At least one trustee, although not an IBM employee, was a director in some of IBM's companies. This would increase the skill required to separate his roles.

Some of the trustees were involved with IBM worldwide financial activities. If there was ongoing activity between IBM and the authorities over these activities, such as the Inland Revenue enquiry into the change of UK subsidiary royalty payments from 8% to 12%, then this would be an incentive to remain in favour with IBM, so as to retain the protection IBM's resources afforded them. I was not able to conclude anything about this.

In support of the statements:

"The trust system was stressed by the actual choice of IBM appointed trustees. The choice of people under extreme pressure to deliver IBM advantages made it excessively difficult for them to put aside those pressures in their trustee roles."

The Trust is not clear about when the decision on transfers was made. It has written:

"The proposal to apply "surplus" to meet IBM's liability to contribute to the M Plan was expressly presented to the Trustee in December 1996 and is reflected in the wording of rule 1(2). The Trustee agreed to this proposal for the reasons set out above and considers that rule 1(2) adequately documents this agreement."

At first sight, this says the proposal was agreed in December 1996. However, the Trust also writes:

"The introduction in 1997 of the M Plan within the existing trust was discussed and agreed at its meeting of 12 December 1996 and at its meeting of 24th April 1997, when the new definitive trust deed and rules was (sic) adopted."

Since any decision has to be made at one meeting or another, not at both, it must be the case that introduction of the M Plan was discussed in December 1996 and decided upon in April 1997. Since the decision about meeting IBM's liability to the M Plan could not precede the decision to introduce the M Plan, that too must have been discussed in December 1996 and decided upon in April 1997.

T F Cadigan was a director of the trust from 23/05/95 to16/12/99. J D Serkes was an alternative director to T D Cadigan over the period 25/04/96 to 20/02/97. J D Serkes was a director from 20/02/97 to 29/07/99. I do not actually know if they were present or what they did, but Cadigan (or Serkes as alternative) was eligible for the December 1996 meeting and both were eligible for the April 1997 meeting.

The attendees at the decisive April 1997 meeting were, as far as I can deduce:

Company appointed:

Elected: