IBM UK Holdings Ltd & Anor v Dalgleish & Ors

This is an application for declaratory relief concerning certain proposed changes first announced by the First andSecond Claimants in July 2009, and announced in their final form in October 2009, in relation to two final salary pension plans that operated for the benefit of their UK workforce, namely the IBM Pension Plan ("the Main Plan") and the IBM IT Solutions Pension Scheme ("the I Plan", together "the Plans"). These changes were announced following a project known internally as "Project Waltz".

In these proceedings, the changes are challenged on behalf of the members of the Plans.

The First Claimant ("Holdings") is the principal employer for the purposes of the Plans (which means that certain important powers and discretions under the relevant rules are vested in it). The Second Claimant ("IBM UKL") also participates in the Plans and is the employer of most of the active members. I shall refer to them together as "IBM UK" in order to distinguish them from the US parent IBM Corporation. IBM Group's headquarters is in Armonk and is known as "CHQ". I use "IBM" as a catch-all to include the Group as a whole.

The First and Second Defendants, Mr Dalgleish and Ms Harrison, have been selected to be representativebeneficiaries ("RBs") for the purposes of the Project Waltz Proceedings. They are respectively members of the Main Plan and the I Plan.

The Third Defendant ("the Trustee") is (and has at all material times been) the sole corporate trustee of the Plans. It takes a neutral role in these proceedings (although it has adduced factual evidence that it considers to be of assistance to the Court).

The Deeds and Rules which currently govern the Plans are described in the following paragraphs.

The Main Plan is currently governed by three Deeds dated 24 April 1997 (as amended) comprising:

There were many earlier editions of the Main Plan Trust Deed and Rules pre-dating 1997, stretching back to

The Main Plan has separate DB and DC sections. The DB sections comprise the following structures:

The main DC section of the Main Plan is known as "the M Plan". There is also a DC section known as "the Enhanced M Plan" which was established in 2006 to which various DB members had a one-off option to transfer at that time as I explain below, an option arising in the context of Project Soto.

The I Plan is currently governed by a single Definitive Trust Deed and Rules dated 25 July 1994 (as amended) ("the I Plan Deed and Rules" or "the I Plan Rules" as appropriate). The I Plan Deed and Rules are the instruments which originally created the I Plan; there were no previous editions. References in this judgment to the I Plan Deed and Rules are (unless otherwise stated) references to the then current version of those instruments as they stood at the time of Project Waltz, as reflected in the informal consolidation.

The I Plan is a separate trust from the Main Plan. It provides defined pension benefits. I will refer to the I Plantogether with the DB sections of the Main Plan as "the DB Plans", or the "UK DB Plans".

Although reference is made in the Main Plan to "sections", it is not in fact a sectionalised plan in the sense ofhaving separate funds appropriated to different sections of the membership and their benefits. There is a single trust and single trust fund albeit that different classes of member have different benefits under the rules of the DB sections and the DC sections.

I will come to the package of changes comprising of the Project Waltz changes in due course. One element("Element 1") was the closing (subject to certain exceptions) of both of the DB Plans to future benefit accrual with effect from 6 April 2011. This was purportedly effected by IBM pursuant to the exercise of a power to direct that any specified person or class of persons shall cease to be a member. There is an issue about whether that power was validly introduced into the Main Plan and there is an issue about whether the powers in both the Main Plan (assuming it was properly introduced) and the I Plan were validly exercised. Although neither power contains the word "exclude" or similar words they have been referred to as "the Exclusion Powers" and that, or its singular, is the expression I will use. I use that expression generally to refer to the power in the Main Plan whether in the 1990 Trust Deed and Rules or in the Main Plan Trust Deed and Rules (ie the 1997 iteration). The wording is the same and it is not suggested that the meaning has changed. These issues have given rise to Issues 1 to 3 in the List of Issues (see paragraphs 46ff below).

IBM's case (subject to a detailed point to which I will come in due course) is that the Exclusion Powers, as amatter of construction, can be exercised so as to direct that all existing active members shall cease to be members. IBM also contends that, having that construction, the Exclusion Power was validly introduced by way of amendment into the Main Plan. The RBs' case is that the Exclusion Powers cannot, as a matter of

Quite apart from those differences between IBM and the RBs in relation to the Exclusion Powers, the RBschallenge each element of the proposals arising out of Project Waltz (including the exercise of the Exclusion Powers) as a breach of what, by way of short-hand, I refer to as the Imperial duty (discussed in detail later in this judgment).

The parties, in their written opening and closing submissions, have all described the essential elements of ProjectWaltz. They have done so focusing in different ways on different aspects. I have taken account of all of it in setting the background. I have found particularly helpful the written opening submissions on behalf of the Trustee which I consider present an accurate description of the background. What follows, all the way to paragraph 59 below, is taken largely from those submissions. Although I interpose my own comments on a number of occasions, the substantial work is that of the Trustee's team. It seemed to me that I would be reinventing the wheel in the most unnecessary way if I were to re-write this work in my own words when it is essentially uncontentious. I do not need to identify, as I go along, where I am taking verbatim or almost verbatim what appears in those submissions but I readily acknowledge that a substantial part is not my own work. I am very grateful to the Trustee's team for their thorough, neutral and objective work which I consider to be wholly reliable.

Holdings presented the proposed Project Waltz changes to the Trustee in May 2009 and to the IBM UKemployees in July 2009. Approximately 5,000 employees were affected by Project Waltz, being, broadly, the then active membership earning DB benefits. The Project Waltz proposals, as presented to the members, were summarised in an email from Mr Brendon Riley ("Mr Riley"), the then general manager of IBM UK, to the employees dated 7 July 2009 which stated as follows:

Project Waltz was finally formulated in 2009. There were five essential elements affecting active DB members:

I now turn to explain each of Elements 1 to 5 in more detail.Element 1 – cessation of DB accrual from 6 April 2011

To accomplish Element 1 of Project Waltz (cessation of accrual in the DB Plans), Holdings served notices ("the Exclusion Notices") on the Trustee on 28 May 2010 purportedly excluding active DB members from membership of the Main Plan and the I Plan with effect on and from 6 April 2011. The Notices were expressed to be served pursuant to the Exclusion Powers which were these:

The Main Plan Exclusion Notice purports to exclude the entire active DB membership from the Main Plan witheffect from 6 April 2011. The I Plan Exclusion Notice purports to exclude all of the active members (apart from 455 named I Plan members) from the I Plan with effect from the same date, the 455 members being those who had contractual entitlements to continued DB accrual or whom IBM decided to keep as active members in order to avoid triggering a statutory debt. Due to doubts about whether the terms of the Exclusion Notices comprehensively covered all the active DB members, supplemental Exclusion Notices were served by Holdings on 28 March 2011 with effect from 6 April 2011.

Taking the original Main Plan Exclusion Notice dated 28 May 2010 as an example, the Exclusion Notices was asfollows:

A similar direction appeared in the I Plan Exclusion Notice of the same date.

In the Main Plan Exclusion Notice, there then followed a statement which did not appear in the I Plan ExclusionNotice as follows:

IBM UK employees were informed on 7 July 2009 that there would be no further pensionable salary increasesfor DB members whilst they were active members of the DB Plans. Holdings sought to accomplish this through the 2009 Non-Pensionability Agreements as follows:

IBM UK had a longstanding practice of allowing or encouraging employees to take early retirement beforeNormal Retirement Date ("NRD"), often during their 50s. I will explain later in this judgment the detailed provisions of the Rules of the DB Plans, but, in summary, under the Rules of the C and N Plans, early retirement pensions for active members were reduced by 0.25% for each month between the commencement date of the pension and the member's 60th birthday, ie an ERDF of 3% pa taken back from age 60 for retirements from active status before age 60, and no reduction for such retirements from age 60. These were advantageous terms for early retirees which were more expensive (for the Plans) than cost-neutral.

Similar provisions applied in respect of some DSL Plan DB members. By contrast, deferred C and N Planmembers who drew their pensions before NRD were subject to a larger, cost-neutral, ERDF applicable to all years preceding NRD (and, again, similar provisions applied in respect of some DSL Plan DB members). However, under the Rules of the C and N Plans as they stood at the time Project Waltz was announced in 2009 (and for some DSL Plan DB members), retirement of active members before NRD was subject to Holdings' consent. Such consent had always been forthcoming in practice in the past.

As part of Project Waltz, Holdings decided to introduce a new restrictive early retirement policy with effect from6 April 2010 which would in nearly all cases prevent early retirement on these advantageous terms for active C and N Plan members (and for some DSL Plan DB members).

The early retirement window was an opportunity for such members to seek Holdings' consent to retire under theexisting permissive and financially attractive policy (from the point of view of the applicable discount factors) before the new policy came into force. The effect of this was that, after the closure of the early retirement window, an active DB member wishing to "retire" early without having been able to obtain Holdings' consent (insofar as consent was needed under the Rules applicable to him) would have to leave service and take early payment of his deferred benefit: this would involve the application of a less generous discount factor than the ERDF that would apply to an early retirement with consent from active membership.

The early retirement window was implemented as follows:

I note that (subject to giving 12 months' notice of their intention to retire) active I Plan members did not requireHoldings' consent for early retirement so, on the face of it, the early retirement aspects of Project Waltz did not affect them in the same way as the members who required consent for early retirement. Similarly, some active DSL Plan DB members did not require Holdings' consent for early retirement. Active I Plan members were also subject to less favourable ERDFs (5% p.a. for each year of early retirement taken back to NRD), so the differential between retiring from active status and retiring from deferred status was less significant for them. Element 4 – new early retirement policy

Holdings' new early retirement policy adopted with effect from 6 April 2010 is recorded in a written policywhich includes the following:

The second paragraph quoted above would have caught C and N Plan early retirements from active status, whichunder the Rules were on non-cost neutral terms and subject to Holdings' consent (and also some DSL Plan retirements).

However, in earlier litigation (see IBM United Kingdom Pensions Trust Ltd v IBM United Kingdom Holdings Ltd [2012] PLR 469 - "the Rectification Action"), I held that the Rules of the C Plan were to be rectified so as to remove the requirement for Holdings' consent for C Plan early retirements from active status at or after age 60.

The new policy nevertheless catches C Plan early retirements before age 60 and N Plan early retirements.

One further complication should be explained: if (under Element 1) the accrual of further benefits by DBmembers has ceased, it might be thought that they are no longer active members and that the early retirement rules for active DB members (and the favourable ERDFs that apply to such members) are no longer relevant. However, as explained under Element 5 below, as part of Project Waltz, Holdings proposed to allow the excluded DB members to retain the favourable early retirement rules that applied to them as active members in respect of their accrued DB service (but subject to Holdings' ability to withhold consent to early retirement, where applicable). Thus the early retirement rules applicable to active members remain relevant notwithstanding the purported cessation of DB accrual under Element 1.

As already noted, the DB members were informed on 7 July 2009 that they would have the option of joining theDC section of the Main Plan (ie the M Plan) once their active DB membership ceased on 6 April 2011. Holdings proposed to achieve this by excluding members from the DB Plans using the Exclusion Powers and then allowing those who had opted for M Plan membership to join the M Plan section of the Main Plan. IBM considered that there was an argument that it would be possible for the affected members to join the M Plan without Trustee consent and wished to ensure that those members would be able to take advantage of the option if that argument was a good one.

Thus for C, N and DSL Plan DB members, they would be excluded from the Main Plan and then be immediatelyreadmitted to the Main Plan but in the DC section. As quoted above, the Main Plan Exclusion Notice dated 28 May 2010 specifically provided that it was without prejudice to the ability of the excluded DB members to join the M Plan. For I Plan members, they would be excluded from the I Plan and then admitted, for the first time, to the Main Plan. The excluded DB members' benefits in respect of future contributions under the M Plan would be governed by slightly different terms from the terms governing ordinary M Plan benefits. Those terms would be known as the "Hybrid M Plan".

Holdings also proposed to create a new personal pension plan, the IBM UK Personal Pension Plan ("the IBM UK PPP"), a contract-based arrangement administered by Standard Life. This would be a DC arrangement outside the Plans, which the excluded DB members would have the option to join on 6 April 2011 as an alternative to the M Plan. They would also have the option of having no IBM pension provision at all for future service.

In addition, Holdings proposed to create "hybrid deferred" status in both the Main Plan and the I Plan for the excluded DB members' accrued DB benefits. This would involve amending the DB Plans to modify the excluded members' past service DB benefits that had been accrued up to the date of their purported exclusion. Rather than simply becoming ordinary deferred benefits, those benefits would be "hybrid deferred" and would have the following features:

Further, the excluded DB members would be entitled to 2 years of enhanced employer contributions to their newDC arrangement for future service (known as "step down" contributions) of 14% of pensionable salary in 201112 and 10% in 2012-13.

The proposed terms for hybrid deferred status were set out in Holdings' letter to the Trustee dated 23 September2009. Draft Interim Deeds have been prepared to give effect to hybrid deferred status but nothing has yet been executed.

The creation of hybrid deferred status, and (at least arguably: see paragraph 38 above) the readmission of theexcluded Main Plan DB members to the M Plan, could not be achieved unilaterally by the Company but would involve the Trustee exercising its powers under the Plans. The classes of member affected by Project Waltz

The classes of member directly affected by the Project Waltz proposals were, broadly speaking, the then activemembership of the DB Plans comprising approximately 5,000 employees, that is:

The parties have agreed a list of the issues for decision in the present trial and, depending on the result of this hearing, for decision following a further hearing. I attach as Annex A a list of the issues for decision in the present trial ("the List of Issues"). There was a further issue, Issue 4A, but that has now gone. It relates to section 91 Pensions Act 1995. In the light of my decision in Bradbury v BBC [2012] EWHC 1369, [2012] PLR 283 ("Bradbury") the Issue is not pursued although the point is left open to be taken should the present case ever come before the Court of Appeal. Issues 1 to 3 raise what are essentially legal arguments or, at least, arguments where the facts are not a matter of any real contention. Issues 4 and 4B are concerned with the Imperial duty of good faith and the contractual duty of trust and confidence owed by an employer and an employee to each other; they raise many issues of fact on which I have heard a large amount of evidence over a number of weeks, including disputed expert evidence from accountants and actuaries.

Following Holdings' presentation of the Project Waltz proposals to the Trustee in May 2009, and after aconsiderable amount of discussion between them in the ensuing months (in particular, various meetings and correspondence between Holdings and representatives of the Trustee during May-July 2009), the Trustee informed Holdings that there were sufficient doubts about the lawfulness of the Project Waltz proposals that the Court would have to be asked to rule on the issues. The Trustee's concerns arose partly as a result of the previous history of the Plans and in particular the Ocean and Soto projects, matters which I will come to later in this judgment. In a letter dated 18 June 2009, the Trustee informed Holdings of the legal advice it had received (from Leading Counsel) raising its concerns over a number of issues which are now encapsulated in the List of Issues.

The Trustee nonetheless continued negotiating with Holdings to see if Holdings would alter the Project Waltzproposals in such a way that would enable the Trustee to conclude, subject to legal advice, that IBM had complied with its Imperial and contractual duties. Although the Project Waltz proposals went through various iterations and Holdings made changes (such as the introduction of hybrid deferred status) these changes did not overcome the Trustee's concerns.

Consequently, the Trustee informed Holdings' representatives at the Trustee meeting on 22 October 2009, andsubsequently confirmed in writing, that it would not be prepared to proceed without an application to Court to affirm the lawfulness of the Project Waltz changes. Emergence of the 1983 C Plan issue

By this stage, matters were developing on the ground. As of October or November 2009, Holdings was seekingto initiate the early retirement window (Element 3) as well as inviting members to enter into the 2009 NonPensionability Agreements (Element 2). In addition, an issue of rectification of the C Plan arose. This was the subject matter of the Rectification Action (see paragraph 35 above), in which I gave judgment last year. In that case, it was asserted that members who joined the C Plan when it was created in 1983 had done so on the basis of announcements promising them a right of "flexible retirement", ie a right to retire from active status at any age from 60 to 63 at the member's election on an unreduced pension.

Individual employees raised this point with Holdings from July 2009 onwards when Project Waltz wasannounced. The issue took on prominence on 5 October 2009 when an in-house solicitor employed by IBM UK, Mr Tony Ford, emailed senior executives of Holdings (copying in Mr Newman of the Trustee) setting out a detailed account of the 1983 announcements, offering copies, and calling for the C Plan Rules to be rectified.

On 9 November 2009, DLA Piper UK LLP ("DLA"), by this stage acting for a group of members, wrote to Holdings (with a copy to the Trustee) referring to an Opinion from Leading Counsel in support of the existence of a rectification claim. DLA also referred to evidence from some of the key witnesses in the Rectification Action (who eventually gave witness statements in support of the rectification claim relied on at the May 2012 trial).

The emergence of the rectification issue was a further cause for concern on the part of the Trustee, in particularbecause it was directly relevant to the early retirement window, for which Holdings had set the rapidlyapproaching deadline of 4 December 2009 for members to decide whether to take early retirement.

By a letter dated 13 November 2009, the Trustee's solicitors formally raised the rectification issue with Holdings'solicitors and asked for an extension of the early retirement window. Holdings rejected this request and, relying on the evidence of Mr Sam Ellis (who gave evidence in the Rectification Action in May 2012), its solicitors wrote to DLA on 4 December 2009 contending that there were no grounds for rectification and that there was no legal right for C Plan members to retire from age 60 on an unreduced pension. Holdings did, however, make a qualified concession on or about 4 December 2009 in relation to certain C Plan members, namely that employees who joined the C Plan at its inception on 6 July 1983 would, once they reached age 60, be given consent to retire early with an unreduced pension, other than in the most exceptional, business critical, circumstances (this concession was slightly widened in early 2010 to those who joined the C Plan up to 30 November 1983 when the November 1983 update to the Employee Handbook was published).

Eventually proceedings were commenced by IBM on 28 May 2010. I do not need to go into an explanation ofwhy it was IBM, rather than the Trustee, which did so. There are perfectly good reasons for this and no criticism is made by anyone about it. Nor do I need to go into the history of the exchange of evidence including the difficulty faced by the Trustee in remaining, on the one hand, entirely neutral as between IBM and the RBs and, on the other hand, explaining to the Court how it is said that IBM was in breach of its Imperial duty. All this is water under the bridge since the Trustee's witnesses were cross-examined by Mr Tennet and Mr Stallworthy (for the RBs) and by Mr Simmonds (for IBM) so that, neutral or not, their evidence was exhaustively considered.

Another aspect of the procedural background which I should mention is that IBM applied to expedite the trialbut that application was rejected at a hearing on 25 October 2010. Following that, the Trustee requested Holdings to postpone the implementation date. Ignoring the warning of Irving Berlin that there may be trouble ahead, Holdings decided to dance. It declined the Trustee's request; and, in the face of what it perceived as a threat by IBM putting at risk the availability of hybrid deferred status or M Plan membership for former DB employees, the Trustee, under protest, agreed to implement Project Waltz on an interim basis. This was on the footing that the steps taken would be unravelled if the Court ruled that the project was unlawful. And so the Waltz had begun.

As I have already mentioned, the interim proposals have not been formally implemented by an Interim Deed. Inpractice, Holdings gave DB employees the option, as from 6 April 2011 (with hybrid deferred status for their accrued DB benefits):

The current position, assuming the validity of the Project Waltz changes is as follows:

If, in contrast, the Project Waltz changes are invalid and unlawful, the true position may be that all DB membershave continued to be members of their DB Plans and have accrued DB benefits linked to final salary. Overview of the Plans

It is useful to have an overview of the governance, structure, terms and administration of the Plans, this beingthe context in which the disputed issues arise. This section of my judgment is again based largely on the written opening on behalf of the Trustee for which I am again very grateful. This description is not contentious (unless otherwise indicated).

The Trustee is the sole trustee of both the Main Plan and the I Plan, although as already noted they are separatetrusts. During the period relevant to the proceedings and up to 2012, the Trustee had a board of 12 directors, of whom 4 were member nominated, the rest being nominated by Holdings.

The board of the Trustee delegated all functions to a committee of the full board known as the TrusteeManagement Committee. Meetings of this committee are referred to in the contemporaneous documents and evidence as Trustee Management Meetings or "TMMs". The board of the Trustee also established a subcommittee known as the Investment Committee which was authorised to decide and implement the Trustee's investment strategy. Until about July 2004, the Investment Committee decided on the overall asset allocation for the Plans, but from that time the basic allocation decision as to the split between reward-seeking assets (equities/property/alternatives) and bonds was decided by the TMM, with the Investment Committee implementing the detailed investment strategy in accordance with the parameters set by the TMM.

The Main Plan and the I Plan each had their own TMM and Investment Committee, and separate minutes werekept for each. However, in practice the TMMs had the same membership and met on the same occasions and their meetings were held concurrently, and likewise for the Investment Committee. As can be seen from an examination of the actual minutes, those relating to the Main Plan and those relating to the I Plan mirror each other to a large extent, since the Trustee's decisions applied equally to both Plans, although, naturally, Planspecific discussions appear only in the minutes applicable to that Plan.

A number of the Trustee directors were IBM executives who from time to time represented IBM's interests inpensions matters. To accommodate this, these executives were designated as "conflicted Trustee directors" (if and when the issue being considered required it) and they were (on those occasions) treated as representatives of Holdings, whereas (in relation to such issues) the other directors were designated as "non-conflicted Trustee directors" who were able (on those occasions) to represent the Trustee in its discussions with Holdings. As is clear from the evidence, on various occasions, discussions between Holdings and the Trustee actually took place within the TMMs, with the conflicted directors representing Holdings and the non-conflicted Trustee directors representing the Trustee.

It is to be noted, however, that not all IBM executives on the Trustee board were treated as conflicted directors.Thus a number of non-conflicted Trustee directors were in fact IBM executives.

During the period of principal relevance to these proceedings (2004-09, when the Ocean, Soto and Project Waltzchanges were announced):

The day-to-day administration of the Plans was carried out by Holdings' in-house pensions administrationorganisation, known as Pensions Trust, staffed by Holdings' employees and led by the Pensions Trust Manager. From 1999-2011, the Pensions Trust Manager was Mr David Newman ("Mr Newman") who gave evidence on behalf of the Trustee. During this period, Mr Newman was also company secretary of the Trustee and the secretary of the TMM and Investment Committee. As Pensions Trust Manager, Mr Newman had delegated authority, within certain parameters, to manage the financial business of the Plans and oversee the investment of the Plans' assets.

Pensions Trust had various operating departments headed by other managers (eg Ms Suzanne Ross, the Pensions Investment Manager), but ultimate responsibility for all aspects of Pensions Trust's activities remained with the Trustee.

The same actuary was appointed for both the Main Plan and the I Plan. During the 2000s, the appointed actuarywas Mr Stephen Gooch of Aon (until 13 December 2002), Mr David Eteen of Aon (14 December 2002 to 7 December 2003), Mr Greg Alexander of Watson Wyatt (8 December 2003 to 11 May 2009) and Mr Graham McLean of Watson Wyatt (renamed Towers Watson in 2010) (11 May 2009 to present).

The principal investment advisers were Aon and (from December 2003) Watson Wyatt (later Towers Watson). Itis to be noted that from time to time Holdings also retained the services of Watson Wyatt, but such services were provided by different teams and offices within Watson Wyatt and were separate from the advice given to the Trustee.

I have already introduced the Main Plan and I now give some more detail. The Main Plan was established by anInterim Trust Deed dated 3 April 1957 and took effect from 1 January 1957; its original Definitive Trust Deed was dated 19 May 1959, and was replaced by subsequent definitive editions in 1977 and then on a number of occasions in the 1980s and 1990s.

One of those definitive editions was the deed and rules dated 14 December 1983 ("the 1983 Trust Deed and Rules") which introduced the C Plan, a contributory DB arrangement, on 6 July 1983 and closed the N Plan to new members after 5 July 1983. The 1983 Trust Deed and Rules were in turn replaced by a new form of Definitive Deed and Rules dated 1 October 1990 ("the 1990 Trust Deed and Rules"). The 1990 Trust Deed and

The next major round of changes was in 1997 when the previous trust documents were replaced and split intothe three current Deeds referred to in paragraphs 6ff above.

A number of additional Deeds making discrete amendments have been executed from time to time. Amongstthese amendments, two changes of note have been made since the current versions of the Deeds and Rules were introduced in April 1997:

The I Plan was created by Holdings in 1994 to be used as the DB pension scheme for DB members whotransferred to IBM employment following outsourcing by the public sector or certain corporate acquisitions. The I Plan was established with effect from 28 July 1994 by a Definitive Deed and Rules dated 25 July 1994 made between Holdings as Principal Employer and the Trustee.

Unlike the Main Plan, the I Plan does not have different sections. However, like the Main Plan members, as partof the Soto changes in 2006, active I Plan members were given the option to transfer to the new Enhanced M Plan section of the Main Plan on 6 July 2006. The transferring members similarly became "Enhanced Deferred Members" of the I Plan in respect of their accrued DB benefits which were left in the I Plan, pursuant to a Deed dated 30 March 2006.

All of the Main Plan Deeds since 1990 have adopted the style of the 1990 Deed with essentially similar provisions and language. Likewise, the I Plan Deed and Rules were based on the Main Plan Deeds and contain similar provisions and language to the Main Plan, although with a different benefit structure. In fact, the I Plan Deed and Rules (prepared in 1994) were closely modelled on the then current Main Plan Definitive Trust Deed and Rules (the 1993 version) and adopted an identical drafting structure. Thus many of the standard provisions of both the Main Plan and the I Plan are the same. In particular, in both the Main Plan and the I Plan, "Member" is defined as meaning, broadly, active members, that is to say current employees, and deferred members are described in both Plans as "Deferred Retirees". "Deferred Retirees" are not to be confused with another class, namely "Postponed Retirees", who are former members who remain in service after their NRD with Holdings' express consent.

Both Plans contain powers of amendment which are subject to restrictions. The terms of the amendment powersare relevant to Issues 1 to 3 of the List of Issues. The Main Plan amendment power is now found at Clause 1 of Part II of the Main Plan Definitive Trust Deed and the I Plan amendment power is at Schedule B rule 1 of the I

The I Plan amendment power, at Schedule B rule 1 of the I Plan Rules, is in the following terms:

The legal background to the discussions between the Trustee and Holdings in relation to scheme fundingincluded the balance of power as to the setting of employer contributions. This is described in the Trustee's written opening submissions and was, in summary, as follows:

Mr Spink has included in his opening written submissions a review of Mr Lamb's evidence concerninginvestment strategy. In relation to the matters set out below, I accept Mr Lamb's evidence.

As of the early 2000s, the Trustee was still following an investment strategy strongly focused on equities. As of2003, the Main Plan DB sections were invested approximately 80:20 in equities/property versus bonds, whilst the I Plan was invested 100% in equities/property.

After the dot.com bubble had burst, markets fell and became volatile, and the DB Plans moved into deficit in

Thereafter, discussions during 2002 and 2003 about asset allocation took place against the background of thedeficit referred to above, which prompted a more fundamental review of the equity/bond mix, a temporary suspension of the gradual move from equities into bonds decided upon in May 2002 and then discussion as to whether to recommence that transfer.

During 2003, in view of the emerging deficit, the Trustee sought and obtained reassurance about the Company'sintention to support the Plans. Watson Wyatt became the new actuarial/investment advisers in December 2003. The newly appointed actuary for the Plans, Mr Alexander, adopted a new, more conservative approach than his predecessor for the AVR as at 31 December 2003, with a focus on (amongst other things) employer covenant and "risk budgeting" (Mr Alexander's consideration of the latter led to a conclusion that 97% of the Main Plan's investment risk was attributable to its equity holdings and a recommendation to the Trustee that it think seriously about a move into alternative asset classes).

The draft AVR as at 31 December 2003, in which the assumed future investment returns included a "margin forprudence", showed a very large deficit of £1.2bn.

This led to extended discussions, the eventual outcome of which was that the provision of a funding guaranteedated 9 December 2004 ("the Funding Agreement") (see paragraph 111 below) in exchange for which (inter alia) the Trustee accepted the "best estimate" basis for the 31 December 2003 AVR, leading to a lower deficit.

At the same time as these discussions were going on, the Trustee was reconsidering asset allocation, which atthis stage was 75:25 equities/property versus bonds in the Main Plan. From around July 2004, the responsibility for making the fundamental decision as to the equities/bond split was transferred from the Investment

It was at this time that Holdings proposed the Ocean changes to the DB Plans to increase employee contributionrates (or reduce accrual rates). The result of these various developments was that in October-December 2004, the Trustee agreed a linked package of measures with Holdings which included: i) support for the Ocean changes;

I consider in detail below the basis on which the Trustee thought it was entering into this bargain.

The Trustee was aware of the risks involved in the investment strategy but was prepared to take these risks onthe basis that, because of the Funding Agreement, it would be able to rely on the support of IBM World Trade Corporation ("IBM WTC") (a subsidiary of IBM Corporation) if investment returns were poor.

In November 2005 to January 2006, Holdings proposed and discussed the Soto changes with the Trustee forconsideration.

The Trustee reconsidered its investment strategy around the time of the Soto changes, and was advised that if

As a result of the large cash contribution made to the DB Plans by Holdings as part of the Soto changes, the DBPlans had moved into surplus by October 2006. The finalised AVR as at 31 December 2006 (completed in 2007, on the "best estimate" basis) showed that both Plans were in surplus.

During 2007, in connection with the preparation of the 31 December 2006 AVR, the Trustee initiated discussionsabout obtaining a 3 year extension of the Funding Agreement to replace the 3 years that had already expired. IBM agreed to this; it was also agreed that IBM Corporation would replace IBM WTC as the guarantor of the Company's funding obligations, and that for a 3 year period the Trustee would limit further increases in the bond allocation of the Main Plan to 3% p.a. (and to 20% over the 3 year period for the I Plan). The Trustee indicated to Holdings that the extension of the guarantee would put the Trustee in a better position to consider IBM's requests for the Trustee to invest in reward-seeking assets.

By this stage (late 2007), the Main Plan had an asset allocation of 35% bonds (as had been agreed in late 2004). Thus the effect of agreeing to limit the move to bonds to a gradual shift of 3% p.a. for 3 years meant that the Main Plan allocation to bonds was targeted to be 44% by late 2010.

The Trustee thought that it had agreed to these restrictions to the asset allocation on the basis that it wasimplementing a long-term investment policy, rather than one in which short-term asset falls might encourage Holdings to close the DB Plans. The Trustee and Holdings were well aware of the risks involved in the investment strategy (including their susceptibility to financial "shocks" due to equity market falls). I will be considering these matters in detail later when I will make relevant findings.

One feature of the Trustee's investment strategy was to target "self-sufficiency" (ie solvency on a discontinuance basis) at the end of the period of the Funding Agreement as a result of the gradual increase in the allocation to bonds over the period. The background against which the Funding Agreement was extended in 2007 was that the Trustee was proposing to target a discontinuance funding level of 105% by the end of the guarantee period. Mr Alasdair MacDonald said that an allocation of 55% bonds would be appropriate to achieve this.

Also in 2007, the Trustee introduced a "liability driven investment" programme, ie a programme of investing in swaps to reduce the interest and inflation risks resulting from the duration mismatch of the bond portfolio.

During the 2007-08 period, the Trustee sought and received further assurances from Holdings that there were nodiscussions about or plans on foot to change the Plans (albeit there were no absolute guarantees).

By autumn 2008, with the Lehman collapse, markets were in a period of extreme volatility and there were largefalls in the value of the Plans' assets. The fall in equity values had the unintended consequence that (due to the shrinking value of the Plans' equity portfolios) the intended target of 44% bonds in the Main Plan was reached by autumn 2008.

The Plans' actuary, Mr Alexander, advised that the Plans had moved back into deficit, and there was discussionof bringing forward the next AVR (scheduled to be performed as at 31 December 2009) to 31 December 2008.

By the end of 2008, the asset allocation for the Main Plan was 52:48 equities/property versus bonds, and the IPlan was 92:8 equities/property versus bonds.

In terms of investment strategy and scheme funding, the announcement of Project Waltz in May 2009 caused theTrustee (in accordance with its professional advice):

Following the announcement of Project Waltz, there was extensive discussion between Holdings and the Trusteeabout investment strategy, in which the consistent position of the non-conflicted Trustee directors was that the Plans should be significantly de-risked, whereas Holdings' position was that the existing asset allocation should not be changed.

This disagreement about investment strategy can be seen in the discussions between Holdings and the Trusteeabout the AVR following the announcement of Project Waltz. In fact, the out-of-cycle AVR as at 31 December 2008 had to be abandoned as the Trustee was unable to agree the assumptions with Holdings in time. During the subsequent discussions about the AVR as at 31 December 2009, Holdings argued against de-risking in the shortterm, whereas the Trustee proposed a substantial de-risking.

Ultimately, the negotiations between Holdings and the Trustee went up to the statutory deadline. The finalagreement was, broadly, to de-risk on a straight line basis to target a portfolio equivalent to 20:80 equities/property versus bonds by 31 December 2016 in both Plans.

There is one other point to note. Although the Trustee was generally aware of NPPC, the Trustee's focus was oncash contributions and scheme funding, rather than on Net Period Pension Cost (an accounting concept under US GAAP: see paragraph 496 below).

In order to put the Project Waltz changes in their context, it is necessary to consider earlier changes to the DBPlans from 2004, in particular two sets of changes known respectively as Project Ocean (in 2004-2005) and Project Soto (in 2005-2006). I start with the actual changes to the relevant Trust Deeds and Rules turning to consider later the reasons for the changes and how they were presented by IBM and implemented. The Ocean changes

The Ocean changes were proposed by Holdings to the Trustee in October 2004, they were approved by theTrustee in October-December 2004 and they were given effect as from 6 April 2005 (or 6 July 2005 for DSL Plan members) by Deeds dated 9 December 2004 and 10 January 2005. The Ocean changes are set out in the table at Annex B to this judgment. It can be seen that Ocean involved, principally, changes to the employee contribution rates.

As part of the Ocean changes, IBM agreed to provide the Trustee with a parent company guarantee in respect of the funding obligations of Holdings (the Funding Agreement). I will come to how this came about in due course. The guarantee was provided by IBM WTC, in a Deed made between Holdings, IBM WTC and the Trustee dated 9 December 2004. Under the Funding Agreement, which was governed by English law, IBM WTC guaranteed certain of Holdings' funding obligations in respect of the cost of ongoing accrual, deficit repair contributions and

The Soto changes were proposed by Holdings to the Trustee during late 2005 and early 2006. They wereapproved by the Trustee on 19 January 2006 and they were given effect as from 6 July 2006 by Deeds dated 30 June 2006.

The principal feature of the Soto changes was that DB members (in both the Main Plan and the I Plan) weregiven an option exercisable by 30 June 2006:

Another feature was that IBM would inject a cash lump sum into the Main Plan and the I Plan by 31 March 2006 in order to extinguish the funding deficit and to fund the 2006 service costs. In the event, the figures were £544m for the Main Plan and £10.25m for the I Plan.

The affected DB members (about 6,000 in number) were invited to make the election by 30 June 2006 using anonline tool which gave them the option to remain in their DB Plan subject to the non-pensionability agreement or to transfer to the Enhanced M Plan. The tool explained that a member who failed to respond would be left in his DB Plan but, in respect of any future salary increases, would not be offered the non-pensionable 50% salary supplement. Of the affected DB members, 78% elected to remain in their DB section, 20% elected to transfer to the Enhanced M Plan and 2% did not respond.

The Soto changes are set out in more detail in Annex C to this judgment.The Enhanced M Plan

The members of the DB Plans who transferred to the Enhanced M Plan on 6 July 2006 became:

The terms governing Enhanced M Plan Membership and Enhanced Deferred Membership were inserted into theRules of the Main Plan and the I Plan by the Deeds of amendment dated 30 March 2006.

The principal rights of Enhanced M Plan Members under the Main Plan DC Rules are, in summary, as follows:

The principal rights of Enhanced M Plan Members as Enhanced Deferred Members of the DB sections of theMain Plan and of the I Plan are, in summary, as follows:

The creation of Enhanced Deferred status applied equally to the accrued DB benefits of DSL DB Plan memberswho elected to transfer to the Enhanced M Plan.

One complication to note is that, because Enhanced Deferred Members are still "Members" of the DB Plans,they had to be carved out of the Exclusion Notices, so that they were not excluded from the DB Plans along with the other active DB members. Thus the Exclusion Notices contain provisos exempting Enhanced Deferred Members from their scope.

As part of the Soto changes, it was agreed that Holdings would make additional contributions to eliminatethe deficit in the Plans. These contributions could have terminated the Funding Agreement (because of a clause terminating it if the Plans became fully funded), so it was agreed to amend the Funding Agreement to remove this difficulty. Accordingly, by a Deed of Amendment between Holdings, IBM WTC and the Trustee dated 30 March 2006, the Funding Agreement was modified on the following terms:

The Funding Agreement was amended again in 2007 as part of the discussions surrounding the triennial AVR asat 31 December 2006. It was agreed that the Funding Agreement (which was due to expire in March 2014) should be extended by 3 years and that IBM Corporation should replace IBM WTC as the guarantor.

Accordingly, by a Deed of Amendment and Novation between IBM WTC, IBM Corporation, Holdings and theTrustee dated 15 November 2007, the Funding Agreement was modified on the following terms:

It is against the background of the Ocean and Soto changes that Project Waltz falls to be considered. The dryexplanation which I have given of the result of Ocean and Soto hides an important (and very large and complicated) part of the story when it comes to assessing the impact of the Imperial duty in relation to Project Waltz. This will be seen when I turn to consider Issues 4 and 4B. The Project Waltz changes

I have dealt with these in outline already: see paragraphs 17ff above. The Trustee's written opening submissions contain a table of drafting amendments and the mechanics for implementing Project Waltz. I do not think I need to set them out.

I would nonetheless note one feature which is that "hybrid deferred" membership is conceptually the same asEnhanced Deferred membership created in 2006, so that the ex-DB members affected by Project Waltz will still be a type of "member" of the DB Plans in respect of their accrued DB benefits and governed by the same rules as those applicable to active members. Hybrid deferred membership is, however, less valuable than Enhanced Deferred Membership because the final salary link is broken as a result of the 2009 Non-Pensionability

With that background, I turn to Issues 1 to 3.

Resolution of Issue 1 requires consideration of (1) the powers to terminate accrual of benefits under the 1983 Trust Deed and Rules (2) the scope of the amendment power in the 1983 Trust Deed and Rules ("the 1983 Amendment Power") and (3) the true construction of the Exclusion Power in the Main Plan. Each of those items informs the others. Thus, if the pre-existing powers of termination were very wide, the restrictions on the 1983 Amendment Power would be less likely to be infringed than if the pre-existing powers were very narrow; and items (1) and (2) are likely to have an impact on item (3) since, if possible, the Exclusion Power should be given a construction which reflects a valid, rather than an invalid, exercise of the 1983 Amendment Power. Powers to terminate accrual of benefits under the 1983 Trust Deed and Rules

So far as item (1) is concerned, the powers to terminate accrual of benefits can be understood by reference toRule 2, in Part I of the 1983 Rules (containing definitions including "Member", "Eligible Employee" and "Pensionable Service"), Rule 3(A) in Part II (the principal eligibility provision) and Rule 3(D) (relating to cessation of membership). Those provisions are as follows:

The result of these definitions is, at least for present purposes, clear. Under Rule 3(A)(1), it was envisaged thateach Participating Employer would keep a register of permanent employees. Such employees would be Eligible Employees if they fulfilled the other conditions in Rule 3(A). A person would cease to be a Member on ceasing to be an Eligible Employee (eg because he or she has attained age 63 and thus ceases to fulfil the requirement of Rule 3(A)(2) ("he has not attained the age of 63 years")).

The definitions of "Eligible Employee" and "Service" contained within them powers for the Principal Employerto determine certain matters. Under the definition of "Eligible Employer" it was given the function of deciding whether a person was in Service and whether his name should be recorded in the register. Under the definition of "Service", it was given the function of determining the date of commencement of Service and the number of years to be counted as Service.

In my judgement, these functions were vested in the Principal Employer in order to deal with cases where theremight otherwise be doubt about whether a person is (i) in Service or (ii) a permanent employee or doubt about the date of commencement of the last period of continuous Service or the number of years of Service with, for instance, to be brought into account under paragraph (b) of the exceptions listed in the definition of "Service". These functions did not allow the Principal Employer to determine that a person who was in fact a "permanent employee" on any sensible meaning of those words to determine that, for the purposes of the Scheme, that person was not to be treated as a permanent employee. Nor did it allow it to determine that the service of such a permanent employer during a specified period or, perhaps, at all is not to be treated as within the definition of "Service".

Further, it is in my judgement implicit in the definitions of "Service" and "Eligible Employee" that the name of aperson who was in Service and who was in fact a permanent employee would appear in the Participating Employer's register of permanent employees. The Principal Employer could not, in my judgement, exclude from membership of the Scheme (or exclude an entitlement to become a member) a person who was in fact a permanent employee in Service by the simple expedient of declining to place his name on the register. If the Principal Employer was able to exclude a person from membership of the Scheme by, in effect, making a decision to that effect, I would expect to see such a power clearly expressed which it was not. Thus, if Rule 3(A)(1) had referred to a "a register of such of its participating employees as it determines to admit to the Scheme", the position would be clear. But the Rule did not make that provision or anything like it. Instead, it referred to a list of persons capable of identification by reference to a given criterion, namely status as a permanent employee. In cases of doubt about whether an individual was a permanent employee, the Principal Employer was given the function of resolving the doubt; but where there was no doubt, it had no function to perform. It would, in any case, be a remarkable situation if the Principal Employer had been able to dictate the names which appear on the register of permanent employees of each Participating Employer: that register and the names to be placed on it was surely a matter for the Participating Employer concerned, subject to resolution, by the Principal Employer, in cases of doubt.

I reach the conclusion which I have without reference to other provisions of the Scheme. Reference to thoseprovisions only goes to reinforce the conclusion. In particular, the Scheme contained entirely separate provisions allowing a Participating Employer to terminate accrual (either for all of its employees or for a specified category or categories) of employees, such provisions appearing in Part VII of the Rules relating to "Suspension or Termination of the Scheme". The relevant Rules are Rules 56 and 57. Rule 56 (so far as material) provided: "56. A Participating Employer's contributions

On its own, Rule 56 was a partial termination provision, leading to a partial winding up in accordance with Rule57 by the purchase of annuities, subject to the unilateral discretions on the part of the Trustee to provide for affected members' benefits by transfers (Rule 58(A)) or to continue the Main Plan as a closed fund (Rule 58(B)).

However, Rule 57 (so far as material) provided:

Subject to continuation of the Scheme as a closed scheme, Rule 57 provided, on the termination of the entireScheme, not only for members to be bought annuities in respect of their accrued benefits under the Scheme (see Rules 57(d)(1)-(5)) but also for any surplus to be applied in augmentation of members' benefits: see sub-Rule 57(d)(6), as follows:

Given that the power under Rule 56 is a unilateral power exercisable by each Participating Employer in respectof its own employees and given that Rule 57 (subject to the Trustee's discretions under Rule 58) makes detailed provision for how the Scheme is to be dealt with (by way of partial or total termination), it would be an odd result if a Participating Employer was entitled, in addition, to remove from the list of permanent employees the names of a whole category of employee or indeed of its entire workforce thus bringing about a cessation of accrual but making absolutely no provision at all about how the Scheme was to be administered in circumstances equivalent to partial or total termination under Rules 56 to 58.

I do not understand Mr Simmonds to argue for a different result. Whilst drawing attention to the apparentlyconclusive nature of the decision of the Principal Employer concerning the names to be recorded in the register, and suggesting that there are superficial similarities between those provisions and the Exclusion Power, he acknowledges IBM's acceptance that the scope of the Exclusion Power exceeds that of the eligibility provisions in the 1983 Trust Deed and Rules.

I have, nonetheless, gone into the point myself and reached a conclusion in relation to it because it is, I consider,relevant to understand precisely how the provisions of the 1983 Trust Deed and Rules differ in their substantive effect from the Exclusion Power and why Mr Simmonds is right not to advance the argument which he has referred to.

I turn now to the scope of the 1983 Amendment Power and the restrictions placed on it. The RBs' position is thatthe Exclusion Power can be valid only if, on its true construction, it has a scope and/or purpose that is consistent with the 1983 Trust Deed and Rules and the fetters on the 1983 Amendment Power. IBM's position is that the Exclusion Power, construed as IBM would construe it, is consistent in that way, so that the introduction of the

The 1983 Amendment Power, which is to be found at Clause 4 of the 1983 Trust Deed and Rules, is set out atparagraph 78 above.

It is common ground that provisos (a) and (b) to the 1983 Amendment Power were not engaged by theintroduction of the Exclusion Power. Proviso (f), which concerns the effect of amendments pending the formal documentation of the same, is similarly not in issue. Furthermore, it is no part of the RBs' case that the introduction of the Exclusion Power was defective for want of compliance with the formalities required by the 1983 Amendment Power.

It was pursuant to this power of amendment that the Exclusion Power in the Main Plan was introduced, it firstappearing, as I have mentioned, in the 1990 Trust Deed and Rules. The 1990 Trust Deed and Rules made no provision for the consequences of an exercise of the Exclusion Power (such as the triggering of a partial or total winding-up); accordingly, the Main Plan would simply continue to run according to its terms. The Exclusion Power would, on IBM's approach, have represented a significant change from the pre-existing provisions since Holdings would be able to terminate future accrual and break the final salary link in respect of pensions earned by service to date; and it would be able to do so without triggering a winding-up. Previously, termination of contributions would have involved a total or partial winding up, on which event the Members concerned would have been entitled to augmentation of benefits out of surplus (if there was any) after the making of provision for benefits. Such provision would have been effected by the purchase of annuities in accordance with the winding-

In my judgement in the Rectification Action, I discussed the circumstances in which the 1990 Trust Deed andRules were drafted and executed: see in particular at [420]ff. Mr Stallworthy relies on my finding, at [423], that the instructions given to Nabarro (who carried out the drafting) were:

He also relies on [428], where I was dealing with a report prepared by Mr Quarrell (of Nabarro) for the Trustee,during the course of which I stated:

I said that in relation to "early retirement" and "flexible retirement" but it applies equally to any change to thesubstance of the termination provisions and the balance of power between IBM on the one hand and the Trustee on the other.

In these circumstances, Mr Stallworthy submits that "one would not expect to find (unannounced) a significantnew power introduced amongst the eligibility provisions of the 1990 Trust Deed & Rules". That point demonstrates the close connection between Issue 1 and Issue 2. One can only decide whether the Exclusion Clause was validly introduced if one knows what it means: but one can only decide what it means once one knows the scope of the fetters on the 1983 Power of Amendment because one should, if possible, construe the Exclusion Power in a way which is consistent with those fetters. Nonetheless, I propose to address Issue 1 and Issue 2 separately, as did both Mr Stallworthy and Mr Simmonds.

It is, however, important to note that I said what I did in the Rectification Action in the context of a rectificationclaim. It was a material factor in ascertaining the intention of the members of the board of the Trustee in their consideration of what became the 1990 Trust Deed and Rules. I very much doubt that evidence of Nabarro's instructions (or of the fact that the 1990 Trust Deed and Rules was intended to be a consolidation in plain English, with substantive amendments being made only in relation to matters on which Nabarro had instructions) is admissible on the question of construction of the 1990 Trust Deed and Rules; and this is so in spite of the current approach to construction under which context appears to be an almost overarching consideration: see in particular the speech of Lord Clarke in Rainy Sky SA v Kookmin Bank [2011] UKSC 50.

I note for completeness at this stage that Mr Stallworthy submits, correctly in my view, that the directors of theTrustee never asked for any new power to close the Main Plan to accrual to be introduced into the 1990 Trust Deed and Rules; and there was no suggestion from any of the Trustee's advisers (eg Nabarro and the Actuary) that any such new power was being introduced. They had no notion that they might be altering materially the balance of powers within the Main Plan by introducing a new power to close the plan to accrual, and thereby break final salary linkage, without triggering a winding up. This is evident from the contemporaneous documents relating to the production of the 1990 Trust Deed and Rules including, in particular, Mr Quarrell's report which I have referred to above. There is no subsequent material which suggests that any of the directors ever thought that that was what had happened.

Mr Stallworthy goes on to submit that if IBM's case is correct, then the Trustee was acting under a fundamentalmistake as to the legal effect of the 1990 Trust Deed and Rules, because the Trustee did not appreciate that such a significant new power was being introduced. However, he accepted that, on the state of the law current at the date of the hearing before me, as stated in Pitt v Holt and Futter v Futter [2011] EWCA Civ 197, the Trustee's decision to introduce the Exclusion Power by amendment would not be void or voidable under the principles previously often referred to as "the Rule in Hastings-Bass". Mr Stallworthy reserved his position in the light of anything which the Supreme Court might decide or have to say on the appeals then before it in those cases. The Supreme Court has now given its decision: see [2013] UKSC 26. Mr Stallworthy has not sought to make any further submissions. The Hasting-Bass point cannot now avail him.

The issue of whether the introduction of the Exclusion Power is vitiated to any extent by mistake was not argued,possibly because the distinction drawn by Millett J in Gibbon v Mitchell [1990] 1 WLR 1304 between effect and consequences was seen as an insuperable obstacle. The law has now been clarified – and to some extent it has moved on – in the light of the decision of the Supreme Court. Again Mr Stallworthy had reserved the RBs' position in the light of anything which the Supreme Court might say but he has not sought to add to his submissions. I therefore say no more about this issue. The Actuary's opinion

Each of the provisos (c) to (e) to the 1983 Amendment Power commences with the words "no such alteration ormodification shall be made which in the opinion of the Actuary shall operate …..". There appeared from their skeleton arguments and their written closing submission, to be a gulf between Mr Stallworthy and Mr

Mr Simmonds had originally taken the line that the absence of negative opinion meant that the amendment wasvalid. Having read and heard Mr Stallworthy's closing submissions, he did not pursue any of the arguments which he had canvassed about the effect of the absence of a negative actuarial opinion. The RBs' case

Mr Stallworthy identifies two elements of the Exclusion Power which he submits breach the fetters found in theprovisos:

The abandonment by IBM of the arguments based on the absence of a negative opinion from the Actuary greatlyreduces the area of contention based on the break in the final salary link. The difference now is this. The RBs contend that the exercise of the Exclusion Power in a way which broke that link is wholly invalid. IBM's contention is that the amendment was valid pro tanto and should be given effect to in relation to service after the date of the amendment whilst preserving the final salary link in relation to service prior to the amendment. But the acceptance of that result by IBM relates only to the exercise of the Exclusion Power itself and will, in practice, have little impact because, on IBM's case, the 2009 Non-Pensionability Agreements will have restricted the part of pay which is pensionable to a figure less than the definition of Final Pensionable Earnings would suggest. The loss of opportunity to benefit from augmentation remains, however, an area of contention.

Mr Stallworthy begins with proviso (e) which is the most important proviso for present purposes. This protectednon-pensioner Members (ie active or deferred members) as regards "the aggregate value of the retirement benefits payable under the Scheme …in respect of contributions already received by the Trustee". In contrast with provisos (c) and (d), there was no test of substantiality – any reduction in the aggregate value of the retirement benefits payable in respect of contributions already received by the Trustee would engage proviso (e).

Mr Stallworthy's argument is as follows:

In support of his submissions, Mr Stallworthy has referred to a number of cases in this jurisdiction and othercommon law jurisdictions, where the courts have had to grapple with the effects of various fetters on powers of amendment. He prepared a short analysis of the cases relied on. I have taken that analysis into account in reaching my conclusions.

Mr Stallworthy's argument is that, assuming the Exclusion Power to have the scope for which IBM contends, theActuary should (and if he had appreciated that scope, would) have formed and expressed the opinion that provisos (e) and (d), and perhaps even proviso (c), were infringed. He says "would" because such an opinion was given in relation to the amendments under Project Soto implementing partial non-pensionability of further pay rises. Those amendments were only rendered valid by the consent of the affected members. Those are bold submissions in the light of the absence of any expert evidence directed at the question whether the fetters were infringed, not by the Soto changes but by the making of the 1990 Trust Deed and Rules.

The point is also made that these conclusions are not affected by the fact that the amendment introduced a powerfor future use, rather than immediately itself terminating accrual. Mr Stallworthy relies on Re Courage Group's Pension Schemes [1987] 1 WLR 495 at 513C-F which demonstrates that Millett J was indeed concerned about the future exercise of the powers inserted by amendment; and in Bradbury v BBC [2012] EWHC 1369, [2012] PLR 283 at [67] I reached a similar conclusion (albeit obiter) in relation to the particular power concerned in that case. As Mr Stallworthy observes, an amending party cannot achieve in two steps what he cannot achieve in one (eg by purporting to delete a fetter to an amendment power and subsequently making a second amendment which would have been precluded by that fetter): see Air Jamaica v Charlton [1999] 1 WLR 1399 at 1411G and HR Trustees Ltd v German & IMG [2009] EWHC 2785, [2010] PLR 23 at [115]-[125], especially [123].

In relation to proviso (d), Mr Stallworthy submits that this clearly protects contingent benefits, includingdiscretionary benefits, referring as it does not simply to "benefits" but to "interests under the Scheme". He is clearly correct in saying what he does in relation to contingent benefits. Whether he is right that it includes discretionary benefits (such as augmentation in a winding-up) is something I will come to.

He finds support in the decision of the Court of Appeal in New Zealand in UEB Industries Ltd v W S Brabant [1992] 1 NZLR 294, [1991] PLR 109 where the amendment power in clause 10 precluded any amendment which would "reduce or adversely affect that Member's interest in the Fund at the date of alteration". The issue was whether it was possible to amend the winding up clause to provide for any surplus to be paid to the employers rather than applied to augment benefits. The Court unanimously held that this fetter precluded an amendment, because the possibility of augmentation constituted part of a "Member's interest in the Fund at the date of alteration". A pithy passage from the judgment of Cooke P at [34] is worth quoting, It reads as follows:

Accordingly Mr Stallworthy submits that any introduction of a new power by which Holdings could unilaterallyclose the Main Plan to all accrual, breaking the final salary link, should (and would) have been invalidated by an opinion from the Actuary under proviso (d).

In relation to proviso (c), Mr Stallworthy submits that the Actuary might even have been of the opinion that theintroduction of the Exclusion Power engaged, and should be invalidated by, proviso (c), on the basis that such a power unilaterally to close the Main Plan to all accrual without triggering a winding up operated "substantially to prejudice the pension payable" to pensioners and the contingent beneficiaries of Member pensioners because it permitted Holdings to avoid augmentations from surplus on winding up. IBM's case

Mr Simmonds breaks the issue of the validity of the Exclusion Power down into three questions.

These seem to me to be the right questions to ask although, in answering the third question, consideration needsto be given to the role of the Actuary and the matters of which he needs to be satisfied. Mr Simmonds sees the issue between him and Mr Stallworthy as resulting simply from a different approach to the way in which an amendment which admittedly breaches the fetters should be treated – whether it is wholly invalid or only pro tanto invalid. But that does not do justice to Mr Stallworthy's appeal to the role of the Actuary: on the basis that there is a prima facie breach of the fetters, the Actuary could not have done anything other than provide a negative opinion, in which case the argument is that the purported exercise of the Amending Power must be wholly void.

In relation to his first question, Mr Simmonds' focus is on the point concerning access to surplus. He accepts thatthe final salary linkage is something that would cease to apply on an exercise of the Exclusion Power as a matter of construction of the Main Plan and the I Plan documentation (although it should be remembered that the final salary linkage point is not relevant to persons who are already pensioners or deferred pensioners at the time of the exercise of the Exclusion Power). He accepts that the prospective right to a pension based on final salary is part of the interest of a Member. It follows that the elimination of this interest would prejudice the interests of a person who is an active member at the time of the exercise of the Exclusion Clause and would thus be a breach of proviso (d) to the extent that it is "in respect of contributions received by the Trustee prior to 1 January 1973". He does not now seek to argue that the exercise of the Exclusion Power would not reduce the aggregate value of the retirement benefits payable under the Scheme to a person who is an active member at the time of the exercise of the Exclusion Power. It follows that the elimination of the final salary link would be a breach of proviso (e) to the extent that the aggregate value is "in respect of contributions received by the Trustee".

So far as concerns augmentation of benefits Mr Simmonds acknowledges that, under the 1983 Trust Deed andRules there was no free-standing power to terminate contributions in respect of a Member; termination of contributions would lead to a partial winding-up (albeit subject to the power (whatever it actually means) for the Trustee under Rule 58(B) to continue the Main Plan as a closed fund) and that, on winding-up, there was a requirement to use surplus to augment benefits. But he does not accept that this entitlement is within the scope of the protection afforded by the provisos to Clause 4; nor does he accept, even if that is wrong as a matter of construction, that the introduction of the Exclusion Clause did infringe any of those provisos.

Mr Simmonds says that proviso (c) is concerned only with pensioners and dependants of pensioners. He isclearly right about that.

Proviso (d) is concerned with active members and deferred pensioners. The temporal limitation demonstrates thatthis proviso is directed only at persons who were Members prior to 1 January 1973. The protection it affords is only in relation to "interests…..in respect of contributions received…. prior to 1 January 1973". The proviso is curiously worded (although I understand quite a common form) and is open to different interpretations.

Proviso (e) is also concerned with active members and deferred pensioners. Here there is a different protectionand a different temporal limitation. The protection relates to the aggregate value of the retirement benefits payable. And that value is not ascertained by reference to "interests… in respect of contributions received prior to 1 January 1973" but rather by reference to "retirement benefits payable…. in respect of contributions already received" (ie received by the date of the amendment).

Mr Simmonds addressed the argument made by Mr Stallworthy, relying on UED, that "interests" in proviso (d) include prospective interests in surplus on winding-up. That case was, on its facts, far away from those in the present case, and the material restrictions on the amendment power very different. The power of amendment contained a proviso to the effect that no amendment could be made in relation to a member if it "would reduce or adversely affect that member's interest in the fund". The actual amendment made was to replace an obligation to apply surplus in winding-up in the augmentation of benefits with a requirement to pay it to the employer. The Judge held this to be an impermissible derogation from the members' rights: the members did not have a mere expectancy nor even a mere right to be considered for any discretionary allocation.

And so in the present case, Mr Stallworthy argues that the members have an interest in surplus in the same waythat the members in UEB had an interest in surplus.

There are, however, two important and material distinctions to be drawn in the present case. The first is that eachof the provisos has to be construed as part of the scheme of protection which is found in Clause 4 taken as a whole. Mr Simmonds submits that when the provisos are taken as a whole – particularly provisos (c), (d) and (e), it can be seen that the "interests" in proviso (d) cannot be given the wider interpretation given in UEB. The second distinction, which may simply be a particular example of the first, is that, in the present case, provisos (d) and (e) both contain a temporal limitation which fits uncomfortably with the idea that an opportunity to share

Those are important distinctions. As to the first distinction, the first point to note is that proviso (c) only protectsthe pension payable to a person who is entitled to a pension at the effective date of the alteration. I agree with Mr Simmonds when he says that the protection relates only to the pension payable at the date of the alteration. It does not extend to any increase in the pension which might be made if the scheme were to go into winding-up and if it were then in surplus.

The second point to note is that proviso (e) does not use the word "interests" at all; rather it relates to the"retirement benefits payable under the Scheme to any Member" and what is protected is the aggregate value of all such benefits. The idea behind that, it seems to me, is that the Actuary will be able to identify the benefits which the Scheme provides and will be able to place a value on them making appropriate assumptions as he would do in an actuarial valuation of the Scheme as a whole. It is the aggregate value which has to be maintained, but that does not mean that some benefits may not be reduced provided that others are increased. That does not fit at all comfortably with that idea that the prospect of sharing in surplus, if there is one, on a winding-up, if there is one, is somehow part of the retirement benefits which have to be valued under the Scheme.

The third point is that both provisos (d) and (e) each treat active members and deferred members in the sameway. If Mr Stallworthy is correct in his construction of provisos (d) and (e), the result is that the opportunity for a deferred pensioner to share in surplus is protected as an "interest" under proviso (d) or as part of "the retirement benefits payable" under proviso (e). But, for reasons already given, proviso (c) does not extend to the protection of that opportunity to pensioners. In other words, deferred pensioners obtain a protection which pensioners do not. That would be a very surprising result.

In my judgement, Mr Simmonds' approach is more coherent and leads to a more sensible result. The protectionafforded to pensioners relates to their actual benefits in payment and extends to contingent benefits in respect of pensioners. Active members and deferred members are treated together because their benefits are not in payment and the eventual amounts payable are not known and indeed a particular benefit (eg a pension) may not even come into payment, for instance in the event of death before the pension comes into payment. Instead protection is afforded under proviso (e) in respect of the retirement benefits which will become payable under the scheme in different events (eg surviving to normal pension age or dying prematurely); those benefits can be ascertained and a value can be placed on them. It is not clear why a different formulation is found in proviso (d) nor why the majority can bind the minority by giving consent to an amendment in contrast with the position in relation to proviso (e). Proviso (d) would appear to afford protection in relation to each interest under the Scheme separately, whereas proviso (e) is expressly directed at the aggregate value of all retirement benefits.

The second distinction is that both provisos (d) and (e) contain a temporal limit to the protection, either 1 January 1973 (proviso (d)) or the date of the amendment (proviso (e)). Mr Stallworthy has drawn attention to the fact that the protection is given in relation to interests or benefits payable "in respect of contributions received by 1 January 1973" (or "already received"). The protection is not expressed to be given by reference to benefits earned by service up to those times. I find the actual wording of the provisos in that regard curious. I suspect that it was a form of words adopted many years ago which has found its way into common usage. It seems to me, however, that it can be referring only to interests or benefits arising under the Scheme in respect of periods for which the contributions (by employer and member) have been paid. In practical terms, there should be little, if any, difference in the result from a formulation based on benefits earned by service up to 1 January 1973 (or the date of the amendment). But I suppose that if contributions are paid in advance so that the Trustee has received contributions in respect of benefits accruing for a period after the date or payment, there could be a difference. But that is fine distinction and I do not consider that it has a material effect on the argument.

The protection thus afforded can therefore be seen to be one which relates to interests or benefits earned byreference to a period of service. It is to my mind a more natural interpretation of both provisos to see that protection as excluding something as uncertain as an augmentation on a winding-up which may never happen out of a surplus which may or may not exist even if it does. I do not say that such a protection could not be

I would therefore answer Mr Simmonds' first question this way. The interests and the retirement benefitsprotected by the fetters in Clause 4 of the 1983 Trust Deed and Rules include the final salary link but do not include any opportunity to share in surplus on a winding-up. The second question

Mr Simmonds' second question is whether the Exclusion Power prima facie prejudices the rights that are protected by the fetters in Clause 4. I proceed, in dealing with this question, and the third question on the footing that the Exclusion Power, on its true construction, does have the scope for which IBM contends (a matter to be decided under Issue 2). The assumptions, therefore, are (i) that IBM can serve a notice in relation to the entire membership of the Main Plan and the entire membership of the I Plan save for excepted individuals and (ii) that there are no implied restrictions on the apparent width of the power read literally. By way of reminder, I set out that power, as it appears in the Main Plan Definitive Trust Deed, again:

Mr Simmonds accepts that the answer to his second question is "Yes" so far as concerns final salary linkage.That is the clear effect of the Rules which apply when a person ceases to be a member.

But he maintains that the answer is "No" in relation to access to surplus. The question only arises if his (and my) answer to the first question is wrong. But the point has been fully argued and I propose to deal with it. The reason for his answer is that the right is only to participate in surplus if the employer decides to put the scheme into winding-up. A member, of course, had no right under the 1983 Trust Deed and Rules to put the scheme into winding-up, nor could the members collectively, or the Trustee do so. But if the scheme does go into windingup, the opportunity to access surplus remains precisely what it always was.

In that sense, the present case is not merely different but wholly distinguishable from UEB where the amendment purported to divert the surplus from the members to the employer. I consider that Mr Simmonds is correct to draw that distinction and I gain no assistance from UEB in relation to access to surplus.

It is, moreover, important to recognise that none of the provisos relates to benefits earned in respect of futureservice and, indeed, proviso (d) does not relate to service after 1 January 1973. Quite clearly, future service benefits could be adversely affected: there could be no objection, as a matter of construction of the 1983 Trust Deed and Rules, to IBM and the Trustee agreeing to reduce the accrual rate for future service. Whether the Trustee could properly agree to such a reduction would, of course, depend on the particular facts of the case at the time of the amendment. Such an amendment would be likely to have an impact on how surplus would in fact be distributed in the event of a subsequent winding-up but it cannot seriously be suggested that it would infringe the fetters on the power of amendment.

Once it is accepted, as I think it must be, that an amendment to reduce future accrual of benefits would notbreach any of the fetters on the power of amendment, I see no reason in principle why IBM and the Trustee should not be able to agree to an amendment which does not bring about an immediate reduction in the accrual rate but which introduces a new power for IBM unilaterally to do so. The rate of future service accrual is clearly not an entrenched provision so that an amendment which adversely affects future service accrual is not to be ruled out as a possibility; such an amendment does not raise the sorts of issue which arise when an amendment is sought to be made to an amendment provision which is hedged about with restrictions so as to remove those restrictions, with the amended power subsequently being used to do something which would have been prevented under the original amendment power.

Moreover, I do not see any reason in principle why the reduction in benefits under such an amendment shouldnot, as a matter of construction of the provisions of the 1983 Trust Deed and Rules, be reduced to nil. The structure of those provisions did not, it is true, include a power for IBM unilaterally to reduce future accrual to nil or indeed at all, although IBM did have power to terminate contributions (and thus accrual of benefits) either in relation to a specified category of member or in relation to a Participating Employer, in the latter case bringing about a partial winding-up. I see no reason to think that that part of the structure was a necessary and enduring part of the architecture. I do not consider that the provisos to the power of amendment preclude a new way of dealing with what is to happen in the event of the termination of contributions. Thus, had an amendment been agreed by the Trustee (again, the facts would have to justify the Trustee acting in this way) under which the default position was changed so that, instead of termination of contributions resulting in a total winding-up with an option for the Trustee to continue the scheme as a closed scheme, the scheme was to continue as a closed scheme with the Trustee having an option to place the scheme into winding-up, I do not see why this should be seen as infringing the fetters on the amending power. The amendment will give rise to a change in the structure but will not compromise the essential architecture of the scheme.

Mr Simmonds' third question is what, as a matter of law, is the impact of the fetters on the 1983 AmendmentPower on the Exclusion Power in terms of the validity of the rule and whether it is wholly invalid (as the RBs contend) or can be saved in part. I refer to paragraph 157 above for the basis on which I deal with this question.

In the light of my conclusions in relation to precisely what it is that the fetters protect, the only issue on thisquestion relates to the final salary linkage, although I will say a little about access to surplus in case I am wrong in my conclusions in relation to that.

There is one point of construction which I wish to dispose of at the outset. The structure of the 1983

Some of the cases to which I have been referred, where defective amendments have been given partial validity,contain wording which has a slightly different focus from the present case. In those cases, the restriction is along the lines "no such alteration shall operate" in contrast with the present case where the restriction is "no such alteration or modification shall be made which shall operate". In the former case, it could be said that what is restricted is the operation of the amendment once effected and that a failure expressly to spell out in the amendment the relevant restriction does not preclude an exercise of the amended provision in a way which would not operate in the forbidden manner. In contrast, in the present case, it can be said that the restriction is on the alteration itself (and not merely on the operation of the amended provision) so that a failure to spell out the restriction in the amendment is fatal.

In my judgement, this is a distinction without a difference. In applying the principle (when the facts justify it) that an apparently excessive exercise of an amending power is valid to the extent that it is not excessive, it would be wrong to allow this narrow difference in wording to defeat the principle of validation (subject to the constraints within which that principle can be applied, to which I come in a moment). Accordingly, I do not consider that the presence of the words "shall be made" affects the outcome.

This is a convenient place to mention one aspect of Millett J's decision in Re Courage Group's Pension Schemes [1987] 1 WLR 495. That case concerned a prospective exercise of a power. The Judge held that the power could not properly be exercised in the way proposed. He did not address – the question did not arise – what if any effect would have been given to the amendment if it had already been made. It does not follow from the fact that, prospectively, a power of amendment cannot properly be exercised in a particular way that, if it has been exercised in that way, the exercise is wholly invalid.

Similarly, in the present case, it would no doubt be the case that, had the Actuary been asked for his opinionknowing that the final salary link had to be preserved but was not in fact preserved, he would have expressed a negative opinion in relation to the introduction of the Exclusion Power. But it does not follow from that conclusion that the Exclusion Power was wholly invalid.

In support of his submission that the court can and should give partial effect to the Exclusion Power, MrSimmonds relies on three authorities: the decisions of Neuberger J (as he now is not) in Bestrustees plc v Stuart [2001] PLR 283, of Lightman J in Betafence Ltd v Veys [2006] PLR 137 and of Arnold J in HR Trustees Ltd v German & IMG [2009] EWHC 2785, [2010] PLR 23.

In Bestrustees, Neuberger J felt able to treat an amendment as valid insofar as it was prospective in effect but invalid insofar as it infringed a proviso protecting accrued rights. The amendment related to the definition of normal retirement age and was intended to bring about equalisation between men and women. The amendment was valid so far as it concerned future service but not past service.

In Betafence, a proviso precluded amendments prejudicially affecting benefits secured up to the date of amendment. The amendment which had been made introduced a requirement for employer consent in relation to early retirement where none had existed before. Instead of striking down the offending amendment altogether, the Judge upheld it to the extent that the proviso was not infringed, holding at [69] that the amendment "must be construed as having effect subject to the overriding limitation on the power of amendment contained in the proviso" although, as the Judge observed "All that is required is that the distinction between what is and is not objectionable is clear and that the meaning and application of what is unobjectionable is clear".

In HR Trustees, the constraint on the amending power precluded amendment which had the effect of "reducing the value of benefits secured by contributions already made". Arnold J held that this protected the value of members' accrued rights calculated by reference to their pensionable service at the date of the amendment and their final pensionable pay. An amendment to convert these benefits to money-purchase benefits was permissible "but only subject to an underpin preserving the future monetary value of the proportion of Final Pensionable Pay which the member has accrued in respect of pre-amendment service".

Mr Stallworthy argued that those three cases dealt with amendments which were not concerned with theintroduction of new powers exercisable in the future, but instead dealt with specific situations at which a onceand-for-all amendment was directed. I have already addressed this argument in paragraph 163 above. My conclusion is that it would have been possible to create, in 1990, an amendment which then terminated future accrual (without triggering a winding-up) provided that it had been made subject to a proviso protecting the final salary link. Mr Simmonds asks rhetorically "Why should it not be possible to create a power for IBM to do just that at a later date?". My answer is that there is no reason at all.

But that is jumping ahead, and I shall mention two other cases relied on by Mr Stallworthy in this context whichare Air Jamaica and IMG. The relevant passage appearing in the speech of Lord Millett in Air Jamaica is this, at

The question was whether one amendment power could be replaced by another, the exercise of which wouldbreach the fetters on the original amending power. The real point, it seems to me, is that the original amendment power was effectively entrenched as an essential element of the architecture of the scheme there in question.

This was the approach adopted by Arnold J in IMG in the passage relied on by Mr Stallworthy at [123]:

So there we see Arnold J applying the principle that you cannot get round a prohibition on doing somethingdirectly, in one step, by doing it indirectly, in two steps. That principle must not be pushed too far. As Millett J recognised in Courage, a series of perfectly valid amendments over a period of time may, at least where not all planned as a single scheme of amendments, result in an end-point which it would not have been possible to reach by a single amendment at the beginning: see his reference to Thellusson v Viscount Valentia [1907] 2 Ch. 1 and the Hurlingham Club at p 506. That is not the present case, however, because the question is whether the first (and only relevant) amendment, the introduction of the Exclusion Power, was valid. It does not seem to me that the proposition that you cannot do in two steps what you cannot do in one has any scope for application. An amendment in 1990 to terminate future accrual would have been valid provided that the final salary link had been preserved. So too the introduction of the Exclusion Clause would have been valid if it had contained an express override preserving that link.

The limitation which Mr Simmonds submits should be applied is similar to the one implied by Arnold J in IMG. Although the drafting of such a limitation is not entirely straightforward, the concept is, I think clear: namely that a Member concerned should be entitled, if it produces a better result for him than statutory revaluation of his leaving service benefit, to a pension based on his period of service to the date of the exercise of the Exclusion Power and on his salary at the date when he actually ceases to be an employee.

I agree with Mr Simmonds' submissions, and, for the reasons which he gives, I reject Mr Stallworthy's argumentthat the Exclusion Power is wholly void.

In practical terms, Mr Simmonds submits that such a limitation would be of no practical effect because of the2009 Non-Pensionability Agreements. Under those agreements, each affected Member agreed to limit the amount of his pay which would be pensionable in respect of his past service. If these agreements are valid, the final salary link will have been broken by agreement. Whether those agreements are in fact valid is a different question, but the answer to that is not relevant to the issues now under consideration. I only comment here that the answer to that question may be influenced by the fact that the Members concerned may well have entered into the agreements on the basis that the Exclusion Power was valid and that the final salary link was broken.

If I am wrong about access to surplus being outside the fetters on the exercise of the 1983 Amendment Power,then it is necessary to analyse what precisely it is that had to be preserved. My own view is that it would be the opportunity to access surplus on a winding-up if and when that took place on the basis of benefits accrued to the date of exercise of the Exclusion Power. That is precisely what is in fact preserved so long as the proviso which I have discussed in relation to final salary linkage is to be implied into the Exclusion Power.

The alternative view is that the only way in which the opportunity to access surplus can be properly preserved isto permit that to be done at the time of the exercise of the Exclusion Power at least in a case where that exercise relates to all active Members. This is on the footing that such a comprehensive exercise of the Exclusion Power is a de facto termination of contributions which would have triggered a winding-up under the unamended 1983 Trust Deed and Rules.

But in that case, Mr Stallworthy is met with what I think is an unanswerable point made by Mr Simmonds. It isthat, as a matter of fact, there was no relevant surplus in either 1990, when the amendment was made, or in 2011, when the exercise of the Exclusion Power took effect. The winding-up rule provided for the purchases of "Paidup Policies securing benefits", that is to say the purchase of annuities and deferred annuities, for pensioners, deferred pensioners and active members (and their dependants): see Rule 57(d)(1) to (5). Any remaining surplus

I have already noted that this Issue and Issue 1 are interdependent. But it is Issue 2 which raises more directlythe proper meaning of the Exclusion Power. The discussion of Issue 1 has really side-stepped that question and has focused on the position which would obtain if, as a matter of construction, it apparently allowed IBM to terminate the membership of all active members of the Main Plan or all of the active members of the I Plan save for the excepted individuals (subject to the discrete point about Postponed Retirees to which I will come). Principles of construction

The principles of construction applicable to pension scheme documentation are now well established. I do notneed to say a great deal about them. The classic statement remains that of Millett J in Re Courage Group's Pension Schemes [1987] 1 WLR 495 at 505:

I considered the principles and re-stated them in my judgement in PNPF Trust Company Limited v Taylor [2010] EWHC 1573 [2010] PLR 261 at [127] to [145]. I see no reason to depart from what I said there and I see no merit in repeating it. For present purposes, it is necessary only to mention two aspects.

The first is to re-iterate that, under the process of construction, apparently wide words may be "read down" inone of three ways:

The second aspect is to draw attention to what Arden LJ said in [29] of her judgment in British Airways Pension Trustees Ltd v British Airways plc [2002] EWCA Civ 672, [2002] PLR 247 which I cited in [129] of my judgment in PNPF, where she noted that pension schemes are often subject to considerable amendment over time. The general principle is that each new provision should be considered against the circumstances prevailing at the date when it was adopted rather than as at the date of the original trust deed. Likewise, the meaning of a clause in a scheme must be ascertained by examining the deed as it stood at the time the clause was first introduced.

There is a preliminary point which needs to be made. Issue 2 as formulated relates to both the Main Plan and theI Plan. Both parties agree that the same answer should be given to Issue 2 in the case of both Plans. But they say so for very different reasons. Mr Stallworthy starts with his interpretation of the Exclusion Power in the Main Plan and says that the Exclusion Power in the I Plan should be given the same restrictive interpretation. Mr Simmonds originally submitted that the Exclusion Power in the Main Plan was entirely valid; it was only in his oral closing submissions that he realistically accepted, having heard Mr Stallworthy's submissions, that there had to be an implied limitation on the Main Plan Exclusion Power to reflect the fetters on the 1983 Amendment Power. If his original submission had been correct, then there would clearly have been no reason to apply a different interpretation to the Exclusion Clause in the I Plan. The concession which he made in relation to the Main Plan does not need to be made in relation to the I Plan because the I Plan contained, from its inception, the Exclusion Power. It is still open to him to argue that the Exclusion Power in the I Plan has the wide meaning for which he contends and that the meaning of the Exclusion Power in the Main Plan is precisely the same but, in order to reflect the fetters, is partially invalid. Partial invalidity does not affect the meaning of the provision: indeed, it only makes sense to speak of invalidity in the context of a provision which, according to its true construction, infringes the fetters.

Having referred to what Arden LJ said in British Airways at [29] to [30] quoted by me in PNPF at [129], Mr Stallworthy submits that the question is what a reasonable person having all the background knowledge available to the parties would have understood the Exclusion Power in the Main Plan to mean. In the 1990 Trust Deed and Rules, the power was to be found at Rule 6 Schedule C and is in the same form as it is now to be found at Clause 4 of the Main Plan Definitive Deed.

Mr Stallworthy's submission is that Rule 6 is unlikely to have been intended or understood to be a unilateraltermination power to close the Main Plan to future accrual given:

I will come to each of those in more detail later.

Mr Simmonds starts from a different place. The central question, he suggests, is this: Can the Exclusion Powerwhich gives power to Holdings to direct that "any specified person or class of persons shall….cease to be a Member or Members" be exercised in respect of the whole class of Members or, in the case of the I Plan, with the exception of certain specifically excluded Members? He says that it is for the RBs to demonstrate why the whole of the active membership does not constitute a "specified… class of person".

The well-known principles of construction which I have referred to show, of course, that words must beconstrued in their context. A particular word can have different meanings and which meaning is to be ascribed to it depends on its context. Each of those meanings might be described as an "ordinary" meaning of the word; but in context, it might be necessary to give the word a meaning different from each of those ordinary meanings. Similarly, a phrase might have different "ordinary" meanings and the context will disclose which of those ordinary meanings is appropriate; and, again, it may be necessary to depart from each of those ordinary meanings to give the phrase the meaning which it really has in the context in which it is used. In construing a

Thus Mr Simmonds' starting point is that an ordinary meaning of the words "any specified person or class ofpersons", indeed the ordinary meaning of those words, includes the whole of the active membership of the Main

According to him, the fact is that there is no rational or logical point at which to draw the dividing line. This inturn is indicative of the fact that, when given their natural and ordinary meaning, there is nothing in the words of the Exclusion Power that requires one to draw a distinction between a class consisting of the entire membership and a class consisting of a particular cohort or sub-set of that membership. Mr Stallworthy answers that rhetorical question by saying that there is such a dividing line and that it is to be found where the 1983 Trust Deed and Rules draws it.

Before coming to the arguments in more detail, it is to be noted that Mr Simmonds' approach as revealed in thatshort summary of his position proceeds on the basis that it must be accepted that Holdings had the power to terminate future accrual in respect of any given person. I am not sure that Mr Stallworthy would accept that since the old Rule 56 only permitted termination in respect of "persons in a specified category or specified categories". But Mr Simmonds' questions remain pertinent because, both under the old Rules and under the 1990 Trust Deed and Rules, it is possible to terminate contributions in relation to a specified category or class so he could replace reference to an individual with reference to an identified small class (eg employees in a particular location with very few employees).

The scheme of Mr Simmonds' argument has the merit of starting with the ordinary meaning of the words used.He notes the following features:

IBM's position is clear: the entire active membership of the Main Plan comprises a "class of persons" and theentire active membership of the I Plan with the exclusion of the specified persons is, a fortiori, a "class of persons".

What is the RBs' position on this? Mr Stallworthy addresses that by reference to the decisions of Morgan J inCapital Cranfield Trustees Limited v Beck, Re: the A.C. Skelton Pension & Life Assurance Scheme [2008]

Mr Simmonds is highly critical of that. He says that there is no convincing explanation of why there is a need forthe cohort to be limited and no attempt to explain where the dividing line is to be drawn. In particular, no explanation is given why such an approach is appropriate in the context of a power that, on its face, is expressed to "override any provisions of the Plan that are inconsistent with it". That is not quite fair to Mr Stallworthy. He does, in his detailed comparison of the old rules and the new rules, explain that those words are "required to allow Rule 56 to operate notwithstanding Rules 1 & 2(1), not because Rule 6 is intended to confer a wideranging power circumventing and abrogating the amendment power". Whether that explanation is at all convincing is an entirely different matter. It seems to me to be nothing more than an assertion of the conclusion which Mr Stallworthy wishes me to reach.

It is convenient at this point to look at the authorities relied on by Mr Stallworthy which I have just mentioned.The first is CCTL v Beck. This concerned a definition of normal retirement age at 65 for men and 60 for women "or such day as the Employers shall determine in any particular case and notify in writing to the Member concerned". The important words, for present purposes, are "in any particular case" since Morgan J's judgment focused to a considerable extent on the difference between a particular case (or a number of particular cases) and a class. In the present case, there is no similar limitation. Both Mr Stallworthy and Mr Simmonds cite [37] in their written closing submissions:

Mr Simmonds, hardly surprisingly, says that, given the absence of a limitation to 'any particular case', thepresent case falls within the second of the two categories in the above passage (a class of Members, or all Members). The word 'particular' connotes a part rather than the whole; by contrast, the word 'specified' connotes indication of the person or class to be affected but does not qualify the meaning of the word 'class' in a way that limits it to something less than the whole.

Mr Simmonds also perceives what Morgan J says as unsatisfactory (to use my word): it suffers from the samedifficulties that are faced by the RBs in the present case, namely the failure to identify where a difference in number becomes a difference in kind so that there is no rational or logical explanation as to where the line ought to be drawn. The inability to identify where the line ought to be drawn points, in his submission, towards the conclusion that the distinction is not a valid one.

Mr Stallworthy, relying also on [38] and [39] of Morgan J's judgment, reaches a different result. These read asfollows:

Mr Stallworthy submits that it seems improbable that Morgan J would have reached a different conclusion if thenotice in that case had sought to characterise a class of member as having been specified as 'particular cases' for whom an equalised NRD at age 65 should apply. I have no idea whether that is improbable or not. Morgan J's analysis would have led to the need to determine, in such a case, whether there was, indeed, a collection of particular cases of Members (on the one hand) or of a class of Member (on the other hand). There is no prima facie correct answer to that issue: it would depend on a close examination of the facts.

The next case is CATL v Gellately, in which Henderson J was concerned with a power in a wider form. There the provision at issue defined NRD as age 65 for men and 60 for women "except where otherwise stated or otherwise agreed by the Employers and the Trustees" (without any express requirement that members affected be notified). Henderson J, at [58] contrasted the wording with that found in CCTL v Beck:

Mr Stallworthy submits that in the same way in the present case; the provision "that any specified … class ofpersons … shall cease to be … Members" is not to be interpreted as a general power to specify all or the generality of Members as ceasing to be Members and thereby to terminate accrual without triggering a winding up.

He also says that the description by the RBs of the word "specified" is "effectively akin" to the word "particular"in CCTL v Beck. That point I can reject immediately. What followed from the use of the word "particular" in that case was the distinction between identified individuals on the one hand and a group of individuals identified as a class. I see no reason for drawing the same distinction between a sub-set of active members on the one hand and the set of all of the active members on the other hand. Both sets are a class (or, if the word "specified" in the Exclusion Power is said to qualify "class of persons" as well as "person" both sets are a specified class). Accordingly, I agree with Mr Simmonds when he says that the suggestion that "specified" is effectively akin to "particular" is inappropriate and, in the circumstances, that case is of no assistance to them.

Mr Simmonds submits that CATL v Gellately does not assist. The Exclusion Power provides that a notice served in exercise of it "shall override any provisions of the Plan that are inconsistent with it". In the circumstances, the fact that the Exclusion Power may be exercised without the formalities required by the 1983 Amendment Power is beside the point.

He makes this additional submission which I cannot sensibly paraphrase and so set out in full:

I now return to the first factor identified at paragraph [221(i)] above and the point which Mr Stallworthy makesabout the 1990 Trust Deed and Rules being intended to replicate the 1983 Trust Deed and Rules, with changes specifically addressed by the Trustee and IBM but no other changes, in plain English where possible, reference being made by him to Nabarro's instructions. I have said that I doubt the admissibility of that factor in relation to the exercise of interpretation in contrast with rectification. But even if it is admissible, it is still important to discover what the plain English used means and to start with its ordinary meaning. If it does not mean what the Trustee and IBM intended it to mean, the remedy is rectification and not to give the words a meaning which they cannot bear. Further, Mr Stallworthy's submission has to start with the proposition that everyone concerned knew what the old provisions meant. It is possible that the draftsman of the 1990 Trust Deed and Rules intended them to mean what IBM says they mean because he thought that that was what the old provisions meant. That, however, is speculation. What I must do is interpret the words used in their context. This is, ultimately, why I do not consider that it makes any difference to the point of interpretation whether one admits the Nabarro instructions or not. It is for similar reasons that I would also reject Mr Stallworthy's submission based on what one would or would not expect to see unannounced as a significant new power. In any case, Mr Tennet accepted in his reply (in the absence of Mr Stallworthy) that Rules 5 to 7 were no more than a replication in plain English of what had gone before.

One thing is, to my mind, clear, namely that on any interpretation of the words used (but subject to the Hole v Garnsey argument), the Exclusion Power goes beyond what was contained in the 1983 Trust Deed and Rules.

Mr Stallworthy comments that Rule 5 broadly corresponds to the provision previously appearing within the 1983 Trust Deed and Rules at the end of the definition of "Eligible Employee"; Rule 6 broadly corresponds to the provision previously appearing within the 1983 Rules in the second sentence of the definition of 'Eligible Employee' and Rule 7 builds on those provisions, permitting the Principal Employer to close the Main Plan to new entrants by directing that employees should no longer be eligible for admission to the Main Plan as new joiners without the consent of the Principal Employer.

I do not agree with his comment in relation to Rule 6. What the second sentence of the definition provided wasthat a person would cease to be an Eligible Employee on his name ceasing to be recorded in the register of employees. As I have already explained, it was not possible to preclude a person from joining or continuing as a Member by the simple expedient of removing his name from the register. If a person's name was not included in the register, he was not eligible to be or to continue as a Member with the decision of the Principal Employer being conclusive but only in cases of doubt, as I have explained. It seems to me that Rule 6 cannot be seen

I therefore agree with Mr Simmonds when he says that Rule 6 was a new provision; from which he draws theconclusion that there is no reason why the project to re-write the existing provisions in plain English should have any bearing on the construction of Rule 6, a conclusion with which I agree. I record the forensic point made by Mr Simmonds, which cannot be used as an argument actually to support the conclusion which he draws, that until quite recently both the Trustee and IBM and their respective advisers considered that the Exclusion Power did have the meaning for which IBM now contends.

It follows that I do not consider there is anything in the first factor set out at paragraph [221(i)] above (the lackof any Exclusion Power in the 1983 Trust Deed and Rules).

The next factor relied on by Mr Stallworthy is the effect of the fetters on the 1983 Amendment Power. Hesubmits that on any objective basis, the 1990 Trust Deed and Rules are unlikely to have been intended and understood by its parties as having an effect that would (unless the Actuary made an error of law and therefore failed to express appropriate opinions under provisos (c)-(e)) engage the fetters on the amendment power and invalidate the offending provisions of the 1990 Rules. That is to say that, if an interpretation which results in the fetters not being breached is possible, it is the one which should be adopted in preference to one which does. He puts it rather higher: absent unequivocal terms which cannot be given any other meaning, the provisions of the 1990 Trust Deed and Rules ought to be interpreted on a basis consistent with the restrictions which the Actuary was required to police; and thus on the basis that the Actuary was correct in giving no opinion engaging those provisos. The reasonable person with all the background knowledge available to the parties would have understood them to be producing amendments falling within the ambit of the restrictions which the Actuary was required to police. In other words, one should search for a construction which complies with the fetters rather than imposes a limitation to give partial validity to an otherwise invalid provision, which is IBM's solution. Under the construction for which Mr Stallworthy contends, IBM could not exercise the Exclusion Power in the way it has purported to do.

He goes on to say, for reasons I will come to, that such an interpretation may be reached simply on the wordingof the Exclusion Power by implying a restriction to its scope or (on a Hole v Garnsey basis) construing the scope of the Power confined to such uses as can reasonably be considered to have been within the contemplation of the parties when the deed was made, having regard to the nature and circumstances of the deed. On this basis, the Exclusion Power would be restricted to causing existing Members to cease to be Members only on the basis of objectively demonstrable changes to the terms of their employment (eg a substantial reduction in their contractual hours) such as might even under the 1983 Trust Deed and Rules have justified their removal from the Employer's register of permanent full-time employees.

Mr Simmonds submits that reliance on the fetters does not advance the case of the RBs in relation to theinterpretation of the Exclusion Power. Even accepting that the fetters are engaged so far as concerns preservation of the final salary link, the correct way of continuing the link is by the introduction of the implied limitation which I have discussed at length under Issue 1. This has no bearing, it is said, on the question whether a single notice exercising the Exclusion Power can specify a class of persons comprising the whole of the membership of the Main Plan. I agree with that last proposition. Mr Stallworthy's argument based on the fetters applies as much to an exercise of the Exclusion Power in relation to a group of persons who undoubtedly do form a specified class as much as it does to an exercise of the power in relation to the whole active membership.

As to the I Plan, it should be remembered that the fetters have no relevance to the interpretation of the ExclusionPower in the I Plan. If the meaning of the Exclusion Power in the Main Plan is a factor to be brought into account in the interpretation of the Exclusion Power in the I Plan, it does not, in my view, carry much weight. This is especially so given what the parties to the I Plan actually thought it meant. As to that, the pointers, such as they are, are that everyone involved thought that it had the meaning for which IBM now contends. That makes the actual meaning of the Exclusion Power in the Main Plan a wholly unreliable guide to the meaning of the Exclusion Power in the I Plan. Indeed, if it is right that the Exclusion Power in the Main Plan is to be held valid

The third factor relied on by Mr Stallworthy is the location of Rule 6 within the structure of the 1990 Rules. Itappears within the "Membership" section of the 1990 Trust Deed and Rules, at Schedule D. It does not appear within the "Termination of Plan" section at Schedule M nor even within the provisions within the

Mr Simmonds' riposte to that submission is that the Exclusion Power is not found in the sections of the 1990 Trust Deed and Rules dealing with termination of the Plan or with withdrawal for the simple reason that it is not concerned with termination or withdrawal. It is concerned with membership of the Plan (hence its inclusion in Schedule D makes perfect sense): but the inability to accrue benefits is a natural corollary of not being a member.

In my view, the location of the Exclusion Power is entirely neutral. It is not at all surprising to me that it is foundwhere it is, for the reasons given by Mr Simmonds. Suppose that the Exclusion Power had been drafted in such a way which made the RBs' construction unarguable and IBM's construction obviously correct, would the informed reader respond by saying "What a strange place to find this provision"? I venture to say that the answer would be "No". Having expressed that view, I ought to note that the implied limitation necessary to preserve the final salary link would have an impact on other provisions of the 1990 Trust Deed and Rules because, somewhere, one would need to find a provision (implied) qualifying the level of benefit of a Member who ceased to be a Member as a result of the exercise of the Exclusion Power rather than as the result of, for instance, ceasing employment with IBM to take a job outside IBM. That consideration does not, however, mean that the Exclusion Power itself is found in a surprising location.

The fourth factor relied on by Mr Stallworthy relates to the provisions surrounding Rule 6. The first point is thatwhat has been referred to as the Exclusion Power does not, in fact, include the words "exclude" or "exclusion" in its text. The heading does so, but the Rules contain the common provision that "headings are for ease of reference only, and do not form part of the Rules". Mr Stallworthy says that, since the heading is not part of the Rules, it does not inform or affect their interpretation. I do not derive much from that observation since, even on the construction which Mr Stallworthy invites me to adopt, the result is a power to exclude certain people from membership: it is simply that the circumstances and manner in which such power can be exercised are narrower on the RBs' case than on IBM's case.

It is said by Mr Stallworthy that Rule 7 sheds light on the true scope of Rule 6 both structurally and in itswording. Essentially, there would be no reason to make specific provision for closing the Main Plan to new entrants under Rule 7 if Rule 6 already provided a broader power to direct that all such employees should be ineligible to become Members. Rule 7 would be superfluous if Rule 6 conferred a generalised power to render them ineligible; which makes it even more improbable that Rule 6 was wide enough to terminate accrual by all existing Members. The fact that a separate and explicit power was needed to permit closure to new entrants surely means that a separate and explicit power would likewise have been needed to permit closure to all accrual by existing Members. He relies on the fact that Rule 6 makes no provision for the Members affected even to be informed that their Pensionable Service is being terminated. He suggests that it is surprising, if Rule 6 was

Mr Stallworthy submits that the scope of Rule 6 is also informed by the termination provisions within the 1990 Rules.

It would seem surprising he says for the termination provisions of the 1990 Rules to provide expressly fortermination only with winding up (at least as regards active members); but then to find elsewhere a unilateral power by which the Principal Employer could terminate accrual for all or specified categories of active members without a corresponding winding up for such members.

I do not see much force in that submission. The Exclusion Power is not restricted to the same circumstances aswould give rise to a partial or complete termination of the Plan leading to a partial or full winding-up (and even then, subject to the Trustee's power to continue to run the Plan as a closed scheme). The Exclusion Power can be used in relation to an individual member or specified members in circumstances where the termination power would not apply. In any event, the focus is entirely different. The powers to which Mr Stallworthy refers in this part of his argument relate to one or more Participating Employers whereas the Exclusion Power relates to members. The consequence of the exercise of those two powers may, in some circumstances, be similar, namely the cessation of benefit accrual for a group of members and the end of a contribution obligation in respect of future service. But the route to that result is different.

Mr Simmonds accepts that Rule 6 is of broader scope than Rule 7 and that there is nothing that can be donepursuant to Rule 7 that could not be done pursuant to Rule 6. It is a realistic acceptance for the purposes of the particular question at issue, but I consider he accepts too much. Rule 6 results in an affected Member ceasing to be eligible for Membership. In contrast, Rule 7 says nothing about eligibility: the closure to new entrants means that nobody has the right to join the Plan but it does not result in otherwise eligible persons ceasing to be eligible. The consequence of closure is not that no person shall be entitled to become a Member but that he shall not be so entitled without the consent of the Principal Employer. So Rule 7 permits something to be done which cannot be done under Rule 6.

Quite apart from that, Mr Simmonds says that the conclusion which the RBs seek to draw from all of the aboveis to circumscribe the scope of the Exclusion Power in a manner which goes beyond the natural and ordinary meaning of the words used. It is difficult not agree with Mr Simmonds, even taking his acceptance of the broader scope of Rule 6 on its face, when he says that it is by no means unheard of for the governing documentation of a pension scheme to contain superfluous provisions. Such superfluity is particularly unsurprising given that the governing provisions of pension schemes are the product of amendment and restatement over an extended period. And that is as true of the "plain English" exercise as of any other amendment. Mr Simmonds points to a (small) example of incorrect drafting in Rule 2(2)(e) which refers to a member who is "precluded from continued membership of the Plan by the Employers under Rule 5, 6 or 7 of this Schedule". Rule 7, he correctly points out, has nothing to do with continued membership. It is clear what it means but such less than perfect drafting is only to be expected and no criticism can really be made.

Mr Simmonds also accepts that there is no provision for the affected Members to be notified that theirpensionable service is being terminated. In any case, it is equally true that the rule relating to termination of the

There are two subsidiary issues which arise. The first relates to "Postponed Retirees" that is to say a formerMember who remains in Service after his NRD with IBM's consent. Mr Simmonds no longer contends that the Exclusion Notices were effective in relation to Postponed Retirees who therefore continued as Members unaffected by those Notices. I need say no more about this issue.

The second issue concerns the effect of a notice pursuant to the Exclusion Power. The power, it will beremembered, permits IBM to direct that an existing member "shall cease to be" a Member. The RBs say that this must mean, in the case of the Main Plan, that an individual will cease to be a Member of the whole Plan and not just of the DB section of it. IBM agrees with that proposition.

The RBs submit that it follows that a DB Member precluded from membership of the Main Plan by an exerciseof the Exclusion Power can only resume membership of the Main Plan within the M Plan under the Money Purchase Rules if the Trustee agrees to that. An employee becomes a member of the Money Purchase Section under Rule 1(2)(b) unless he is precluded from membership of the Plan by an employer under Clause 4 of Part V (that is to say the Exclusion Clause) among other matters. And Rule 2(1)(e) provides that a Member (ie of the Money Purchase Section) shall cease to be a Member if he is precluded from continued membership under that same provision. And so it is said a person can only become a Member or resume Membership with the consent of the Trustee. Membership can be resumed under Rule 3 "at the sole discretion of the Trustee" or can be commenced for the first time under Rule 1(5) provided that the Trustee waives the eligibility and entrance requirements.

Mr Simmonds presents a different interpretation. He notes, as the RBs have themselves noted, that the words"exclude" and "exclusion" do not feature in the text of the Exclusion Power. The Exclusion Power contains two limbs: a power to direct that a person or class "shall not be eligible for membership" and a power to direct that a person or class "shall cease to be" a Member or Members. It is therefore possible, he says, for IBM to direct that a member shall cease to be a Member without thereby rendering him ineligible for membership. That, according to him, is exactly what the Exclusion Notices did, in the case of the Main Plan expressly stating that it did not constitute a direction that the persons so ceasing to be members of the Main Plan should be ineligible to join the M Plan.

In any case, he says that even if, as the Trustee now argues, the Exclusion Notice meant that the affectedMembers could not join the M Plan without the consent of the Trustee, that does not mean that the Exclusion Notice was ineffective to bring about a cessation of membership of the Main Plan.

I have to say that the interaction of the Exclusion Power and the eligibility and membership provisions of the MPlan is not entirely clear. The difficulty with Mr Simmonds' primary position is this: If a Member ceases to be a Member as the result of a notice under the Exclusion Power but is not subject to a direction that he is ineligible for membership, what is it that precludes an entitlement immediately to rejoin the Plan without any consent at all, whether from IBM or the Trustee? On the facts of the present case, IBM might be content to say that there is nothing to prevent that but since the C Plan is closed, the individual would only be able to join the M Plan. But suppose that IBM had simply wanted to exclude employees working at a particular location from membership of the C Plan, wishing them to join the M Plan and had served a notice to that effect, what would there be to stop the members concerned from rejoining the scheme? The solution would be for IBM to serve a notice precluding the affected employees from eligibility. But if that is necessary, what then, one might ask, is the content of the power to bring about a cessation of membership?

Accordingly, it seems to me that it is inherent in the power to exclude a person from membership that he therebybecomes ineligible for membership and that the reason for the two limbs of Rule 6 is to deal separately with

This does not mean that the affected Members cannot join the M Plan since the Trustee has sufficient powersunder Rule 1(5) to waive the eligibility requirements to allow the Member to rejoin under Clause 3: see paragraph 266 above. But I agree with Mr Stallworthy when he says that the consent of the Trustee is required for an employee, who has been the subject of an exercise of the Exclusion Power, to join the M Plan.

But in my judgement, Mr Simmonds is also correct when he says that that does not mean that the ExclusionNotice was ineffective to bring about a cessation of Membership of the Main Plan. This may have implications in the context of Project Waltz, but that is a different matter.

I will consider the argument based on Hole v Garnsey later.

The RBs contend, in the alternative, that the Exclusion Notices were a purported use of the Exclusion Powers fora collateral and improper purpose and as such invalid. They assert that on any basis the Exclusion Powers were quite clearly not included within the governing provisions of the UK DB Plans for the purpose of closing the plans to accrual (either partially or totally). And they assert that even if the Exclusion Powers were considered to permit Holdings to exclude specified classes of employee from the description or category of employment to which the plans relate, that is not the purpose for which Holdings has now sought to use the Exclusion Powers.

Rather, the purpose for which Holdings sought to use the Exclusion Powers was a closure of the UK DB Plansto all DB accrual by all Members, save those protected by contractual arrangements namely in both the Main Plan and the I Plan, the Enhanced Deferred Members and additionally in the I Plan, the exempted members listed in Appendices A-F to the Exclusion Notice.

This it is submitted is consistent with how IBM itself has characterised this aspect of Project Waltz. It has beendescribed as a closure of the UK DB Plans to future accrual (save by exempted members), not as an adjustment to the description or category of employment to which the UK DB Plans related. That is certainly how Mr Riley has, I accept, described it; see his email dated 7 July 2009 set out at paragraph 18 above.

Mr Stallworthy submits that the fact that the Exclusion Power was being used for a collateral and improperpurpose is starkly demonstrated in the case of the Main Plan by the form of the original 2010 Exclusion Notice. Exercise of the Exclusion Power sought to achieve two things: first that "all Members currently accruing benefits under the Defined Benefit Rules shall cease to be Members of the Plan with effect on and from the 6 April 2011" but secondly (and he would say inconsistently) "this does not constitute a direction that the persons so ceasing shall not be eligible to join the M Plan …". The lack of consistency stems from the fact that the

And so it is said that IBM's true purpose and objective was not to preclude affected Members from membershipof the Main Plan (as demonstrated by the proposition that they should be eligible immediately to rejoin the Main Plan within the M Plan), but rather collateral and improper purposes of seeking (a) to break final salary linkage (without triggering a winding up); and (b) unilaterally to force affected Members from the DB section to the M Plan in a way that could otherwise only be achieved by amendment with the cooperation of the Trustee.

The Trustee has raised the possibility that it might be relevant that members were not directly transferred byHoldings from the DB section to the M Plan, but were given a choice as to whether to be admitted to the M

Mr Stallworthy next refers back to the fetters on the 1983 Amendment Power, saying that IBM had no power tobreak the final salary link because of those fetters. He points out (i) that the amendment power in the 1997 version of the Main Plan Deed is subject to the same fetters and (ii) that the amendment power in the I Plan (in the exercise of which in the Principal Employer must act "in a fiduciary manner" and only with the Trustee's consent) is subject to a proviso prohibiting retrospective amendments "which would result in the reduction of any rights of [a beneficiary]".

It is submitted, therefore, that there was an attempt by IBM to achieve by two steps (namely an exercise of theExclusion Power followed by re-admittance to the Main Plan within the M Plan) what could not be achieved by one step (namely by an amendment bringing about a break in the final salary linkage applicable to past accrual combined with changing future accrual from DB to DC). That result was not a purpose for which the Exclusion Powers were conferred. Further, to have achieved the change in one step would have required an amendment (even if it could have been done without breaching the fetters) which involved the participation of the Trustee and could not have been achieved unilaterally.

The thrust of those submissions is that the Exclusion Power, assuming it to have been properly introduced, wasactually exercised for an improper purpose, not to bring about an end to the membership of the Plans but only of the DB sections of the Plans in order to force the Members into the M Plan (albeit with a choice to go into the IBM UK PPP or to have no pension provision at all).

Mr Simmonds explains why the Exclusion Notices took the form which they did. The explanation is that IBMwanted former DB members to have the choice between joining the M Plan or entering into the IBM UK PPP arrangement. At that time, given the Trustee's concerns regarding the lawfulness of the Project Waltz changes, it was not prepared to consent to anything. IBM considered it arguable that, in the absence of a direction that relevant members were not eligible to join the M Plan, those members had the right (without requiring Trustee consent) to join the M Plan under Rule 1(1) of Schedule B to the 1997 Money Purchase Rules. I do not propose to go into the merits of that argument other than to say it draws a distinction between being "precluded from membership" and being "precluded from continuing membership" in two different rules, namely Rule 1(2)(b) and Rule 2(1)(e) of Schedule B to the 1990 Trust Deed and Rules. It was to leave that option open as a possibility to the Members concerned – that is to say, to join the M Plan without the need for the consent of the Trustee – that this proviso was inserted into the Exclusion Notice. The argument may be a good one, it may be a bad one: I do not need to decide. If it is a bad one, then Trustee consent is required, but it does not impact on the validity or otherwise of the Exclusion Notice in terms of improper purpose, although it may have an impact in relation to the Imperial duty.

I say that the argument just referred to does not impact on the validity of the Exclusion Notice because MrStallworthy's arguments do not depend on whether or not the consent of the Trustee was required to join the M Plan. The point is that the purpose of the exercise was not genuinely to remove the affected members from the Main Plan; it was a different purpose of removing them from the DB sections of the Main Plan.

IBM denies that it was forcing any member to join the M Plan and submits that it cannot on any view have beenimproper to have exercised the Exclusion Power in such a way that kept open that opportunity for Members. For my part, I have found it helpful to consider a slightly different scenario in assessing the validity of the exercise of the Exclusion Power. Suppose that, instead of offering membership of a DC section of the Main Plan, IBM had created a new, free-standing, money purchase arrangement and given Members precisely the same option of joining such new arrangements as was given in relation to joining the M Plan. The fact, if it be a fact, that the Exclusion Power was being used to bring about a situation where Members would, it was hoped, join the new arrangements does not mean that the Exclusion Power was being used for an improper purpose. The Exclusion Power, again assuming it was validly introduced, would surely be being used for precisely the purpose for which it was given, namely to bring about an end to membership of the Main Plan or I Plan and, in consequence, to bring about a cessation of further benefit accrual. If that is right, then I do not see why it would be improper to achieve the same result by providing benefits through the M Plan.

There is one other point made by Mr Stallworthy which I should mention. He says that the governing provisionsof the UK DB Plans did not give Holdings a unilateral power to close the UK DB Plans to DB accrual without triggering a winding up. Accordingly, this was an attempt to use the Exclusion Powers for a purpose foreign to that for which they were conferred.

Mr Simmonds says that that is quite simply wrong. He accepts that there was no such power in the 1983 TrustDeed and Rules, but according to him, the effect of the 1990 Trust Deed and Rules and the introduction of the

Drawing all of this together, I repeat that Issues 1 to 3 (and in particular Issues 1 and 2) are closelyinterdependent. I have set out the arguments and made various comments and observations on them as I have gone along. I therefore propose to express my conclusions briefly.

In my judgement;

In reaching the first three of those conclusions, I have formed the firm view that the narrowest interpretation forwhich Mr Stallworthy contends cannot be right. It is, in effect, to treat the 1990 Trust Deed and Rules as reflecting precisely the provisions of the 1983 Trust Deed and Rules save where Nabarro's instructions were to the contrary. In my judgement, it is not possible to arrive at that conclusion in the light of the language actually used. Whatever the strengths of the arguments in favour of the view that the Exclusion Power does not extend to the exclusion from membership of the whole of the active membership of the Main Plan (other than the Postponed Retirees), I can see no justification for excluding an exercise of the Exclusion Power in relation to a class of persons (eg employees at a particular location).

Nor do I accept Mr Stallworthy's application of the Hole v Garnsey principle which, for reasons already given and which he himself contends, is really to be seen a part of, or certainly closely linked to, the question of

I have also formed the firm view that, once it is accepted that the Exclusion Power can be exercised in relation toa class of Members in a way which is not restricted to the manner in which Mr Stallworthy's primary submission would require, it is permissible to exercise the Exclusion Power in the way in which it was in fact exercised subject to the exclusion of Postponed Retirees. Since the amendment introducing the Exclusion Power is subject to the implied limitation which I have discussed, the result of the exercise will be to preserve the final salary link. For reasons already given, I do not find Mr Stallworthy's submission in relation to CCTL v Beck and CATL v Gallately persuasive. This is a case where, in my judgement, it is appropriate to determine that the introduction of the Exclusion Power was, pro tanto, valid.

The answers to Issues 1 to 3 are as follows:

These answers are without prejudice to any issues which arise in relation to the Imperial duty.

I now turn to issue 4 and 4B. I propose to deal with matters under the following main headings:

So far as witnesses of fact are concerned, I propose, first, simply to identify them and to state the positions they held. I will then go on to make some observations about some of them.

On behalf of IBM, the following six witnesses were called and were cross-examined by Mr Spink and by eitherMr Tennet or Mr Stallworthy:

On behalf of the RBs, the following eleven witnesses were called and were cross-examined by either MrSimmonds or Mr Newman and, in some cases, by Mr Spink:

I have also read a witness statement from Mr Larry Hirst ("Mr Hirst"). He was unable to attend the trial due to health problems. His witness statement was the subject matter of a Hearsay Notice. Mr Hirst was a very senior IBM employee. He is a former director of Holdings and IBM UKL. His career at IBM spanned 1977 to July

On behalf of the Trustee, the following three witnesses were called and were cross-examined by either MrSimmonds or Mr Newman and by either Mr Tennet or Mr Stallworthy:

I have thought it helpful to collect my observations about the witnesses in one place, and to put theseobservations near the beginning of my consideration of the evidence. This section of my judgment cannot be made proper sense of until later sections of the judgment, where I deal with the history of the development of Ocean, Soto and Project Waltz in some detail, have been digested. In particular, there is reference in this section to events and concepts which I identify and describe much later on and, until those events and concepts are understood, I am afraid that certain things will, on a first reading, seem obscure or even incomprehensible to a reader coming afresh to this litigation.

I start with IBM's witnesses.

Mr Chrystie's first involvement of any sort was on 20 February 2009 when he attended a meeting with MrKoppl. This was after the "3 plays" (see paragraphs 1182ff below) had been agreed on 4 February. He was briefed by Mr Koppl about the UK pensions changes that were being made to assist in safeguarding "the 2010

Mr Chrystie himself came across as something of an advocate in IBM's cause in his oral testimony. I do not,however, doubt his honesty or integrity, although as will be seen later, his focus on the UK may have caused him to lose sight of the bigger (from CHQ's perspective) picture if, indeed, he had ever appreciated its full scope.

I did not find Mr Ferrar to be an impressive witness. In certain respects, I found his evidence unreliable and hewas, on any view I think, inappropriately defensive. He was prepared to downplay his involvement in Project Whisper (see 1109ff below) and his understanding of it when, in my assessment, he knew that it was more than contingency planning for hypothetical options as he had sought to explain his understanding of the project. In his original evidence, his summons to the US to take part in the Project Whisper meeting from 15 to 19 October 2008 was not mentioned; nor were the slides for the "work-plan" to reduce DB liabilities under Project Sapporo mentioned let alone disclosed. He failed to disclose to the Trustee board of which he was a member, material which would have been germane to its consideration of Project Waltz. Mr Heath

Mr Heath, as Director of Human Resources for IBM UK from September 2003 to January 2007, was the manresponsible for communicating the Ocean and Soto Proposals to members. He prepared the relevant

I find him to be an honest and straightforward witness whose evidence I can rely on. He attempted to explain hisrecollection of events truthfully and clearly. Inevitably, almost, at this distance of time he was occasionally confused, with one example of a need for further cross-examination again following re-examination after his initial cross-examination. His honesty however is not, in my mind, in question. Further, he was clear about what he could and what he could not remember, albeit that his recollection was not always perfect. I accept the accuracy and reliability in respect of the matters which he said he could remember.

Mr Koppl is in some respects a surprising witness. Not because he is unreliable or in any sense untruthful – quitethe reverse – but because his evidence is not particularly helpful to IBM when it comes to establishing the reasons for the pension changes in the UK. Mr Koppl, as I have said, is and was Director of Pension Analytics within the Finance function at CHQ in Armonk. I think it is fair to say from an assessment of all the evidence that he was pivotal in (Mr Tennet would say the mastermind behind) the development of the Soto changes and that he was heavily relied on by CHQ and by the UK teams in the formulation of Project Waltz.

Mr Tennet's assessment of Mr Koppl was this: he was a frank and honest witness. The impression he gave wasthat he was genuinely trying to explain his own recollection of events and thought processes at the relevant time accurately and fairly. His recollection of events was good, and he indicated clearly when he could not remember a particular matter. I agree with that assessment.

One thing that came across clearly from his evidence was the involvement which he had in formulating theProject Waltz proposals and the readiness with which they were accepted by CHQ. Mr Koppl was concerned with the global picture and not with the development of justifications on a country-by-country basis and in particular with the communication exercise for UKI. He clearly had a great deal of influence in the CHQ decision-making process on the Project Waltz changes.

I have found Mr MacDonald a very difficult witness to assess. He is a man of charm and charisma. But he alsohas, it appears to me, a quite dominant character. He consults with those who work under him and around him but he clearly "calls the shots" in his own area of responsibility and the extent to which he hears, rather than simply listens to, those he consults is perhaps not always clear. I do not question his honesty and integrity. But having said that, I have to say that in some areas his evidence is really based on reconstruction rather than recollection. That he has no actual recollection of some matters is no surprise; the events took place quite some time ago and Mr MacDonald is a senior manager whom one would not expect to be involved in the detail of implementation of pension changes.

His own evidence in that last respect used a colourful (or I suppose colorful in his case) turn of phrase: he had totake a view from no lower than "a 50,000 foot level". Mr MacDonald was, by his own admission, not a "pensions man" to use my phrase; nor was his expertise in finance or accounting. His evidence showed that he was heavily reliant on others (such as Mr Koppl). His lack of financial/pensions expertise led him into error: for example, his witness statement stated that after Soto "costs continued to increase year on year" but, having been shown a document in cross-examination, he accepted that it demonstrated that pension costs would decrease.

Mr MacDonald's recollection was not always accurate. Mr Tennet refers to one example concerning the meetingat the Marriott Hotel in Portsmouth (UK) (as to which see paragraph 813 below). This was an important

I have to treat Mr Riley's evidence with a degree of caution. He was, I think, an honest witness although he didoccasionally need to change his position as a result of cross-examination. Mr Tennet gives one example concerning his "sametime" exchange with Mr Chrystie on 17 March 2009. He initially denied having warned Mr Chrystie not to speak to Mr Wilson about pensions issues. But when shown this "sametime" exchange, he accepted that he could well have told Mr Chrystie that he was not to discuss Project Waltz with Mr Wilson. I do not suggest there was any dishonesty here; indeed, in that relevant exchange, Mr Riley sought to explain what his understanding had been and why he had given the earlier answer which he had.

As with Mr Ferrar, his original witness statement can now be seen to be interesting for what it failed to say. Itspurpose was, as he put it, to focus on the business reasons behind the decision to adopt Project Waltz. But he made no mention at all of one business reason, namely a directive from CHQ that retirement related costs savings be made as a matter of urgency to fix NPPC and ensure that the 2010 EPS Roadmap was met. And yet, in cross-examination, he not only accepted that this was a reason, but he also accepted that it was an important motivation for the changes.

One positive matter which he stated in his witness statement was that he was leading a transformation of thebusiness at the time of Project Waltz:

That sits very uncomfortably with his oral evidence during the course of which it became clear that reviewingthe cost implications of the Plans was no part of the UKI Transformation strategy until CHQ made it a requirement for retirement-related costs savings to be achieved at the end of October 2008. Consistently with that, the minutes of the TMM on 23 October 2008 record Mr Riley as saying that there were "currently no discussions going on as regards the IBM UK pension plans". The RBs' witnesses: the member witnesses

In IBM's skeleton argument, the point was made that the evidence of the member witnesses is not relevant or, atthe very least, their evidence is no more relevant than other evidence to the effect that other members of the Plans did not have the same understanding about IBM's commitment to continue DB accrual. That point is repeated in the written closing submissions. It is one which I will need to address more fully later on.

Mr Simmonds, in addition, draws attention to the following aspects which he detects in the member witnesses'evidence:

Having noted those observations, which have some force, I should add that all of the member witnesses struckme as honest and doing their best to help the court. If their evidence is not relevant, as Mr Simmonds suggests, that is not their fault. But as with some of the evidence given on behalf of IBM, not everything is recollection, some is reconstruction perhaps coloured by the resentment to which Mr Simmonds has referred.

In dealing with these witnesses generally, I have found it helpful to start with Mr Simmonds' assessment as setout in the written closing submissions and to say where I agree or disagree, taking account of everything I have heard from the witnesses and from Mr Tennet. I have found helpful his approach of identifying, in the case of each witness, the source of the understanding of IBM's commitment which the witness relies on, by far the most important sources being Mr Heath's Webcast and the terms of the guarantee. I have described the guarantee at paragraph 111 above. A description of the Webcast can be found at paragraphs 614ff below; a general understanding of it is necessary to understand the evidence of the member witnesses and what Mr Simmonds has to say about that evidence.

Now, it may be that a very detailed analysis, sentence by sentence, of the Webcast (and the Funding Agreementincluding the guarantee) and questions closely focused on each sentence in the light of that analysis,

I also need to point out that much of Mr Simmonds' cross-examination was directed at showing that, on a carefulreading of the Webcast and the guarantee, the understandings which the member witnesses say they had were not justified. To some extent he succeeded in that because, on a sentence by sentence analysis, each witness broadly accepted many of the propositions put to them. He did put to each of them that there was nothing in the Webcast or the guarantee to justify the conclusion that IBM was undertaking that future service accrual in the DB Plans would continue for any period, but he did not suggest that they were making up their evidence about what they understood at the time.

I must add at this point that the members were never shown, nor were they able to obtain, the Funding

Taking the preceding three paragraphs together, I reject any suggestion that any of the member witnesses issimply making things up to suit their case.

It is to the individual member witnesses that I now turn, in each case starting with Mr Simmonds' assessment. Iaccept his assessment as fair, subject to the particular observation which I make.

Mr Simmonds' assessment:

Mr Simmonds' assessment:

Mr Simmonds' assessment:

Mr Simmonds assessment:

Mr Mills understood the nature of Mr Heath's statements as indicating IBM's hope that the Plans would remainopen, whilst not guaranteeing the same.

To similar effect, Mr Mills subsequently described IBM's position "as an intent not to make a change, an intentto make it happen and a commitment to do certain things along the way…like pay in the money". He saw that the only commitment was to fund the Plans; and that the consequent intent to keep the Plans open was just that – a statement of current intention.

Finally, when discussing Project Soto, Mr Mills said as follows:

Nonetheless, Mr Mills' overall understanding as explained in his witness statement was that the Plans would notbe closed during the period of the guarantee which had been put forward as part of the reasons why the Plans were sustainable. It is also appropriate to mention at this point that Mr Mills qualified his answer concerning IBM simply providing a statement of intent by saying that IBM could only change its mind in exceptional circumstances, such as a collapse of its business.

Mr Simmonds' response to that is that an intention not to do something save in exceptional circumstances is notthe same as a warranty that changes will not be made outside of those circumstances: a qualified intention remains an intention, and not an unequivocal representation capable of being given legal effect. I will be dealing with the communications in due course and with this submission in particular.

Mr Simmonds' assessment:

That Mr Turner's evidence may have contained some inconsistancies and that he was unable to controvert someof the propositions put to him by Mr Simmonds is true. If it is suggested that he was deliberately giving evidence which he did not believe was true, I reject it. I do, however, think that some of his evidence is reconstruction and reflects a degree of wishful thinking. Mr Dalgleish

Mr Simmonds' assessment:

Mr Simmonds' assessment:

Mr Simmonds' assessment:

Mr Newton is a member of the M Plan. His evidence went to a rather different aspect of the case. One of thereasons for the Project Waltz changes is said by IBM to have been the need to deal with the disparity of pension treatment as between DB and DC members. It is said that DC members (including, but not restricted to DC members in the UK) were concerned that their prospects of obtaining salary increases were being harmed by the

Mr Simmonds' assessment:

I do not propose to enter into a detailed analysis or discussion of Mr Simmonds' assessment of the RBs'witnesses. What he records them as saying is, I believe, accurate, certainly accurate enough for present purposes, and Mr Tennet has not made any serious criticism about that. Many of Mr Simmonds' observations about inconsistency between the different witnesses, and even within the evidence of an individual witness, are selfevident from what he has recorded them as saying.

It is fair for him to say, as he does, that there was a general confusion amongst many of the member witnessesbetween benefits and contributions, in particular in assuming that IBM's power to alter benefits for the future was limited to increasing employee contributions (as both Mr McRitchie and Mrs Cordina seem to have assumed judging by their answers in cross-examination). Similarly, in 2004 Mr Mills was not sufficiently informed about pensions to understand the difference between past and future service when it came to considering deficit and surplus.

By detailed questioning in relation to the Webcast and other communications (and also in relation to theguarantee which, as I have noted, the members never saw, their only understanding of its nature being what they were told by IBM and the Trustee), Mr Simmonds was able to establish that different members had different understandings of the starting point (in terms of what IBM could and could not do) and different understandings of what the Webcast and other communications were saying: they took different messages away. In those circumstances, he says that the understandable ignorance of at least some members giving evidence should make me very slow to bind IBM to the understanding of such members as to the effect of statements made on the future of the Plans' benefits. I should not be persuaded to construct a legally-binding arrangement from words that are, as he puts it, clearly not fit for that purpose.

I take those cautionary words on board. But, on the other hand, the Webcast and communications were directedat people who, although a comparatively sophisticated group, perhaps, compared with the membership of many other large UK pension schemes, are not pensions experts. I must attempt to interpret the Webcast and the other communications in the way that the hypothetical reasonable and objective member would understand them; and I must do so realising that the members will not have had the benefit of an appropriately qualified professional explaining in detail the nuances of the communications when subjected to detailed examination and realising that the hypothetical reasonable and objective member might not give the full attention to these communications which IBM's case might be seen as requiring them to have undertaken. I deal with this aspect of the evidence further when considering the topic of Reasonable Expectations (see paragraph 386 (iv) below).

What is clear from the evidence is that all of the members who actually gave evidence now regard IBM ashaving breached commitments to the members generally; and there is evidence to suggest that there were very many members (both DB and DC members) who regarded the Project Waltz changes at the time they were announced not merely as unfair but far more seriously, "betrayal" and "breach of trust" perhaps summing up the thrust of the complaints. I am quite sure that all of the members who gave evidence were being honest in the evidence which they gave. But I am also conscious that there is a real danger of a potent mix of reconstruction and wishful thinking about what they actually thought at the time rather than what they now say about the changes. Having said that, I do not doubt that they all genuinely considered at the time that the changes were unfair. I do not know whether even IBM would dispute that: it does not matter, because unfairness, on any view, is not enough to get the RBs home as I will explain. The Trustee's witnesses

I say only a very little about the Trustee's witnesses at this stage. I have no doubt that all of them were honestwitnesses doing their best to assist the court. Mr Lamb was an impressive witness. In contrast with the RBs' witnesses, I do not have any concerns about reconstruction or wishful thinking. I consider that when Mr Lamb says he understood what he was being told in a certain sense, that is indeed what he understood. Whether what he was told justified his understandings gives rise to different issues.

There is one final witness whom I should mention, Mr Wilson. He was called on behalf of the RBs. As with theTrustee's witnesses, I say very little about him at this stage. He was, in my assessment, an honest and reliable witness. As with Mr Lamb, I have no concerns about reconstruction or wishful thinking. I will consider relevant parts of his evidence in due course

I take those in more detail in turn.

A genuine and rational exercise of the power. I do not think that this is a controversial proposition. However far the Imperial duty may go or not go, it is right to conclude that an employer must not act in an "empty or irrational" way. The word "empty" is not mine but has been used on several occasions and denotes, to my mind, an absence of any reason. I think the words "empty" and "irrational" inform each other so that the composite "empty or irrational" has a pretty clear meaning. But this has nothing to do with fairness and little, if anything, to do with objective reasonableness. As Newey J put it in Prudential, assessing whether a decision is irrational or perverse (in the sense in which he was using those words) is not to be equated with the application of an objective standard of reasonableness.

Not a test of fairness. In Imperial Tobaco, Browne-Wilkinson V-C himself rejected the test of fairness and Newey J reached the same conclusion (see for instance at [142] of his judgment). Imperial Tobacco was approved in Malik and although this point did not need to be addressed in that case, there was no hint of qualification or reservation by the House of Lords in what it said. I do not consider that there can be any doubt that the test is not one of fairness. The essential reason is that appeals to fairness do not solve the problem. What is fair to one person may seem unfair to another; both the employer and the employees may have perfectly reasonable views even though those views are diametrically opposed. It is not for the court to take on the mantle of arbiter of fairness. Some other test has to be found.

A severe test. Newey J, at [132] of his judgment, described the test as a severe one, citing Hale LJ (as she then was) in Gogay v Hertfordshire CC [2000] IRLR 703 ("Gogay"). That was said in the context of an employment contract and not of a pension scheme. If the test is severe in that context then, a fortiori, it is severe in a pensions context. However, what Hale LJ meant by severe was that the conduct must be such to destroy or seriously damage the relationship so perhaps that takes one no further. Whether the severity is such that the test is actually one of irrationality or perversity is a matter I will come to.

The test is objective. I do not think that the RBs disagree that the test, whatever it may be, is objective. It is the approach which Newey J took in Prudential and which I took in the second rectification judgment.

There is a gulf between the RBs and IBM about the content of the test, but as to objectivity, they seem to acceptobjectivity as a guiding principle. Thus in the RBs' written closing submissions we find this:

I am sure that it is right. It is supported by what Lord Nicholls said in Malik at p 35 ("The conduct, must of course, impinge on the relationship in the sense that, objectively, it is likely to… breach the implied duty of good faith") and by Lord Steyn at p 47 agreeing with an article by Mr Douglas-Brodie who wrote that, in assessing whether there has been a breach:

It follows that, in taking this approach, the court is concerned with a "reasonable employee" and not with anyparticular actual employee or group of employees. No doubt the court needs to be put in a position, so far as it is possible to do so by proportionate means, to understand as fully as possible the context in which the actions (including what was said or not said to employees) said to give rise to breach of the implied duty of good faith occurred. That will include evidence being given to the court by members about what they were told. The court can, of course, look at the documents; it can also look at slides prepared for presentations. But what it cannot do is put itself in the chairs of the employees who attended meetings (save where videos are available). To learn what was actually said, the evidence of those attending is relevant. Further, it is relevant not only to what was actually said, but also to the way in which it was said and the impressions which, objectively, might reasonably have been gained. And it is relevant to establishing what IBM could reasonably have considered the level of understanding of the target audience might be. I do not consider that such evidence is irrelevant.

Nonetheless, I see it as far less relevant, if relevant at all, to know the subjective impressions gained by the smallnumber of individuals who have actually given evidence even if I accept that their understandings are representative of a large number of other members. This is an aspect I deal with below when considering the topic of Reasonable Expectations.

The employer's financial and other interests. Browne-Wilkinson V-C recognised that, in exercising the discretionary powers under consideration in that case and, I would say, in exercising IBM's powers in the present case, the employer is entitled to take account of its own financial and other interests subject to any constraint imposed by a requirement (if one exists) of consistency with any expectations on the part of the members. The employer's financial and other interests are factors to be taken into account in assessing whether there had been any breach of the employer's duty of good faith. Other authorities have adopted the same approach: see for instance Robert Walker J in National Grid Company plc v Laws [1997] PLR 157 at [88], Pumfrey J in Hoover Ltd v Hetherington [2002] PLR 297 at [53] and my own decision in Bradbury at [103]. Other things being equal, an employer is therefore entitled, quite clearly in my view, to prefer his own economic interests to those of his employees. By doing so, he may create all sorts of problems in terms of workforce motivation even leading to strikes and the like, but that is a matter for his commercial judgment. If a breach of the Imperial duty is to be established, there has to be more than a decision which the employees consider to be unreasonable: there has to be something more substantial. That is no doubt why the RBs in the present case place so much reliance on what was said to the members in the implementation of Ocean and Soto and do not simply rely on the terms of Project Waltz being in themselves so monstrous, to use my own word, that taken in isolation they breach the Imperial duty.

The parties have a very different view of the role of legitimate expectations in the scope of the Imperial duty. The RBs submit that it is central. IBM says that the concept, at least as developed in public law, has no part to play: the (only) test is whether the employer has acted irrationally or perversely. It follows that unless it would be irrational or perverse to depart from the members' legitimate expectations, there would be no breach of the duty. I return to legitimate expectations later in this judgment.

I should here record that there are two strands to the RBs' case. There is the first limb, the contractual arm,where reliance is placed on the well-established implied term in an employment contract, that neither employer nor employee will, without reasonable and proper cause, conduct himself in a manner calculated or likely to destroy or seriously damage the relationship of trust and confidence between employer and employee; and there is the second limb which is that an employer and a member of a pension scheme are under the Imperial duty. Both of these duties are admitted as matters of principle and it is also admitted that they apply to IBM's powers under, and in relation to, the Plans. Both strands are relevant in the present case, because not all of the Project Waltz proposals were to be implemented under powers in the Plans. They involved other aspects, such as the 2009 Non-Pensionability Agreements and the alleged pressure placed on members to enter into them.

In relation to a power vested in an employer under a pension scheme, the effect of the Imperial duty is to place a limitation on the manner in which an apparently unfettered power may be exercised. Quite what the

Mr Tennet submits that the origin of the Imperial duty entails that the two duties I have just identified have the same content in terms of what sort of conduct would be likely to destroy or seriously damage the relationship of trust and confidence between an employer and its employees. One consequence of that, if it is correct, is that the authorities concerning the implied duty in an employment contract will be relevant to the Imperial duty in a pensions context as well. Mr Tennet developed that submission in the RBs' written opening submissions. Without retreating from those submissions in any way, his case has moved on with a new focus which I shall come to in a moment. I do not therefore propose to address in any detail what was said in the opening although I have taken all of it into account in reaching the conclusions which I do.

Mr Simmonds' position is that little assistance can be gained from the employment law cases and that they arenot binding on this Court in relation to the issue of the scope and application of the Imperial duty. It is neither sensible nor appropriate simply to bolt on employment law to the law relating to the exercise of discretionary powers under an occupational pension scheme. Moreover, to do so would be contrary to Newey J's decision in Prudential.

The RBs' case on the scope of the Imperial duty now rests on the identification of a "two-pronged test" to use

Since the hearing before me, the second of those cases has been to the Court of Appeal: see [2013] EWCA (Civ) 394 ("Attrill"). Elias LJ, with whom Beatson LJ and Maurice Kay LJ agreed, said this at [101]-[102]:

Mr Tennet relies strongly on that to demonstrate that the two-pronged test is the correct approach. But for theavoidance of doubt, he makes clear that the RBs continue to maintain that a breach of the Imperial duty should be found even if I were to reject the two-pronged test notwithstanding the guidance given in Gogay and Attrill.

Mr Simmonds submitted in his oral closing that this approach by the RBs was simply an attempt to justify theintroduction of the test of reasonableness which had been rejected right from the beginning by BrowneWilkinson V-C in Imperial and consistently by other Judges since then. He submits that there is no basis for a two-stage approach even after the decision of the Court of Appeal in Attrill. The correct approach in relation to the exercise of discretionary powers (he makes no submission about other contexts) is what I can describe as a multifactorial approach (a jargon phrase familiar to intellectual property lawyers) or, more colloquially, everything goes into the melting pot. Everything is taken into account but the ultimate question is whether the employer has acted irrationally or perversely. His key point is that it is "simply not appropriate to enter into an analysis, an evaluation of the merits of the employer's decisions by reference to criteria that are essentially designed to answer the question: was this a reasonable thing to do?". He could make the same points even if a two-pronged approach is to be adopted, but on the basis that the "reasonable and proper cause" prong is satisfied provided that the employer can at least demonstrate a rational reason for his decision.

In the absence of authority, I would certainly agree with Mr Simmonds. I have already suggested that there is aconceptual problem about dividing the issue into two separate components. It is not unhelpful to do so as one way of illuminating the answer. But, to repeat, the words of Browne-Wilkinson V-C are not legislation and he was simply attempting to articulate, at the very beginning of the development of the duty, its core elements. Logically, if an employer has a reasonable cause or excuse to conduct himself in a way which might give rise to a question as to whether what he is doing is proper because on the face of it the result looks surprising, then he is not in breach of duty. Indeed, whether Mr Tennet is right or Mr Simmonds is right on this issue, the result will be, if there is a reasonable cause or excuse, that the Imperial duty is not breached at all; it is not that there has been a breach of duty which the Court then goes on to treat as no breach or as a breach which carries no remedy.

That is not to say that the two-pronged approach is not of assistance at all. In Gogay, for instance, the facts were fairly extreme. An allegation of sexual abuse is clearly likely seriously to undermine the relationship of trust and confidence in both directions. If the allegation turns out to be true, the relationship of trust and confidence is likely, nonetheless, to have been seriously undermined if not altogether destroyed. But there will have been no breach of the implied duty of good faith on the part of the employer because his conduct (the making of the accusation) would have been with proper cause (unless, for instance, made unnecessarily public until proven).

However, where the conduct is not of that nature but is such that a careful enquiry is needed to weigh theconduct in the balance against, for instance, reasonable expectations, the two-pronged approach appears to me to be a less than helpful approach. My strong inclination is to reject it, in a case such as the present, as a reliable guide.

Does Attrill require me to conclude otherwise? Nowhere in the judgments, in particular not in the passage relied on by Mr Tennet which I have quoted at paragraph 376 above, did the Court of Appeal say anything about this aspect of the Imperial duty, nor did it need to do so. Given that reasonableness is not the test for breach of the Imperial duty, as Imperial itself and Malik make clear, Mr Simmonds is correct to say that the approach adopted by Owen J and, without further analysis, by the Court of Appeal should not be allowed to re-introduce the rejected test.

Since the Court of Appeal simply followed Owen J, it is helpful to consider why Owen J articulated the test inthe way which he did. Mr Simmonds submits that Owen J in reality applied a unitary test. He adopts what Mr Spink said in his oral closing submissions.

The case concerned employee bonuses. As Mr Spink helpfully described the case, bonus promises had beenmade to a pool of employees which were then watered down contractually by the insertion of a rather significant clause the effect of which was that the bonuses would only be paid in certain circumstances. In that context, Owen J did seem to suggest that there was a two-stage test to be discerned from Malik: see [199] of his judgment. Mr Spink's submission is that when Owen J actually came to what constituted a breach of duty, he really applied a rolled up unitary test, referring me to [200] to [202] and [222] and [223]. Mr Spink is correct, in my view, when he says that Owen J did not make a finding that the relationship of trust and confidence was damaged until he had been through all aspects of the matter, including the question whether or not the offending contractual clause was inserted with reasonable or proper cause. And I agree with him when he suggests that Owen J's decision is not a clear authority for the requirement that there must be a formal two-stage test. It was not until [223], having been through the evidence, that Owen J was able to say that the introduction of the offending clause was regarded at the time as likely seriously to damage the relationship of trust and confidence. He did not subsequently go on to address whether that conduct, even though likely seriously to damage the relationship could nonetheless be conduct with a reasonable and proper cause.

In my judgement, the authorities do not force the court into the adoption of a formal two-pronged approach asadvocated by Mr Tennet. That approach will, no doubt, sometimes provide a sensible reality-check in relation to a decision reached on a unitary approach adopting the multifactorial analysis already referred to. In some cases – and Gogay would appear to be one example – it may provide more than a reality-check and in fact represent the correct approach: if conduct cries out for an explanation and none can be given, the conclusion may well be that there is a breach of duty.

Whether I am right or wrong on that, Mr Tennet makes a number of submissions about the Imperial duty which I should in the first place identify:

Mr Tennet also has a number of submissions in relation to Prudential and why it does not detract from the submissions above.

If, contrary to my view, a formal two-pronged approach is to be adopted, Mr Tennet also makes these points(although I should add that each of those points can I think be made, albeit in very slightly different language, if the unitary approach is correct):

Having identified the propositions which Mr Tennet seeks to establish, I remind myself what this case is reallyabout so that my focus is on the legal aspects which really matter. As Mr Simmonds puts it, and I agree so far as their primary case is concerned, the RBs case rests on two pillars:

That is how I perceived the nub of the case, particularly in the light of the written closing submissions from allparties by the time oral submissions began. As I put it to Mr Tennet as a summary:

In his closing oral submissions Mr Tennet confirmed that the RBs were not submitting that the severity of thecuts, in the absence of Reasonable Expectations, which were imposed on members, was a breach of the Imperial duty. He does, however, rely on the severity of the cuts as a factor "suggesting….. that there was a failure to maintain trust and confidence in this case". This is important because, as Mr Simmonds submits, if there were no reasonable or legitimate expectations which had been engendered by IBM prior to Project Waltz and which Project Waltz would have dashed, then Project Waltz, of itself, is not a breach of the Imperial duty. As to the

In the context in which I asked Mr Tennet the question which I did, he did not take exception to the way I put it.But it would be entirely wrong to view the entirety of the case that way. Representations giving rise to expectations certainly have a part to play in the present case; indeed, the issue is of central importance. But that is not the beginning and end of the matter since serious challenges are made by the RBs to the alleged justifications for Project Waltz which they say are no justifications at all. If that is right, then the purported reliance on such matters is at the very least a factor to be taken into account in assessing whether or not a breach of the Imperial duty is established. Further, the RBs say that IBM's statements about the sustainability of the Plans means that IBM must have believed, on the basis of the information available to it at the time, that its statements were true. Or, if IBM did not believe those statements or had no justification for making those statements because it had not taken steps to confirm them, it would be a breach of the Imperial duty to act in a way that is inconsistent with the truth of those statements: in other words, if there was no significant change in financial circumstances between the time when the statements were made and the time of Project Waltz, IBM cannot, according to the RBs, rely on non-sustainability as justification for the later changes.

At this stage, it is convenient to give further consideration to the severity of the test applicable to the Imperial duty. As I have mentioned, the RBs rely on some of the employment law cases to identify the scope of the Imperial duty but IBM submits that these cases do not assist. Mr Simmonds' position is that the bar has been set very high – he would say at the level of perversity or irrationality – and that the RBs seek to lower the bar by reference to the employment law authorities which do not admit of sensible application in the context of a pension scheme.

It is certainly correct that there is no reported pensions case where the Imperial duty has been applied which cannot be explained by reference to the test of perversity or irrationality. Perhaps Imperial itself is the nearest to a counter-example but there the facts were quite extreme and the duty was breached, if not because of perversity, then because the power of amendment was being used for an inadmissible purpose. As I said at [100] of my judgment in Bradbury:

Mr Tennet is, of course, right to say that the employment law cases were the foundation on which BrowneWilkinson V-C built the Imperial duty. But equally, the Imperial duty is not simply an application of employment law to a particular facet of the employment contract. It has to be recognised that it was a new duty whose justification and rationale was the parallel employment law duty. But what Browne-Wilkinson V-C did not do, at least expressly, was to import lock, stock and barrel, into the new duty all of the attributes of the established employment law duty. Had he done so, he would no doubt have explained in more detail than he did the parallels which could properly be drawn.

That he did not do so is entirely unsurprising. As I remarked in my judgement in Bradbury, Browne-Wilkinson V-C gave his judgment under considerable time-pressure and, against his inclination, gave some guidance about how the new principle would work in practice. And just as some of those examples are not entirely easy to understand as fitting within the new principle, perhaps because thought of in some haste, so too his statement of the new principle cannot be taken as a rigorous articulation of the scope of the Imperial duty. Moreover, even in an employment law context, there is support for the conclusion that, in some cases, the test of irrationality or perversity is the correct test: see Horkulak v Cantor Fitzgerald International [2005] ICR 402 ("Horkulak") and Clark v Nomura International plc [200] IRLR 766 ("Clark v Nomura"). Mr Simmonds also relies on what Newey J said in Prudential at [142]. I will come to the relevant passages in a moment.

Thus, although the duty in the employment law context and the duty in the pensions context can be stated in thesame language (ie an employer should not without reasonable and proper cause conduct himself in a manner calculated or likely to destroy or seriously damage the relationship of trust and confidence between employer and employee) the content may differ. It is certainly true that the Imperial duty is not contractual not least because otherwise it would not be owed to persons (such as a member's dependants) who had never been in an employment relationship with the employer.

Further, as Mr Simmonds points out, there cannot be a complete elision between the Imperial duty and the duty of good faith in an employment context for two other reasons. First, the latter comes to an end when the employment relationship comes to an end: see for instance Lord Millett in Johnson v Unisys Ltd [2003] 1 AC 518 ("Johnson v Unisys") at [78].

The second reason is that, in an employment law context, express and unrestricted powers cannot in the ordinaryway be circumscribed by an implied qualification; and that applies as much to the implied duty of good faith as to any other provision. In that context, Mr Simmonds relies on Reda v Flag Ltd. [2002] IRLR 747 ("Reda") where the Privy Council refused to apply the implied term to constrain the exercise by the employer of its express and unqualified right to terminate an employee's employment without cause. He submits that assimilation of the employment law duty and the Imperial duty would prove too much because, in a pensions context, the employer's powers are nearly always unrestricted or restricted in specific ways. On the basis of Reda, such unrestricted (or specifically restricted) powers would not be subject to the Imperial duty which would then be of very limited scope indeed.

In my view, that last submission goes too far. The point which was being made in Reda was that the implied term would not constrain the exercise of the power in a way which went against its express terms: in effect, a power to dismiss without cause could not be subject to a constraint which required cause to be shown. In the case of an ordinary power of amendment in a pension scheme, the power is unrestricted only in the sense that, on a literal reading, there are either no express constraints or express constraints usually imposed to protect members' accrued rights. It would not be inconsistent, in the sense that the question was addressed in Reda, to subject the amendment power to the Imperial duty.

Mr Simmonds then places considerable reliance on Prudential. To repeat what I said in the second rectification judgment:

In moving the camera forward in time, Newey J considered the development of the obligation, drawing adistinction between the employment law context and the pensions context, the two streams as Mr Simmonds describes it. He certainly considered the two separately, considering, in [122] to [128], the following cases: Stannard v Fisons Pension Trust Ltd [1991] PLR 225, National Grid Company v Laws [1997] PLR 157 (Robert

Malik recognised the emergence of the implied contractual obligation of mutual trust and confidence as a sound development; as Lord Nicholls (with whom Lords Goff and Mackay agreed) said the implied obligation was required "if the employment relationship is to continue in the manner the employment contract implicitly envisages". And, as I have already noted, an objective approach is required. Newey J referred to Gogay for its endorsement, if further endorsement were required, of the existence of the implied obligation and to Hale LJ's description of the test as a "severe one" such as to "destroy or seriously damage the relationship".

In Johnson v Unisys at [24], Lord Steyn described the Imperial duty as "an employer's obligation of fair dealing". He also noted that Imperial "did not involve trust law and the employer was not treated as a fiduciary"; his view was that the case was "decided on principles of contract law".

Newey J derived further guidance (see at [133]) about the scope of the contractual duty from a passage in thespeech of Lord Hoffmann (with whom Lords Bingham and Millett agreed), at [43]. In this passage, Lord Hoffmann indicated how judicial creativity might be able to provide a remedy in a case such as Johnson:

In Eastwood, Lord Nicholls (with whom Lords Hoffmann, Rodger and Brown agreed) said that the implied contractual duty means:

And Lord Steyn referred to a power to suspend an employee being exercised "with due regard to trust andconfidence (or fairness)".

Newey J also referred to two other cases concerning the payment of discretionary bonuses to employees: Clark v Nomura and Keen v Commerzbank AG [2006] EWCA Civ 1536, [2007] ICR 623. In Clark v Nomura, Burton J rejected the tests of capriciousness, on the one hand, and absence of reasonable or sufficient grounds on the other. He considered that the correct test was one of irrationality or perversity (of which capriciousness would be a good example), adding that this meant "that no reasonable employer would have exercised his discretion in this way". He regarded this as the same test as that applied in what is now the Administrative Court.

In Keen, Mummery LJ made a number of observations about the provision of information in relation to the award of bonuses. The relevant passages are set out in [136] of Newey J's judgment and I do not repeat them here.

It was on the basis of the employment law cases that Mr Rowley, for the beneficiaries of the pension scheme inPrudential, argued that the Imperial duty involved an obligation to act fairly. He relied not only on Johnson v

Counsel for Prudential (as it happens, Mr Tennet) put the matter very differently. He recognised that the Imperial duty meant that Prudential did not have an absolute discretion and accepted that an irrational or perverse decision could be attacked. The conduct in question must be serious: where a member remains an employee, the conduct must, viewed objectively, be so destructive of the relationship of trust and confidence that the employer may be taken to have repudiated the contract of employment.

In his discussion of the cases starting at [142], and after referring to the references to fairness and fair dealing inJohnson v Unisys and Eastwood, Newey J said this:

In [143], Newey J noted that in Imperial¸ Browne-Wilkinson V-C had said that the Imperial duty could be breached if an employer had acted "capriciously" but that with discretionary bonuses a slightly different test had been adopted, a test of irrationality or perversity. He agreed with Mr Tennet that an irrational or perverse decision "by an employer in a pensions context is likewise capable of offending the obligation of good faith". He referred to the judgment of Potter LJ in Horkulak (a discretionary bonus case where no issue of expectations, reasonable or otherwise, was in issue):

And so Newey J concluded in [143] that a power to increase a pension similarly requires "a genuine and rational,as opposed to an empty or irrational, exercise of discretion". This was something different from the application of an objective standard of reasonableness as he said in the opening words of [144].

Now, it is to be noted that Burton J in Clark v Nomura had expressly adopted his version of the test of irrationality or perversity which should apply in that case. He did not simply say that irrationality or perversity would be sufficient although it would have been sufficient on the facts. It is clear that the focus of Newey J's judgment was on the same test. Not only is that the test which he identified at the end of [143] and contrasted with an objective standard or reasonableness at the beginning of [144], but he also considered other submissions relating to the Imperial duty, testing them against the test of irrationality or perversity: see [146] (saying that there might be cases where it would be irrational or perverse to override expectations which the employer had engendered; and acknowledging that the entitlement of the employer to take into account his own interests must limit severely the circumstances in which a decision could be said to be irrational or perverse), [147] (contrasting a fiduciary power), [148] (the manner of internal decision-making may shed light on whether an employer has acted irrationally or perversely) and [149] (even an irrational or perverse decision may not necessarily breach the Imperial duty, for instance if it related to a trivial matter and would not seriously undermine the relationship of trust and confidence).

Inevitably, however, the researches of counsel in the present case have revealed one further authority on whichMr Simmonds relies and now is as good a place to mention it as anywhere. It is the decision of the Supreme Court of New South Wales in Re KCA Super Pty Limited [2011] NSWSC 1301 ("KCA"). There are similarities with, as well as differences from, the present case. There was an application by the trustees of the relevant scheme for directions as to whether or not it could give consent to a determination by the employer to reclassify existing DB members as DC members. In the absence of consent, the employer had indicated that it would wind up the scheme. The purported justification for the changes (which have parallels in the present case) were (i) to eliminate fluctuations in the employer's financial results directly due to the uncertainty of funding DB arrangements (ii) to reduce administration costs (iii) to ensure that all employees enjoyed consistent employment conditions and (v) to align the scheme with "global best practice, both within KC [the employer group] and elsewhere, in providing [DC] arrangements". This was, again, not a case where any issue of expectations arose.

Brereton J in referring to the winding-up provisions of clauses 48 and 49 of the scheme's trust deed and to thecontracts of employment under which employees agreed to be bound by the deed said this at [71]:

And so Mr Simmonds submits that there are clear similarities between that case and the present case. Inparticular, he says that it is clear that, regardless of what the employment-law cases might say about the effect of reduction in an employee's pay, the exercise of a (non-fiduciary) power to cease DB accrual, in circumstances where the purpose of such an exercise is to improve the employer's financial position on an on-going basis, will not constitute a breach of the Imperial duty.

I now return to the first prong of the two-pronged approach and the five elements identified at paragragh 386 above and which I deal with in turn.

There is no limit to the type of conduct which is capable of being destructive of the relationship of trust and confidence. I agree with that as a general proposition provided that it was remembered that the conduct must be such as to impinge on the relationship of trust and confidence and that the impact must be assessed objectively: see the passages from Malik quoted at paragraph 366 above. So any type of conduct is capable, in principle, of being relevant. But that tells one nothing about the test to be applied in assessing whether conduct of a particular nature does, in fact, breach the Imperial duty.

The duty to maintain trust and confidence applies to the exercise of powers under a pension scheme in the same way as it applies to any other conduct by an employer. If by that it is meant that, once one has identified the correct test, there can be a breach of the Imperial duty when that test is applied to the exercise of power under a pension scheme, this proposition is correct but it is trivial because the Imperial duty is, of its essence, directed at the exercise of such powers. If by that it is meant, in contrast, that the content of the Imperial duty is the same as the content of the implied duty of good faith as a matter of employment contract law, the discussion which I have already conducted shows, in my judgement, that this is not so. One cannot blindly apply employment law principles or cases. Indeed, if it were possible to do so, it would be necessary to identify which of the several strands of employment law is to be applied. Clearly the strand which suggests that fairness is the governing feature or a feature of real importance is to be rejected; but another strand, such as that demonstrated in Clarke v Nomura may provide a real parallel. That parallel, at least, would have the advantage that it related to an element of the remuneration of employees as do benefits under a pension scheme.

It does not matter that the conduct complained of falls literally within the employer's powers or that it is not otherwise actionable. Mr Tennet relies on what Lord Steyn said in Malik at 45 C-F. The duty to maintain trust and confidence will apply in cases even if as a matter of construction or otherwise there is no other implied limitation on the power which would prevent the behaviour complained of:

And so Mr Tennet submits that it is clear from this passage that conduct which undermines the relationship oftrust and confidence can, and often will, fall within the strict scope of powers vested in the employer (under the employment contract or pension scheme (see above)). However that plainly does not mean that the exercise of the power is not constrained by a duty to maintain trust and confidence. That is true up to a point.

It would difficult to maintain, however, that an express qualification of what would be otherwise be impliedleaves any room for the implication: in this context, see the discussion of Reda at paragraphs 399-400 above. And when asking oneself whether the term which is to be implied in a particular case has a content which precludes action which falls within the literal meaning of the power being taken, the nature of the power is, it seems to me, of central importance when it comes to establishing the scope of an implied restraint on its exercise. Thus I agree with Mr Simmonds when he says that the nature of the power is relevant. Further the fact that the employer's power is otherwise unfettered, either by an express constraint on the power or by making the exercise of that power subject to another's consent, is, as he submits, a relevant factor to take into account in determining the fetter imposed by the Imperial duty in that context. The terms on which the employees became members of the scheme, contained in the trust deed and rules and notified to them by way of employee handbooks, are also I consider relevant to the scope of the Imperial duty. Newey J clearly took this view when giving his reasons for saying that Prudential had not acted irrationally or perversely or otherwise in breach of duty. One reason (see at [186]) was "crucially, [the relevant rule] conferred on Prudential a discretion which was not subject to any express restrictions". This chimes with the approach of Brereton J in KCA.

Mr Tennet says it is only the start, but not the end, of the enquiry for the employer to show that what he hasdone or proposes to do falls literally within the scope of the power concerned. But that is also putting matters too high because it may very well be the end as well as the start of the enquiry. It is for the employee to show that the employer is in breach of his Imperial duty so the enquiry does not even begin unless and until the employee shows an arguable case that a breach of duty is involved. So although I agree with the principle articulated by Mr Tennet as stated above, it remains for the employee to show that an exercise of the power which, on its face, is compliant with the power gives rise to a breach of the Imperial duty. In assessing whether there is in fact a breach, the court will take account of the nature of the power and the absence of any fetters on the power. Indeed, if there are fetters on the power, the task for those alleging a breach of the Imperial duty may be harder than if there had been no fetters at all. The presence of fetters may be seen as at least some indication that the power is intended otherwise to be entirely unfettered.

Reasonable Expectations are relevant. I will be dealing with Reasonable Expectations in more detail separately. I note that they are relevant at this stage of my judgment.

It is not necessary to demonstrate irrational or perverse behaviour. That is Mr Tennet's position about the test for which Mr Simmonds contends. But he would say that, at the end of the day, it does not matter because

Be that as it may, the only relevant test, according to him, is whether the impact of conduct complained of has,objectively, seriously damaged the relationship of trust and confidence. How one describes the conduct is irrelevant. The behaviour of BCCI in the Malik case could not be described as irrational or perverse but it nevertheless represented a potential breach of the duty to maintain trust and confidence.

At first sight, that looks like a reasonable submission. But on analysis, it actually takes one nowhere. The realquestion is whether there has been a breach of the Imperial duty. The duty is not, however, a statutory duty and one must be very careful about construing the words of a Judge, however eminent, as if they were legislation. But even if one were to take the words "seriously damage the relationship of trust and confidence" as a quasistatutory reference point, it tells one nothing about how to assess whether the conduct complained of can be seen, objectively, as doing so. Instead, it is necessary to turn to again Judge-made criteria. Mr Tennet is right, no doubt, that it is not universally the case that there has to be irrationality or perversity in an employment law context as a matter of contract. Malik may be an example of that. But Malik was not concerned with the exercise of a discretion and gives no guidance about how the exercise of a discretion by an employer is constrained. Again, I suggest that Clarke v Nomura provides a more reliable touchstone than a case such as Malik.

Mr Tennet has referred to various paragraphs of the judgment in Prudential where Newey J referred to conduct which was "irrational or perverse or otherwise in breach of the obligations of good faith". Mr Tennet relies on those references to "otherwise in breach….." to support a submission that Newey J was not adopting irrationality or perversity as the test but only as examples of the wider, proper, test for which he contends. I addressed this point, but did not need to decide it, in the second rectification judgment at [19]. There are several paragraphs in Newey J's judgment where such a reference occurs: [161], [162], [163], [164], [169], [176], [178], [183], [185] and [187]. In each case, Newey J was expressing a conclusion that Prudential had not breached its Imperial duty.

Mr Tennet submits that Newey J's focus earlier in his judgment on irrational or perverse behaviour stemmedfrom the fact that he had already concluded that employees had no reasonable expectation engendered by the employer that the employer would continue to pay discretionary increases at RPI in future. So all that remained to be considered was whether the discretion could be said to have been genuinely exercised at all. He relies on the reference in [143] (see paragraph 361 above) to "a genuine and rational, as opposed to an empty or irrational, exercise of discretion". And so, he says, Newey J was plainly not suggesting that irrationality or perversity was the only relevant test for whether there had been a breach of the Imperial duty.

He also relies on [146], where Newey J said this:

There is reference in [124] to the decisions of Robert Walker J and the House of Lords in the National Grid case. Robert Walker J had stated that an employer was not prevented from looking after its own interests even where they conflicted with those of members and pensioners. Lord Hoffmann observed that the members had accepted, by the time that the case got to the House of Lords, that Robert Walker J had been right to take the view that the employer was "entitled to act in his own interests provided that he had regard to the reasonable expectations of the members": see [2001] 1 WLR 864 at [11]. I do not know the source of Lord Hoffmann's identification of Robert Walker J's view: so far as I can see, Robert Walker J did not anywhere in his judgment

Mr Tennet's reading of [146] of Prudential is that Newey J did not decide that members' expectations (even mere expectations) were irrelevant to the question of whether the employer had breached its duty to maintain trust and confidence; on the contrary, that paragraph makes clear that a failure to give effect to reasonable expectations engendered by the employer itself could amount to a breach of the duty to maintain trust and confidence. He levels the "gentlest of criticisms" at the use of the words "irrational or perverse" in the extract quoted as shorthand for conduct destructive of the relationship of trust and confidence, which is the actual legal test. He says that the use of the shorthand is, however, explicable on the basis of the facts of the case and makes complete sense if references to damaging the relationship of trust and confidence were substituted for the references to irrationality and perversity.

Well, it is often true that if one rewrites what a judge actually says, he or she can be seen as supporting the viewfor which the advocate is contending. I would put the matter the other way round and take Newey J as meaning precisely what he said. If one adopts Mr Tennet's substitution, one is thrown straight back to the questions of how you identify conduct which is destructive of trust and confidence and what is the test to apply. It is not that Newey J was using irrationality and perversity as a shorthand, he was using them as the test. It seems to me that [146] actually supports the conclusion that, whilst the overriding of expectations which an employer has engendered is capable of giving rise to a breach of the Imperial duty, the test for deciding whether it actually does so is irrationality or perversity in the sense explained at paragraphs 443ff below.

Mr Tennet submits that it appears from [185] that Newey J also accepted that members' expectations (evenapparently mere expectations) were relevant to the decision making process and to whether a breach of the duty to maintain trust and confidence had occurred. That paragraph includes the following:

The "very strong expectations" were those asserted in Mr Rowley's words as "generated by [Prudential's] longstanding policy, paying only lip-service to a policy dating back more than two decades".

I agree with Mr Tennet that [146] and [185] show that Newey J did not reject interests and expectations (inparticular Reasonable Expectations and perhaps even mere expectations) as being of potential relevance to the question whether conduct by an employer breaches his Imperial duty. Everything must be weighed in the balance. But he clearly saw that relevance as going to the question whether the employer's actions leading to the disappointment of expectations were irrational or perverse or at least something akin to that; that is what the balance is weighing. So while Mr Tennet may be right to say that Newey J held that the employees had no expectation engendered by the employer, all that follows from that is that irrationality or perversity could not be established on the basis of Reasonable Expectations being disappointed.

In my view, Newey J cannot have been intending to add to the content of the test of irrationality or perversity byusing the sweep-up words "or otherwise in breach of the obligations…". The use of those words is apposite to cover two matters.

It follows from this analysis that I consider Mr Simmonds to be correct, up to a point, in his submission that theemployment law cases do not provide the answer to the scope of the Imperial duty. I say up to a point because his submission as I understand it is, essentially, that they provide no useful guidance at all. If that is his submission, it goes too far. In particular, I do not consider that this particular submission justifies a conclusion that expectations engendered by an employer are necessarily irrelevant to the test for the scope of the Imperial duty.

Moreover, given that Burton J in Clark v Nomura was equating the test which he described as irrationality or perversity with the test applied in granting judicial review, as he put it "that no reasonable employer would have exercised his discretion in this way", it seems to me that breach of expectations is, at root, an aspect of irrationality or perversity. In other words, if expectations have been engendered by an employer, that may have been done in such a way that to disappoint those expectations would, absent some special change in circumstances, involve the employer acting in a way that no reasonable employer would act; in which case, irrationality or perversity, as those concepts are to be understood in this context, is established. To that extent, reasonableness does come into the picture. But this is not to bring in, by the back door, the test of fairness rejected in Imperial and cases following it since there is no question of choosing one person's (the employer's) idea of fairness rather than another person's (the employee's) idea of fairness. Rather, it is an objective assessment of where the range of reasonable perceptions reaches its limits.

Although one must heed Lord Hoffmann's warning in O'Neill v Phillips [1999] 1 WLR 1092 (the well-known case concerning unfair prejudice petitions in a company law context) about importing concepts from one area of law to another in an inappropriate way, I do not see any difficulty in using the concepts of irrationality and perversity, as developed in the context of public law, to identify the test for establishing the scope of the Imperial duty. That is the sense in which Burton J was using the words in Clarke v Nomura. And that is the sense in which Newey J was using the words too. That can best be seen from his reference in [144] to the decision of the Court of Appeal in Socimer Bank Ltd v Standard Bank Ltd [2008] EWCA Civ 116, [2008] Bus LR 1304. It is worth repeating part of the citation by Newey J from the judgment of Rix LJ:

That passage is instructive because it shows that honesty, good faith and genuineness, on the one hand, andarbitrariness, capriciousness, perversity and irrationality are two sides of the same coin in the context of employer/employee (and I would say employer/member) relationships. Adopting the approach of Burton J (the "no reasonable employer" approach) to irrationality and perversity, an absence of good faith as properly understood can be seen to be the presence of irrationality or perversity in this sense since an employer could not, in good faith, act in a way in which no reasonable employer would act.

In my judgement, the test of irrationality or perversity in the sense in which I have described it is the correct testto apply in relation to the scope of the Imperial duty so far as that duty is relevant to the exercise of a discretion by an employer under a pension scheme.

But lest there be any misunderstanding about what follows from my conclusion, this does not mean thatReasonable Expectations have no part to play. Not only must they be brought into the balance, they may, on the facts of a particular case, be critical in the sense that conduct which disappoints such Reasonable Expectations will amount to a breach of the Imperial duty.

It is an uncontroversial proposition that an employer is able to take account of his own financial interests indeciding how to exercise the powers vested in him without thereby putting himself in breach of his Imperial duty. Browne-Wilkinson V-C himself acknowledged that in Imperial. It is straightforward to accommodate that proposition if the test is as I have identified it: it could in some circumstances be irrational or perverse for the employer to give precedence to its own financial interests rather than to the Reasonable Expectations of members although in other cases (eg radically changed financial and economic conditions) it may be entirely reasonable, on any view, to depart from those Expectations. It is not so easy to accommodate the proposition within any other test since whatever the test is, it has to be stated in terms by reference to which the employer may, or may not, properly give precedence to his own financial interests rather than the Reasonable Expectations of members. Indeed, it becomes something of an artificial exercise to attempt to identify such a test: the question whether expectations are Reasonable Expectations or mere expectations must be intimately tied up with what constraints can be objectively established on the extent to which the employer is entitled to take into account his own financial interests. These considerations support, in my view, the conclusion which I have reached about the test to apply and the rejection of Mr Tennet's two-pronged approach to the Imperial duty.

The Imperial duty is expressed as a duty on the employer. There has been more than a hint in some of the submissions which have been made on behalf of IBM that Holdings cannot have been in breach of its Imperial duty since its management acted at all times in entirely good faith and rationally. Many of the criticisms of IBM as a whole are made in relation to what are alleged to be misleading financial statements and lack of relevant information. These, even if justified (it being IBM's case, of course, that they are not) are to be laid at the door of the team at CHQ and not at the door of Holdings or its management. The conclusion which IBM draws is that there can have been no breach of duty on the part of Holdings since it cannot sensibly be said that Holdings' decisions, in the light of its own knowledge, could be seen by the members as amounting to conduct which would destroy or seriously damage the relationship of trust and confidence between them.

It is more appropriate to deal with this aspect of the case in my discussion of Project Waltz rather than to addressit as a point of law to be applied to the facts of the case. But, jumping ahead, I should say that I do not consider that there is anything in the point. For reasons which I will come to, I do not consider that Holdings can shelter behind a business case (if justification is needed at all) based on the need to meet targets imposed by CHQ unless, in turn, a business case can be demonstrated justifying the imposition of the targets. So although the Imperial duty is owed only by Holdings, the requirements of CHQ can nonetheless result in Holdings being in

As to the question, to whom is the Imperial duty owed? it is right to note that the authorities have all dealt with a duty owed to the members and beneficiaries of a scheme; and insofar as the cases deal with expectations at all, they deal with statements and conduct which act directly on the members concerned. In the present case, however, some of the statements and presentations on which the RBs rely were made not to the members (or any class of the members) but to the Trustee. Those statements and presentations were, in turn, relied on by the Trustee in making its decision to agree to the amendments required to give effect to Ocean and Soto. More importantly for present purposes, much of the content of those statements and presentations was communicated to the employees: to the extent that content gave rise to Reasonable Expectations, they are, in my judgement, to be attributed to Holdings just as much as statements made directly on behalf of Holdings, for instance in the Webcast and in roadshows. I can see no relevant difference in principle in the context of the engendering of expectation between statements made directly to the members and statements made directly only to the Trustee but which were then communicated by the Trustee to the members with the knowledge and approval of CHQ. Reasonable Expectations

It is convenient, at this stage, having identified the test, to say something more about Reasonable Expectations.

Although Mr Tennet suggests that it is self-evident that failing to act in a way that you have led people tobelieve you will act is going to be destructive of their trust and confidence in you, he recognises the difference between a Reasonable Expectation and a mere expectation as explained already. His case is that an employer's failure to act consistently with Reasonable Expectations which the employer himself has engendered, in relation to a matter as central to the employment relationship as pension provision, is a paradigm example of a situation that is likely to impact on the relationship of trust and confidence. This is particularly so in the circumstances in which employees are relying on the employer to guide and inform their decision making in relation to future pension arrangements as in the present case. Dishonesty or recklessness is not needed to establish breach of the Imperial duty in relation to statements made; there is no reason in principle why trust and confidence cannot be undermined by making a false or misleading statement without reasonable cause or negligently.

He relies on the decision of the Court of Appeal in French v Barclays Bank [1998] IRLR 646. In that case, Mr French had been asked by his employer, the bank, to relocate and had accepted an interest-free "bridging loan" to enable him to move before his old house could be sold. After 6 months (and at a point where he could not sell his old house save at a substantial loss) the bank sought to charge interest. It was held that the attempt to require interest to be paid while Mr French was genuinely attempting to market his former home was a breach of the loan contract. But quite apart from this, it decided that the attempt to withdraw the interest free loan and charge interest was a breach of the bank's duty to maintain trust and confidence under Mr French's contract of employment. This was because of Mr French's expectations as to the nature of the loan transaction in the context of the bank's relocation request. Waller LJ said this at [46]:

That was an application of established principle to a particular set of facts. It seems to me that it would pass thetest (in the sense of giving rise to liability) of irrationality or perversity. I find it of no assistance in the present context other than to go this far with Mr Tennet: conduct on the part of an employer which is contrary to a Reasonable Expectation is, in principle, capable of giving rise to a breach of the Imperial duty on the basis that the conduct is irrational or perverse. I will be returning to this case later, since Mr Simmonds says that it actually

The way in which Mr Tennet articulates the difference between a mere expectation and a Reasonable

Mr Simmonds attaches a great deal of importance to the distinction which, to use his different language, is:

There is some underlying commonality in the way Mr Tennet and Mr Simmonds articulate the distinctionswhich they draw. But the two distinctions are not identical. Mr Simmonds is concerned only with expectations engendered by conduct on the part of the employer: the distinction is between (i) an expectation engendered by a representation as to intention and (ii) an expectation engendered by a representation amounting to a promise, commitment or guarantee as to future conduct or events. He takes it as a given that an expectation which is not engendered by the employer – what Mr Tennet has called a mere expectation – is irrelevant in the context of the Imperial duty. Mr Tennet's distinction is between an expectation engendered by the employer (the Reasonable Expectation) and one which is not (the mere expectation). For him, Mr Simmonds' distinction appears to be no distinction at all.

Clearly, there is no inconsistency between the two distinctions: it is simply that they cover different ground. In my view, they are both relevant distinctions. The distinction drawn by Mr Tennet is not a particularly interesting distinction in the present context because it cannot be seriously contended that a mere expectation by itself can be the foundation on which a case alleging breach of the Imperial duty is built. It is only once there is some expectation which has been engendered by the employer that the question of breach of the Imperial duty arising out of expectation can arise (always excepting the truly exceptional case, an example of which is difficult to construct). The distinction drawn by Mr Simmonds then becomes important. If, on the facts of a particular case, the evidence establishes that the employer has made a representation only about his current intention, then, assuming that the employer was being honest about its intentions, the most that the members could expect (applying an objective assessment of what it would be reasonable to expect) would be that the employer would not change his intention without some rational ground for doing so. Or, to put it in terms of the test which I have concluded should apply, the members could succeed in a challenge only if it would be irrational or perverse of the employer to change its intention in the sense that no reasonable employer would do so.

But if the representation goes further than that, the members may be entitled to expect more. Whether theemployer's subsequent conduct, or proposed conduct, would be irrational or perverse in the context of the representation which engendered the expectation concerned, will then be heavily fact-dependent. It does not necessarily follow, however, that absent a statement or course of conduct expressing a commitment, promise or guarantee, there can be no breach of the Imperial duty. That conclusion is, however, one which Mr Simmonds urges on me. He contends that the distinction which he draws is fundamental for two reasons:

As to the first of those, he relies on a number of cases:

The first, second and fourth of those cases are examples of factual scenarios where a breach of the duty of goodfaith was found to exist in a context where there were particular factors - breach of contract, or a statement known to be wrong, or reckless conduct – justifying the conclusion. The third case emphasises the hurdle faced by a claimant relying on negligent statements. The fourth example also suggests that something beyond negligence or carelessness is necessary in order for there to be a breach of the implied duty, otherwise it would not have been necessary for the Judge to consider, and make findings about, recklessness. To that extent they support Mr Simmonds' approach. But it is important, I consider, to emphasise what the Judge, in the fourth case, saw as reckless – namely a statement made without any proper confidence that it is truthful and accurate.

Mr Simmonds also referred to the decision in Eyett v. University of Nottingham (No. 1) [1999] 2 All ER 437 at 443-444. Hart J there took the approach which Mr Simmonds now urges on me when considering the application of the Imperial duty to require employers to notify members of their entitlements under the pension scheme. In the present case, he submits that I should not allow the RBs to rely on the Imperial duty to establish a liability

In addressing these competing positions, it is important to note that Mr Simmonds' argument is addressed,essentially, at the question whether statements about the future falling short of commitments can give rise to a breach of the Imperial duty. The cases on which he relies have nothing to do with future conduct: they are all about representations of fact or of the consequences of choices. I accept the thrust of Mr Simmonds' submission insofar as it relates to representations of that sort. The making of a merely negligent statement which is then acted on is unlikely in the ordinary course to be such as to destroy or substantially damage the relationship of trust and confidence which lies at the root of the Imperial duty. If the negligence does not, of itself, give rise to a cause of action, it should not be possible to have recourse to the Imperial duty. In contrast, if the statement is reckless, it will quite likely give rise to a breach of the Imperial duty and as such be actionable when it has been acted on. The remedy may be monetary compensation or unravelling the actions taken by the employee/member. The appropriate remedy will be heavily fact-dependent.

However, that is not an end of the matter. Suppose that a negligent statement is made by an employer and that amember makes important choices in reliance on it. Suppose later that the employer seeks to make detrimental changes to the Plan; and suppose that, had the negligent statement not been made, the member would have made a different choice. Does the member have any claim against the employer based on a breach of the Imperial duty when it comes to the making of the further changes? In principle, I do not see why the member should not have a claim although the issue will be heavily fact-dependent depending, for instance, on the precise nature of the representation, on the time at which the member learned of it, on whether he would in fact have made a different choice if there had not been a misrepresentation and on the scope of the new changes. This, however, has nothing to do with expectations, reasonable or otherwise, about what the employer will do in the future; rather, if there is a breach of the Imperial duty, it arises from the consideration (in accordance with my view of the appropriate test applicable in relation to the Imperial duty) that no reasonable employer would make the changes proposed in the light of the misrepresentation it had previously made.

However, quite apart from that, a misrepresentation of fact or consequences such as that which I have referred tomight also carry with it a message about the future, and give rise to an expectation in the members concerned (or rather the objective, reasonable, member) that the employer will, or will not, behave in a particular way in the future in relation to the relevant pension scheme. If it does give rise to any such expectation, that is a consequence, to state the obvious, of what the communication says. It does not make any difference to the message conveyed and received by the communication whether it is correct, negligently wrong or dishonestly or recklessly given. In all of these cases, the same Reasonable Expectations, if any, will be engendered.

In all of those cases, a question may arise whether a particular course of action subsequently proposed or takenby the employer in relation to the scheme will give rise to a breach of the Imperial duty. The answer to that question must take into account all relevant factors. One of the relevant factors may well be that the misrepresentation relied on was not reckless but merely negligent or perhaps even wholly innocent. If it was wholly innocent, the employer may well be able to argue that its proposed course of action is not one which no reasonable employer would take. It would not, however, be right in my judgement to eliminate, a priori, the possibility of a breach of duty in such a case.

As to the second of the matters raised above (paragraph 458ii), Mr Simmonds suggests that anybody can tell thedifference between something that is a promise or a guarantee, and something that is not, something that is an expression of hope. It is true that that distinction is one which, as a matter of principle, is readily understood. But which side of the line a communication falls may, depending on the facts, be capable of hot disputation among lawyers. The distinction does not, in any case, provide an answer to whether there can be a breach of the Imperial duty in the context of the exercise of a discretion. It is not possible to rule out "a priori" the possibility of an employer engendering such strong expectations, not amounting to promises or guarantees, among the members about how it will exercise a discretion in the future, that to act contrary to those expectations would seriously breach the relationship of trust and confidence on which the Imperial duty is based.

Without trespassing at this stage too much into the facts, Mr Simmonds' position is that the evidence establishesonly an expression of intention by IBM and not a commitment, promise or guarantee. It is not open for the recipient of the relevant statement to treat the former as if it were the latter. This is more so where, as he says in the present case, the statement of intention is coupled with an express disavowal of any commitment or guarantee. I do not expect that Mr Tennet would dispute the proposition that a statement of intention, if that is all that is established, cannot be treated as a commitment, promise or guarantee. However, he does take a very different view of what the statements made actually amounted to, saying that they go far beyond a mere expression of present intention and give rise to Reasonable Expectations including one to the effect that future accrual of benefits would not be curtailed for a period which had not expired by the time Project Waltz was implemented. If I am right in the legal test which is to be applied, this means that the RBs need to demonstrate that IBM was acting irrationally or perversely in the sense in which those words were used by Burton J in Clark v Nomura.

Mr Simmonds notes that it is common ground (which I believe is the case) that the alleged representations – thatis to say the words and conduct on the part of IBM relied on by Mr Tennet to establish the Reasonable

One answer given by Mr Tennet is that it is hardly surprising that an employer's breach of a specific duty tomaintain trust and confidence is capable of being actionable in situations in which the employee would not have a remedy for negligent mis-statement or breach of contract. There would be no point in implying the duty if it only applied in situations in which there was already a cause of action. Mr Simmonds' riposte to that answer is to say that it is not an answer at all to the question why statements made to employees should be actionable, where they would not be actionable if made to anyone else, by reason of a duty to maintain trust and confidence which plainly would still have a purpose in many other situations even if it did not extend to statements which were otherwise non-actionable.

I do not find that riposte at all satisfactory. It is perfectly clear that the Imperial duty applies to the exercise by employers of discretionary powers under a pension scheme. Mr Simmonds attempts to deflect Mr Tennet's argument by saying that the Imperial duty is not without application even if statements by an employer can only give rise to a liability on the part of an employer when they are actionable as a matter of general law. But if that is right, it is not easy to see why the same could not be said in relation to any other situation in which the Imperial duty might be invoked. It would always be possible to rely, in any particular situation under consideration, on the Imperial duty having content in a different context and thus to justify, in that particular situation, the conclusion that the members only have the rights which they have apart from the Imperial duty; but taking the argument to its logical extreme, there would be no residual category where the Imperial duty but no other obligation would arise.

Quite apart from that, the answer to Mr Simmonds' challenge is surely that Imperial Tobacco itself demonstrates that, in a pensions context and by analogy with employment law, a new obligation is imposed on an employer in the exercise of discretionary powers. The content of that duty is, if I am right in my conclusions, established by reference to the test of irrationality or perversity. If, in a particular case, the facts which establish irrationality or perversity also establish a remedy grounded in contract or estoppel or misrepresentation, so be it. But that is no reason to impose a requirement that there can be a breach of the new duty in the case of expectations engendered by representations only where there is already a remedy under the general law.

Mr Simmonds made a point in his written closing submissions about the way in which different beneficiariesunder the Plans should be treated. He submitted that it would be quite wrong to use the Imperial duty to allow all employees (and former employees) to benefit in the same way from statements made to all of them, without evidence of what those employees knew and of what they understood from these statements. Drawing an analogy with group estoppel cases, he correctly observes that there must be evidence that is sufficient to establish what information would have been received by the general body of members; and if different classes receive different information, they may have different rights (and are customarily separately represented in court proceedings). I perceived in that submission the suggestion that the subjective understanding of members was relevant to establishing a breach of the Imperial duty on a member-by-member basis. But that is not Mr Simmonds' case as will appear in a moment.

In contrast, Mr Tennet relies on something which Lord Nicholls said in Malik. The employer had submitted that unless the employee's confidence is actually undermined, there is no breach. Lord Nicholls said this:

And so Mr Tennet submits that the court's task is to ascertain objectively what expectations members wouldreasonably have been given by IBM's communications in 2004 and 2006. This submission draws no distinction between two matters. The first matter is whether Reasonable Expectations on which a member may rely are established. The second matter, assuming that some Reasonable Expectations are established, is whether IBM's conduct was inconsistent with those Reasonable Expectations and if so whether there was a breach of the Imperial duty. The objective standard referred to by Lord Nicholls is, I consider, clearly relevant to the second matter. Whether the conduct of IBM breached the Imperial duty must be judged objectively.

However, before a member or any other person can assert a breach of duty by IBM, he has to establish aReasonable Expectation on which he can rely. I accept that, by analogy with the group estoppel cases, a clear representation can be relied on by the membership as a whole and it is not necessary to produce evidence from each and every member affected that the representation was made to him or her. But in such a case, there needs to be, I consider, strong and clear evidence to establish the estoppel – evidence of a clear representation for instance – in relation to the general body of members. This reflects what Lewison J said in Trustee Solutions Ltd v Duberry [2007] 1 All ER 308 at [50(ii)]. It is also the case, I think, that where distinct groups can be identified, this principle can apply to each group separately. Thus, where different representations are made to different groups, the evidence needs to be of a clear representation to the general body of members of that group.

There remains one further aspect to mention in relation to this aspect of the case. It is whether it is right to applyan objective approach in ascertaining whether there is a Reasonable Expectation in the first place. For an employer to engender an expectation is not a breach of any duty – at least if the expectation is genuinely held by the employer. The communication giving rise to an expectation about future conduct does not – and is not of itself even likely – to damage trust and confidence. It is the subsequent conduct which disappoints that expectation which might give rise to a breach of the Imperial duty. Indeed, the very use of the word

I have reservations about whether he can do so. However, Mr Simmonds does not take this point. He concedes –indeed it suits part of his case positively to assert – that whether Reasonable Expectations are engendered is to be judged objectively. He expressly accepted in response to a question from me that (1) the members can rely on the objective meaning of the relevant communications even if their subjective understanding of them was different and (2) that the answer is the same whether the communications are received by a single member or the whole membership. That is common ground with Mr Tennet.

But, having made that concession, Mr Simmonds goes on to submit that I should only confer benefits on themembership as a whole if I find an unequivocal representation: in the absence of such an unequivocal representation, and in the face of the evidence about various different understandings on the part of members (as was shown by the evidence given by the various member witnesses), there is no basis for finding that every member who stands to benefit read or understood a representation in the way in which the RBs contend it should be read and understood. I do not accept that reasoning. If, objectively, a reasonable member would have understood the communications from IBM in a particular sense, it does not assist either side's case to demonstrate that there are members who understood the communications in a different sense (or did not understand them at all). I do, however, accept that an objective reasonable recipient of the communication will take into account the circumstances in which it was given and assess whether words which have an apparent meaning were actually to be taken as some sort of assurance by IBM that certain actions would or would not be taken in the future.

Mr Tennet makes a number of further submissions about the correct approach which the Court should take to theevidence. First, he says that Reasonable Expectations are "to be ascertained in the light of IBM's communications, not narrowly as a process of construction of those communications". He says that the Reasonable Expectations which members would have derived from communications must be determined by assessing those communications:

I accept the general thrust of those submissions although their application on the facts is not straightforward.They amount, perhaps, to no more than a list of some of the factors applicable in the present case when I come to apply the general principle that the meaning of documents and statements must be ascertained in the context in which and the purposes of which they were produced or made. I would question, however, the example he

That is not to say that there is to be imputed to the members the sort of understanding which people fully versedin pensions law and terminology might gain from the communications having read documents carefully and having viewed the Webcast several times. As I said in my judgement in the Rectification Action:

Mr Simmonds submits that the reasonable employee will read all of the material provided to him in order toinform himself about a particular event, particularly where he is specifically directed to that material as being relevant to that event. Mr Tennet suggests that this is simply an unrealistic assessment. He says that members cannot be expected to read everything, for instance, as he put it, "to wade through all 236 Project Soto Q&As, or all the UK Forum minutes made available on various websites". He points out that Mr Heath, who presided over the Project Ocean and Project Soto communications, accepted this reality: Mr Heath said in his oral evidence that he would expect members to place more importance on the key announcements, and that he knew that not all members would read the Q&A documents.

It is common ground that the reasonable employee will read the documents as a whole, and not cherry-pick ortake out of context phrases that suit his argument and ignore those that do not, although Mr Tennet contends that the flavour of the communications taken as a whole is clear and is such as to give rise to Reasonable

There is force in Mr Tennet's submissions concerning what the reasonable member will read or watch and theattention which he will give the material as a matter of practical reality. I am concerned, however, not with what different actual members might have done: one can speculate (with confidence I think) that some members would have not read or viewed anything, that others would have looked at the key documents with varying degrees of attention and that others would have read all the material assiduously. However, in relation to the issue of Reasonable Expectations, it is what the hypothetical reasonable member would have done which is relevant. That member must, it seems to me, be a person who is concerned to inform himself about the content of communications about pensions and the impact on him for the future. He would listen to (perhaps more than once) the Webcast; he would read written communications devoting enough time and attention to form an understanding of them. He would not simply skim though them saying to himself that he had no time understand them properly. Further, even if he thought that he understood the key documents, he would know that there was ancillary material (such as the Q&As). He would give some attention to that ancillary material and, although he may not be expected to read every word, he would devote enough time them to establish whether particular aspects of importance to him had been dealt with.

Part of the context in which the communications are to be assessed includes IBM's Business Conduct Guidelinesand its Core Values document and in particular:

I have no doubt that the guidance and values apply to IBM itself. I agree with Mr Riley who accepted in crossexamination that "IBM's conduct within the employment relationship should be assessed by reference to the standards IBM itself set for its employees" and "it would be perverse for IBM to require adherence to such conduct guidelines for its employees but not to adhere to those guidelines itself".

I am of the view that these express principles are to be taken into account in assessing whether IBM has been inbreach of its Imperial duty. Mr Simmonds is rather dismissive of any argument which relies on these principles. At the very least, however, they are surely capable, at the margins, of rendering conduct by IBM a breach of its Imperial duty when, absent these principles, there would be no breach of duty; and for my part, I attach more weight to them than that.

Mr Tennet submits that IBM knew that there was a perception among members, before each of the Soto andWaltz changes were effected, that the Plans would be kept open to future accrual for the longer term and, as I understand his case, that no detrimental changes would be made in the foreseeable future to IBM's early retirement policies. IBM says that that perception was incorrect and, if there was such a perception, it was not engendered by anything which IBM said in its Ocean and Soto communications.

Mr Tennet, in contrast, says that it was incumbent on IBM to correct any misapprehension of which it knew,especially since, on the RBs' case, the misapprehension will have arisen from IBM's own communications to members; an employer cannot excuse itself for giving out confusing or misleading messages by requiring the employees on the receiving end of those communications to make further enquiries.

I am not sure that he is correct when putting the matter so generally and widely. But what I do think is correct isthat if IBM was to comply with the Guidelines and Values, it would be incumbent on it to correct any widely held (or perhaps even not so widely held) misperceptions of which it knew about IBM's intentions in relation to the Plans before implementing Project Ocean or Project Soto. In that way, the membership would be given the opportunity to make informed decisions in relation to the Plans.

As I have already recorded, IBM's position is that the evidence of individual members about what theyunderstood the various communications to mean and their understanding about continued accrual of benefits under the Plans is not of relevance: what the Court should be concerned with is the hypothetical reasonable member. Enlarging on that, Mr Simmonds draws what I think is a useful analogy with survey evidence going to the issue of confusion in trademark cases. The "average consumer" is a legal construct. The approach to survey evidence was considered by the the Court of Appeal (in particular in the judgment of Lewison LJ) in Marks and Spencer plc v Interflora Inc [2012] EWCA Civ 1501. His approach is reflected in another appeal between the same parties in Marks and Spencer plc v Interflora Inc [2013] EWCA Civ 319. More recently, there has been yet further consideration of survey evidence in the context of passing-off: see Zee Entertainment Enterprises Ltd v Zeebox Ltd [2014] EWCA 82. In the first Interflora appeal, Lewison LJ quoted from earlier judgments of Morritt LJ and Chadwick LJ in Bach and Bach Remedies Trade Marks [2000] RPC 513:

And so Mr Simmonds submits that, in the context of the Imperial duty, I am not greatly assisted in determining the response of the reasonable employee by repetitious evidence of the subjective response of individual members when it comes to assessing the impact of what IBM told its employees.

Mr Tennet accepts that such evidence cannot be determinative, but submits it provides helpful insight into thecontent, prevalence and strength of member expectations. I think this is really only a forensically attractive way of saying that the attributes of the reasonable member should reflect what a number of ordinary (and apparently reasonable) members actually understood. Mr Tennet makes three submissions in support of the relevance and helpfulness of member evidence:

It is necessary to have some understanding of the US accounting requirements in relation to pensions costs. Theyare complex and to some extent counter-intuitive, at least to someone versed in the approach to pensions costs in the UK. Both of the accountancy experts have included sections in their reports explaining how the rules work, or are meant to work. For the moment, I set out as brief a description as I can to give a very broad picture (which at almost every detailed level I suspect is not entirely accurate). I have drawn heavily on Mr Robbins' report. That is not to say that I consider Mr E Bradley Wilson's description to be any less accurate, but I have found it difficult to attempt to merge their two descriptions without a very great deal of (unnecessary) work. What I say is intended to be, and I hope is, uncontroversial.

The accounting "cost" of a DB plan that is recognised in the accounts in each year is a calculated accountingfigure, designed to accrue the cost of employee pensions payable over the approximate service life of employees

Under US GAAP, a company is required at each financial year-end, to undertake a valuation of:

Prior to the Financial Accounting Standard, FAS 158, the accrued or prepaid pension cost that was incorporatedinto a company's balance sheet was equal to the accumulated difference between past NPPCs and past employer contributions (ie total employer contributions paid, rather than the fund value of the plan). A prepaid pension cost (a balance sheet asset) arose when the total past plan contributions exceeded the past NPPCs. An accrued pension cost (a balance sheet liability) arose when the past NPPCs exceeded the total past plan contributions.

There may have been additional amounts that the company was required to reflect on its balance sheet in certaincircumstances. These additional amounts did not affect the company's current or future I&E statements.

Effective for listed companies as of fiscal years ending after 15 December 2006, FAS 158 requires companies torecognise on the balance sheet the funded status of their plans (ie the excess or shortfall of the plan's assets over the PBO).

There are two themes of US GAAP pensions accounting that are relevant to the present case: attribution andsmoothing.

The NPPC is made up of the following principal components: i) Service Cost; ii) Interest Cost; iii) Expected Return on Assets; iv) Amortisation of Unrecognised Prior Service Costs; and v) Recognised Actuarial Gains and Losses.

The Service Cost is the actuarial present value of benefits earned by plan members for their service over the accounting period. This value is calculated based on several assumptions. The most significant assumptions are for the discount rate, salary increases, pension increases after retirement and mortality rates. The discount rate used for the calculation is the corporate bond rate used for the calculation of the PBO.

The Interest Cost is the increase in the PBO due to the passage of time. The conceptual basis for the figure is that it represents the increase in the actuarial value of the benefits under the plan as a result of the benefits being one year closer to payment. As the payment of liabilities becomes closer, there are fewer remaining years of discounting, resulting in an increase in the present value of the liabilities. The interest cost over a year is calculated by the discount rate at the start of the year multiplied by the PBO. This can be described as the unwinding of the discount in PBO. There are three points to make in relation to PBO:

The Expected Return on Assets ("EROA") is a credit component within the NPPC. It is the return that a company would expect on its pension plan's assets over the accounting period. This is the product of the company's expected long-term rate of return on assets ("the EROA assumption") and the MRVA, thus the higher the EROA assumption and the higher the MRVA, the higher the credit to NPPC. This component of NPPC is not affected by actual asset returns earned over the year. The conceptual basis for the figure is as a projected annual increase in value from the plans' assets; Therefore it acts as a credit figure (ie a cost reduction) in the calculation. Mr Simmonds draws attention to the following points with which I agree:

The Amortisation of Unrecognised Prior Service Costs is the cost to be recognised in a particular accounting period in respect of Prior Service Costs. Prior Service Costs (or credits) occur when plan amendments are made which result in an increase (or reduction) in benefits already accrued in respect of service rendered before the plan amendment. In general, FAS 87 permits the cost to be amortised over the future working lifetime of the plan members affected by the amendment. (This is an example of the attribution feature discussed above.) In the case of the Project Soto changes, the benefits were initially amortised for the period after Project Soto, but the cessation of accrual in the Plans as a result of Project Waltz meant that most of the PBO credit had to be recognised in a single year. This is the "curtailment gain" referred to below.

Recognised Actuarial Gains/Losses are the gains/losses recognised in a particular accounting period. This is an important concept in the context of the volatility of NPPC relevant to the present case. Actuarial gains/losses experienced by a pension plan can fall into two categories:

Mr Simmonds adds the following comments with which I agree. The accounts for the Plans show a running tallyof unrecognised actuarial losses and gains. The phrase "actuarial gains and losses" is correct but misleading: the tally is not confined to changes in actuarial assumptions but includes what are in practice real world changes (such as falls in the Plans' assets), as measured against the way they have been recognised for the purposes of the calculation. For the purposes of this case, the most significant additions to the tally are: falls and gains in the value of the Plans' funds; increases in PBO due to movements in the corporate bond market (and therefore the applicable discount rate) and any changes to actuarial assumptions (eg potential changes in longevity assumptions or similar). Where the running tally exceeds 10% of the greater of PBO or MRVA (known as "the corridor"), the quantum of the tally in excess of the corridor is amortised and recognised in the accounts. The relevant amortisation period is prima facie the average future working lifetimes of the active members of the plan.

There is an asymmetry between the way in which changes in PBO and asset movements are recognised in thefigure to be amortised: changes in PBO (eg as a result of movements in corporate bond yields or actuarial assumptions) are added to the figure which is amortised immediately; whereas changes in asset value feed into the figure which is amortised over time. I agree with Mr Simmonds when he says that including this is one of the reasons why even investment in a portfolio which tracked the discount rate for PBO exactly (ie a UK corporate bond portfolio) would not eradicate (or even massively reduce) NPPC volatility.

Mr E Bradley Wilson separately categorised Curtailment gains or losses which are recognised decreases or increases in employee benefits accrued due to closure of the plan to future accrual. A Curtailment gain (or loss) is an occasional component of NPPC, affecting the I&E as a result of a curtailment effected by the company. Both a curtailment and a plan amendment can have the effect of reducing employee benefits for future service. The distinction is that a plan amendment (leading to a Prior Service Credit) arises from a reduction in benefits for future service, while a curtailment (leading to a Curtailment gain or loss) arises from a reduction of number of years of future service. Both a Curtailment gain (or loss) and a Prior service credit arise from some sort of

US GAAP permits companies to recognise immediately only a portion of these gains and losses through theNPPC. (This is an example of the smoothing feature discussed above.) Any unrecognised gains or losses are carried forward to future years, and augmented with any new gains or losses that arise, in order to determine the amount to recognise in NPPC in future years. There are various options available for amortising gains and losses. The above components are aggregated and reported as a single net amount, the NPPC, in the I&E statement in the company's financial statements.

In any year, to the extent that expected return on assets, any prior service credits and recognised gains exceedservice cost, interest cost, any prior service costs and recognised losses, there will be an I&E credit in relation to pensions.

Conversely, to the extent that service cost, interest cost, any prior service costs and recognised losses exceedexpected return on assets, any prior service credits and recognised gains, there will be an I&E cost in relation to pensions.

Under US GAAP, companies have flexibility in applying the above principles in a number of areas. The threeprincipal areas of flexibility are:

It is not necessary to go into any of those at this stage.

NPPC has featured large in the present action because, on the RBs' case, IBM's agenda was driven by a need tominimise the figure for NPPC in order to produce a particular level of earnings per share ("EPS"). It is that allegation which forms a main plank of the case.

There is no difference between the parties' accountancy experts as to how NPPC works. However, the RBs makeseveral references to a comparison between service cost and NPPC. Their proposition is that NPPC might be expected to be at least as much as service cost over a period. I do not accept that proposition. Instead I accept Mr Simmonds' analysis in the light of the expert evidence. It is in essence as follows, although more detail can be found in paragraph 290 of his written closing submissions and the footnote references. He contends that the relationship between serice cost and NPPC depends on the relationship between the expected EROA (deriving from the scheme's asset allocation) and the proportion which service cost bears to the PBO of the scheme.

Mr Simmonds concludes from the above that, in truth, any statement as to whether or not a scheme might beexpected to produce an income for NPPC purposes is a statement about the relationship between the scheme asset allocation (ie the EROA assumption) and the maturity of the scheme (the quantum of service cost). In this case, as the actuarial experts accept, there was a mismatch between liabilities and asset allocation, which was nevertheless within the spectrum for UK funds. In those circumstances, any statement that "you might expect NPPC to be at least equal to service cost" or "service cost was a bargain" or the various statements of that nature made on behalf of the RBs should properly be viewed as unfounded assertions masquerading as common sense. Whether it is right, as Mr Simmonds suggests, that to the contrary you would 'expect' a mature scheme invested with this mismatch to produce an NPPC profit, I do not need to decide. But what I do conclude is that the arguments raised by the RBs based on the proposition that NPPC cost can be expected to equal or exceed service cost over a period are unsound. The argument that NPPC should only become a problem for Holdings if it were consistently greater than its share of service cost does not address the relationship discussed in the preceding paragraph.

"Volatility" is a word which has featured a great deal in this litigation. It is, unfortunately, not always clear inwhat sense it is being used. That is even more so the case in relation to its use from time to time in various contexts – by CHQ and in its communications, by the Trustee and the UK team and in their communications with members. The word itself is not difficult to understand: it means, for present purposes, being likely to change suddenly and unexpectedly or unpredictably. The lack of clarity in the present case comes not from the word itself but from what it is being applied to. There are at least three different relevant aspects to which it has been applied. First, to the value of scheme assets and the amount of scheme liabilities; secondly to the gap between that value and that amount (sometimes referred to as "breadth of variance volatility"); and thirdly to NPPC. The second, the deficit, was a matter of concern to the Trustee. Volatility of that sort could, to some extent at least, be managed by appropriate investment policies and in particular matching (although in the real world, 100% matching is impossible). The third was the matter of greatest concern to CHQ because of the impact NPPC would have on EPS.

There is nothing to suggest that the meaning of volatility in any sense was ever explained to members. I canmake no assumption at all about what the reasonable member would have understood by volatility, although it seems highly improbable that he would have had any inkling about NPPC volatility, something which, speaking for myself, I had no understanding of before this case began. However, by the time of Project Soto (or at least during the course of it) the members of the Trustee board and of the UK team would have gained an understanding of NPPC volatility and that CHQ's concerns related to NPPC volatility.

One finds the use of the word volatility being used in relation to economic forces outside IBM's control in whichcontext the meaning of the word is tolerably clear. But that is not the sense in which volatility (of a type capable of being brought under some sort of control) is being used. As used within CHQ and in the communications between CHQ and the Trustee, volatility was always (there may be exceptions, but if so they are few and far between) used as referring to NPPC volatility. The UK team's paper prepared for the presentation to the Trustee on 8 December 2005 focused on NPPC volatility. Thereafter, my impression of all of the material is that everyone involved in preparing communications was using the word in the sense of NPPC volatility. As I have said, whether the reasonable member reading or hearing the few references to volatility would have appreciated the reference seems to me to be unlikely.

I have received lengthy submissions both in opening and in closing about Project Ocean. There is much that isuncontentious. My statements of facts in the following paragraphs are intended to be my findings of fact save where otherwise appears.

Mr Simmonds' opening written submissions focused, in his description of Project Ocean, on the financialposition of the Plans from a UK perspective. He did not say a great deal in the context of Project Ocean about the position in relation to US GAAP and NPPC (although he had much more to say about them in the context of Project Waltz and to some extent Project Soto). The RBs' skeleton argument focused immediately on this, noting that the effect of the 2002-2005 financial crash in world financial markets resulted in substantial falls in the asset values of the Plans. For purposes of US GAAP accounting, it was not just negative ROA that caused a problem;

The result was that in 2000, 2001 and 2002 the Main Plan incurred actuarial losses (namely a shortfall of actualROA from EROA) of £488m, £699m and £865m respectively, totalling £2,052m for the duration of the 20002002 crash. In 2003, there was the start of a recovery: the Main Plan began to make actuarial gains as ROA well exceeded EROA. These losses for 2000-2002 were, however, smoothed over the coming years, delaying their effects. The funding implications of the 2000-2002 crash was one the factors which led to the Ocean Changes.

Thus, it was becoming apparent by mid-2004 that the triennial actuarial valuation of the Plans as at 31 December 2003 would show a significant funding deficit in respect of past accruals (for which, given the balance-of-cost nature of the Plans, IBM would be responsible). In fact, the valuations (which were signed off on 24 December 2004) showed deficits of £900m in the Main Plan and £19m in the I Plan.

Further, there was financial pressure from CHQ on the IBM UKI businesses which were seen asunderperforming as compared with other parts of the worldwide business. This prompted the UK management team (led, at that time, by Mr Hirst, then the UK general manager) to put together a business recovery plan. As part of the review that this entailed, UK management began, in February 2004, to consider making changes to the design of the UK DB Plans. A "pensions project" team, led by Kevin Waller and also including Mr Wilson, Mr Newman and Mr Heath was set up to consider pension policy changes, including changes required by new Government legislation, increases to employee contributions, a policy for increases in pensions in payment and cost of living adjustments ("PIP" and "COLA") and questions of long term funding.

Given the system of "powers reserved" to IBM Corporation which I described in detail in my judgement in theRectification Action, CHQ became involved in the formulation of these possible changes at an early stage. The RBs put this rather higher than mere involvement. They say that even though the pensions project was initiated by the UK team, CHQ retained ultimate control over any pensions design changes. It is clear, in my view, that there was more than simple involvement. The "powers reserved" meant that CHQ had to approve rather than merely be consulted. In that sense, the RBs are correct to say that CHQ retained ultimate control; but it is also the case that CHQ did not impose its view in the sense of compelling UK management to take particular action. In practice, this was never an issue because UK management knew where the ultimate power lay and would follow CHQ policy and directions: there was no occasion on which they were faced with implementing something which they considered was wrong, in the sense of being against the interests of the UK companies, although I am bound to say that matters came pretty close to that later on in relation to Project Waltz. The practical recognition of this power structure is reflected in the meeting of the UK Pensions Project Team on 28 June 2004 where it was noted that approval from CHQ via Damian Glendinning (Director of treasury operations and pensions strategy in the Corporate Treasury Department at CHQ), and review by Paul Michaud (Program Manager, Global Benefits, CHQ), would be required for any proposals the UK team made, which might take some time.

In any case, the evidence shows that there was considerable involvement in Project Ocean on the part ofCHQ. Mr Tennet gives the following examples:

Before I come to the events rather closer to the adoption of the Project Ocean changes at the end of 2004, thereare some earlier documents which have been relied on by the RBs which I should deal with and which Mr Simmonds has addressed in his written closing submissions. They are relevant to establishing precisely what it was that the board of the Trustee, and in particular Mr Lamb, thought they were agreeing to in late 2004, especially in relation to the commitment alleged by the RBs that the Plans would be kept open.

The first document is a letter from Mr Lamb to Mr Hirst dated 31 July 2003. Mr Tennet described this letter inhis oral opening submissions in this way: "effectively Mr Lamb is seeking reassurances from IBM about its commitment". At this stage however, I consider that it is clear, as Mr Simmonds submits, that Mr Lamb's concern was with the funding of the past service deficit, thus he wrote:

Mr Lamb and Mr Hirst met on 10 September 2003 to discuss the concerns raised in Mr Lamb's letter. Followingthat meeting, Mr Hirst wrote to give some reassurance:

That exchange of correspondence says nothing about any commitment to keep the Plan open to further accrualof benefits. It may well have taken place in a context in which no-one was considering or concerned about the Plan being closed to further accrual. The point is that this exchange of correspondence cannot be read as giving rise to any increased expectation let alone a commitment that the Plan would be kept open: both before and after this exchange, there was simply an assumption that the Plans would continue because no-one had suggested they would not. This is consistent with what the minutes record Mr Lamb as having reported to the TMM on 23 October 2003 where he referred to IBM's intention:

Whatever Mr Lamb may have hoped, or even expected, there was, in my view, clearly no commitment given byMr Hirst that the Plan would remain open for any particular period or at all. The reassurance given by Mr Hirst related to funding, not to any future accrual of benefits. Moreover, Mr Hirst expressly stated that it was IBM's current intention to continue to support the Plan through the payment of contributions. No doubt the Trust Deed and Rules (together with statutory provisions) would have imposed some obligations on IBM from which it could not escape. But that does not lead the conclusion that IBM was obliged to keep the Plan open. I can detect nothing in the correspondence which could reasonably be taken as a commitment to keep the Plan open to accrual of future benefits either at the then current rate of accrual or at any other rate. To the extent that IBM could properly bring about a closure of the Plan to future accrual before this correspondence and the September meeting, it continued to be able to do so; the correspondence and the meeting did not constrain IBM's behaviour in that respect. I ought to note, in any case, that there is nothing to suggest that this correspondence or the contents of the meeting were communicated to members: it cannot therefore have engendered directly any expectation, reasonable or otherwise, in them. That is not to say that it had no impact on the members. The subsequent communications from the Trustee to the members were based on what the Trustee had been told and was capable of giving rise to expectations on the part of the members.

The RBs have also relied on an email from Mr Heath to a Mr Glazerman (a C Plan member) dated 21 September2004. It appears that there had been some press speculation about the future of the C Plan and that IBM was considering winding it up. Some members were concerned that IBM might do just that. Mr Glazerman drew attention to these concerns. Mr Heath's (draft) email in response to this refers to this speculation but does not mention member concerns. Mr Heath stated that IBM did not comment on media rumour and speculation but went on to say that it "may be helpful however to draw your attention to a statement contained in the most recent IBM Pension Members' Report" where Mr Lamb had commented:

In his opening submissions, Mr Tennet said that Mr Heath was addressing, in the passage just quoted, theongoing future of the scheme, not the bailing out of the Plans in the event of a wind up. Mr Simmonds submits that this is a misreading of Mr Hirst's letter because:

When giving oral evidence, Mr Heath confirmed that he was not stating that there would never be a change: "That did not mean that we would not address issues around the plan, as a company". I accept that that is so: but it has never been the RBs' case, and Mr Lamb has never suggested, that there would never be a change. Indeed,

The next significant document is an email from Mr Hirst to Mr MacDonald dated 18 October 2004. This e-mailrefers to consultation with employees. IBM accepts that the communications to employees were seeking positively to encourage members to accept the Project Ocean Changes, but it is submitted that it is important to see the way in which Mr Hirst explained how the communications process would seek to do this:

I agree with Mr Simmonds that this passage, read in isolation, did not involve any representation or commitmentregarding the continuation of future accrual of benefits in the C Plan. It is entirely consistent with IBM's case concerning the Project Ocean changes.

In May 2004 IBM Retirement Funds EMEA prepared UK pensions figures, showing actual and projected NPPCfor the UK plans. The assumption was that EROA would be 8% and the discount rate was 5.75%. The figures revealed that, owing to the 2000-2002 crash, UK pension costs (including DC costs) were already projected to increase by $84m between 2005 and 2006, and that total UK pensions costs were projected to be $237m in 2006:

Even though there was already a potential issue over the projected increase in NPPC for 2005 and 2006 andfuture service costs, UK management's attention was focused on the issue of scheme funding (and the deficit) at this time. In early July 2004 the new Scheme actuary, Watson Wyatt, produced preliminary valuation results for the plans as at 31 December 2003. The preliminary results indicated that there was a Main Plan deficit of £1.2bn on an ongoing basis, which would necessitate annual company contributions of around £244m, in addition to I Plan contributions of around £40-45m, and DC contributions. As set out in the Actuarial Valuation Report which was later produced on 24 December 2004, the funding position had deteriorated since 2000 because:

The preliminary valuation was carried out using what have been termed "central basis" assumptions, which inparticular set the pre-retirement discount rate at 6.8%, and the post-retirement discount rate at 5.25%. In calculating the post-retirement discount rate, the actuary assumed that retiree members' liabilities would be backed by an asset mix of 10% equities, 10% real estate and 80% bonds, although the actual asset mix of the fund was more heavily weighted towards equities. I do not propose to go into the technical reasons for this or into an examination of the Actuary's paper explaining his methodology.

In carrying out the valuation, Watson Wyatt used assumptions that were more prudent than the previous SchemeActuary, Aon, had used. Watson Wyatt's approach led to the reduction in the EROA assumption to 6.025%, which had the effect of decreasing the discount rate, and hence increasing the valuation of the DB Plans' liabilities. It is not suggested by IBM that Watson Wyatt's approach was unreasonable.

At this stage, I say something about asset allocation.Asset Allocation

Mr Lamb exhibited to his second witness statement a table showing the asset allocation policy (and actualallocations) at the year-end for each year from 2000. As at 31 December 2000, when the Main Plan had a surplus of £585m and a funding level of 118% on an ongoing basis, the asset allocation strategy in relation to the DB section of the Main Plan was 80% in equities and other assets (primarily property) and 20% in bonds. The Trustee considered, on the basis of expert advice, that this was an appropriate mix but recognised that, as the Main Plan matured, there would need to be a move towards bonds. The policy of an 80%/20% mix continued in 2001.

At a meeting of the Investment Committee held on 2 May 2002, it was decided (following expert advice) that,although there was no need for major action, the asset split ought to be changed – by way of a de-risking strategy whereby a greater proportion of the fund was invested in bonds – in order to reflect the increasing maturity of the Main Plan. In particular, it was decided to implement a 5% move from equities to index-linked bonds by means of a 1% move per quarter for five quarters. After the first 1% movement, however, it was decided (at the meeting of the Investment Committee held on 25 July 2002) to delay the next switch pending further examination of the appropriate ratio in the light of a volatile, declining investment market. The asset allocation policy as at 31 December 2002 was, therefore, a 79%/21% split. In 2003, further consideration was given by the Investment Committee to the appropriate split and it was intended that a full asset/liability modelling ('ALM') study would be carried out in 2004 following the valuation as at 31 December 2003. At the end of 2003, the policy remained a 79%/21% (although the actual split as at 31 December 2003 was 81%/19%).

On 8 December 2003, Mr Alexander (of Watson Wyatt) took over as scheme actuary from Mr David Eteen (ofAON). The first major task for Mr Alexander, and of Watson Wyatt as the Trustee's new actuarial adviser, was to prepare the formal triennial valuation as at 31 December 2003. It was during the preparation of this valuation that IBM and the Trustee became aware that the Plans were likely to show the substantial deficit which I have referred to and which, in turn, was a primary impetus for the Project Ocean changes, not least because of the effect it was likely to have on IBM's future contributions to the Plans.

By July 2004, Watson Wyatt had completed its first ALM study for the Main Plan, which was discussed on 22 July 2004 by both the Investment Committee and the TMM. As at that date, the asset allocation mix was 75%/25%. At about that time, Watson Wyatt had also produced a draft valuation as at 31 December 2003 that, contrary to AON's previous approach to such valuations, was based on (1) the market value of assets and (2) more conservative investment return assumptions. This draft valuation projected a deficit of some £1.2bn which Mr Simmonds describes as "alarming". Given this projected deficit, one of Watson Wyatt's concerns was the restoration of the Plans' solvency, which would require, in particular, significantly increased contributions from Holdings which, in turn, would require the Trustee to assess the strength of Holdings' covenant.

In July 2004 Watson Wyatt produced notes on the effect of various design options and on the effect of cashcontributions to the DB Plans, to assist IBM with its consideration of a solution to the problem.

Returning to pension costs, and putting matters neutrally, Holdings needed, if it was to reduce pension costs asCHQ clearly wanted, to address two issues namely (i) how to deal with past service deficits and (ii) how to reduce the costs of future service benefits.

Mr Simmonds explained IBM's approach to the solution in this way:

Mr Tennet undertook a rather fuller review of what happened. He explained that the preliminary valuationresults were immediately forwarded to Mr Greene, David Hershberg (CHQ Legal) and Mr Glendinning at CHQ on 5 July 2004.

Following that, Mr Glendinning described, in an email to Mr Wilson and Mr Newman on 15 July 2004, thediscussions he had had with Mr Greene:

An additional cash contribution would necessarily be at the expense of CHQ's share repurchase scheme, sincethere would be fewer funds available to repurchase shares. Mr Glendinning emailed Mr Wilson and Mr

In its pleading (the Response), Holdings seeks to characterise this email as an example of their general approachof seeking to balance the interests of members and shareholders when setting its preferred investment strategy. But the RBs do not accept that and say:

Mr Glendinning, Mr Greene and Mr Hershberg discussed some time in mid-July 2004 how IBM was to handlethe discussions with the Trustee. On 18 July 2004, Mr Glendinning sent to Mr Greene an email suggesting that, "rather than have someone from IBM UK argue the case", it would be a good idea for CHQ to draft a paper to distribute to the Trustee. The email contained a suggested draft which explained that the new actuary's approach represented a "fundamental shift in the planning for, and funding of, pension plans" and that the effect of the actuary's conservative approach reduced EROA from IBM's current assumption of 8% to 6.025%.

So what Mr Tennet draws from the above is that, rather than listening to the advice of the new Scheme Actuary,and taking the opportunity to reconsider the true liabilities of the DB Plans, CHQ's approach to the UK plans was to seek to influence the assumptions and investment strategy underlying them, with a view to keeping EROA and discount rates high, and hence NPPC and employer contributions down. I agree that CHQ was, indeed, concerned to keep NPPC and contributions down and, consistently with that, wanted to see what it regarded as more realistic assumptions and a less conservative investment strategy. This was an approach driven by commercial considerations and not, of itself, improper. I do not accept that this approach is to be categorised as a refusal, or even disinclination, to listen to the advice of the Scheme Actuary. Indeed, the advice of the Scheme Actuary was not only listened to but heard. What he said was not welcome. But there was nothing improper about IBM adopting the approach which it then did and to seek to persuade, provided it did so without improper pressure and without making misleading statements or providing false information, the Trustee to act in a particular way.

It is, I think, fair to say that these desires of CHQ – to achieve a greater investment in return-seeking assets andto obtain some control over the assumptions to be used for funding purposes – lay behind the idea of a holding company guarantee. It enabled an investment policy to be adopted under which the margins of prudence previously adopted in relation to EROA could be weakened or even eliminated. The result was, in the event, the adoption of the "best estimate basis" in place of the "central basis".

The Trustee in turn considered the position. Watson Wyatt's preliminary results were discussed by the TMM on22 July 2004. I am satisfied that the Trustee's aim was to work with IBM to find a solution to the funding problem and not to provoke IBM into extreme measures. I am sure that Mr Lamb is accurate when he says, as he does in paragraph 1.29 of his second witness statement:

It is easy to be wise with hindsight, of course. At that stage, it was the shared assumption of Holdings and theTrustee that the Plans would continue: closure or winding-up was not on the agenda. There is nothing to suggest that Holdings (I shall come to CHQ later) had given any thought to what would happen if their aggressive investment strategy failed to deliver in the short term any more than Mr Lamb had done so. In any case, the Trustee was secured in relation to past service, even if the Plans were closed or wound-up, as a result of the guarantee. I accept, nonetheless, Mr Lamb's evidence about what the Trustee would have done if it had realised that the Plans might be closed if the anticipated returns did not transpire. But I do not take that as evidence that the Trustee actually thought about the point at the time. It was an unspoken assumption that the Plans would not be closed.

The investment strategy agreed is not, however, therefore without significance in the context of ReasonableExpectations. The fact that the investment strategy was agreed as a long-term answer to a commercial requirement (I deliberately do not say a solution to a problem) can only strengthen the message heard by the Trustee, such as it was, about sustainability and putting the Plans on a firm footing for the future.

Further, although investment strategy was not something which was mentioned in member communications (sothat members cannot rely directly on assurances made to them on the topic), it does not follow that it was irrelevant to the message which was given to the members. As will be seen, Mr Lamb communicated the Ocean proposals to the members by a letter dated 9 November 2004. That letter was designed to inform the members about the Trustee's decision in relation to the proposals and why it had decided to support them. The fact that the Trustee supported the proposals would clearly have been a factor to be taken into account by the members in making their decisions and in their approach to the Plans for the future. Indirectly, therefore, communications with the Trustee which gave (if they did give) a message to the Trustee about future benefit accrual are part of the material to be taken into account in deciding whether subsequent actions by IBM UK amounted to a breach of the Imperial duty.

Test the matter this way. Suppose that Mr Lamb had been given an assurance by Holdings that it would not seekto argue that a failure of the investment strategy in the short term would cause it to act in relation to the Plans otherwise than it would have acted absent such a failure – an assurance too vague to give rise to a binding contract but sufficiently strong to show a very strong commitment on the part of IBM UK. And suppose that the Webcast (as to which see paragraph 613 below) and Mr Lamb's letter had taken precisely the form which they did. Can it be said that the assurance is simply to be ignored because (i) the Trustee itself has no cause of action based on it and (ii) the members were not told about it, so that the assurance could not have engendered any reasonable expectations? The answer to that question is, in my judgement, a resounding "No". To fly in the face of an assurance of that sort is, in my judgement, relevant to the issue of breach of the Imperial duty in the same

Expanding on what Mr Simmonds said about formulation of the proposals and the rejection of any suggestionthat the Plans should be closed, I note the following. Mr Waller presented the UK pension project team's proposals for changes to scheme design to Mr Hirst on or about 18 August 2004. Holdings' proposal was to increase employee contribution rates to the C Plan and I Plan and to reduce the accrual rate in the N Plan. The UK team dismissed the option of restricting pensionable earnings on the basis that it would not have a uniform effect on all members. Further, although it was actively considered, closing the plans to future accrual was not recommended by the team for a number of reasons. One reason, as Mr Simmonds himself acknowledged, was that such a course would be controversial amongst employees and detrimental to IBM from an employeerelations perspective.

Mr Waller's paper expressly relied upon Watson Wyatt's note of 12 July 2004 to show that the cost of accrualhad increased. That paper calculated the increase in the costs of accrual explaining that the reasons for the increase included allowance for more members to retire early. Further, there is nothing to suggest that COLA increases had been treated any differently from the past in ascertaining the contribution rate. Accordingly, Mr Tennet argues that the proposal to increase employee contribution rates, namely because the costs of accrual had increased, implicitly assumed that IBM would maintain its COLA and early retirement terms and practice. That must be correct; and it is certainly the case that everyone was proceeding on the basis that the Plans would not close and would provide further benefit accrual at the current rates. But that cannot be taken as a commitment that things would not change.

From late August 2004, CHQ engaged Towers Perrin to assist the UK pension project, ahead of the Trusteemeeting which had been set for 21 October 2004 at which it was intended by IBM that the proposals to increase contribution rates would be discussed. Towers Perrin, it must be emphasised, were retained by IBM; they were not advising the Trustee. Although they were asked to consider a whole range of redesign options, they were not instructed to provide advice "as if" they were the Scheme actuary. It had been Mr Glendinning's suggestion that they should provide "as if" advice, but Towers Perrin had said that this "would force them to be more conservative than is useful for our purpose", as Diane Gherson (Vice-President, Compensation and Benefits) told Mr MacDonald in an email dated 1 October 2004. Mr Tennet is right, I consider, to identify CHQ's aim as being to find a way to encourage Watson Wyatt and the Trustee to adopt a stance which was less conservative than the Scheme Actuary would ordinarily recommend.

I have already considered the genesis of the guarantee. As Mr Simmonds explained, the guarantee had the twofold advantage of enabling the scheme actuary to adopt the less conservative "best estimate" approach to the actuarial valuation and to postpone the date for completion of the move from equities to bonds. It would also, as Mr Tennet correctly noted, have a significant effect on company contributions by (i) encouraging the Trustee to adopt a more risky investment strategy and (ii) the removal of the margin for prudence in the assumptions used; but those two results came with an increased risk to IBM which carried the ultimate funding obligation (via IBM WTC), if matters did not turn out as hoped for.

I should mention at this stage a complaint which was made to the Pensions Ombudsman by certain members ofthe Plans. The complaint was that Holdings had failed to share the surplus in the Plans with members by awarding benefit increases, instead preferring to use the surplus to fund its own liabilities for contributions in relation to the DC Plans. In its submissions to the Pensions Ombudsman, Holdings had rejected the proposition that there was any link between surplus and its policy in relation to pension increases, pointing out that benefit increases had been awarded in the past even when there was an actuarial deficit and arguing that the decision to meet DC contributions out of surplus did not reduce pension costs but dealt merely with the question of cash flow.

The UK Pension Project Team was alive to these submissions. They considered the submissions at a meeting on2 September 2005 concluding that to fund the current deficit by stopping increases to benefits would create a link between funding and increases, invalidating the argument which had been presented to the Pensions

Holdings was therefore faced with a presentational problem in relation to the new proposals to increasecontributions as part of the solution to the past service deficit. At the start of October 2004, CHQ, which was itself aware of the submissions made to the Pensions Ombudsman, began to consider how the increase in contribution rates might be presented to members. Mr Michaud noted the problem in a draft note to Mr MacDonald at the end of September 2004.

Ms Gherson, too, noted the problem in an email dated 1 October 2004 to Mr MacDonald. She wrote: "Unfortunately, the idea of connecting funding status to employee contributions turns out to be problematic because of the many years when IBM had a funding holiday while employees made contributions". That is entirely consistent, and indeed confirms, the proposition that CHQ's concerns were about the funding deficit in the UK Plans and how this would impact on IBM Corporation's I&E. Ms Gherson's conclusion was that:

On 27 September 2004, Mr Lamb met Mr Hirst and Mr Wilson to discuss the guarantee and future companycontributions. There is no note of that meeting. It was followed by a letter to which I turn in the next paragraph of this judgment. Shortly after that, work began on producing a first draft of the guarantee. It is interesting to note (I say interesting because it shows who really called the shots) that CHQ was at this stage keeping open the possibility of ignoring the recommendations of the Scheme Actuary on contribution rates even if adopted by the Trustee. Towers Perrin had already indicated to Ms Gherson that if they had had to produce a valuation for the Trustee, it would be more conservative than the indications which they had previously given to CHQ; but they felt that IBM was entitled to base its funding decision on its own interests and the independent advice which it had received (ie from Towers Perrin).

Mr Lamb followed up that meeting and the provisional agreement with a letter dated 4 October 2004 to Mr Hirst.This explained that Watson Wyatt had agreed to use the "best estimate" valuation basis, rather than the "central" basis, if a guarantee were provided, thus reducing company contributions from £289m pa to £200m pa. Mr Lamb also noted that "if an agreement along these lines can be reached the Trustee will be better placed to try to accommodate Company wishes on investment policy and structure", the reference to structure being, in context, a reference to the proposed changes to the contribution rates.

In that letter, Mr Lamb explained that he wanted to:

It is clear from those passages that it had not occurred to Mr Lamb that the benefit structure of the Plans mightchange, and so the explicit discussion concerned the funding of the Plans and did not expressly address ongoing benefits. But a discussion of funding only makes sense if those involved understand precisely what it is the funding relates to. In that respect, there can be no doubt that one element of the funding discussion related to the deficit in respect of past service. Past service benefits were a given: at least in relation to Project Ocean there was no suggestion that the final salary link might be broken.

But the position in relation to the future is less clear. It is apparent, in my view, that Mr Lamb was making theassumption that benefits accrual would continue unchanged. It is hardly likely that Mr Lamb would have used the language he did if it had occurred to him that the Plans might be closed in the short term (whatever that may mean) particularly given the use of the words "ongoing contribution" and "commitment to ensuring that ongoing contributions will be met". He cannot, in my view, be read as speaking only of the contributions needed to fund the past service deficit or, in my view, of contributions to some hitherto unmentioned reduced accrual rate. I think Mr Hirst must have appreciated that that was the basis on which Mr Lamb was proceeding. Indeed, Holdings itself had no intention at that stage to reduce benefits, even if it had been discussed within CHQ as a possibility.

Moreover, a feature I will address later, the contribution figure actually agreed for the years 2005, 2006 and2007 included an amount in respect of benefit accrual during those years, a fact which the Trustee would have been well aware of. That is consistent with the second paragraph just quoted from Mr Lamb's letter. In that paragraph, his reference to "ongoing contributions" must be read as contributions to reflect precisely the same obligation as the £289 million was required to meet: that is to say a contribution which included an amount in respect of the relevant year's future benefit accrual.

But none of this means that Mr Lamb thought that his proposal would give rise to a guarantee that benefitaccrual under the Plans would continue unchanged indefinitely or even for a fixed period come what may. To be fair to him and to the RBs, that is not their case. But it is an aspect of this litigation to which a considerable amount of time and energy has been devoted (in relation to Soto as well as Ocean I must add). I need to deal with it in some length because the results of that time and energy inform the debate about what message Holdings and IBM Corporation/CHQ actually did give to members.

The context of the letters passing between Mr Lamb and Mr Hirst was a funding problem for IBM UK. Thefunding rate which the Scheme Actuary would be recommending was more than CHQ wanted Holdings to pay, or perhaps which it could afford, to pay. Had Holdings been willing and able to pay, however, there would have been no need for a funding agreement with the Trustee. Nor would there have been any question of a binding commitment that future accrual would remain unchanged (although, of course, any changes to the Plans in the future would have to be effected within the constraints of the Trust Deed and Rules). Whatever may have been Mr Lamb's assumption, I do not consider that it was implicit in Mr Lamb's proposal that IBM would guarantee not seek to change the accrual rate in the future either at all or for some defined period. The proposal, insofar as it can be ascertained from the letter and the attachments (which I come to in a moment), was to replace one funding proposal (namely the recommendation of the Scheme Actuary) in relation to the deficit and an assumed future accrual rate with another funding proposal (namely that proposed by Mr Lamb as a satisfactory middle ground between the extremes of the Scheme Actuary's recommendation and CHQ's wishes). Each proposal was based on the same (assumed) future accrual rate. There is nothing in Mr Lamb's proposal to give rise to a commitment by IBM that the sort of changes for the future which could have been made had the Scheme Actuary's recommendations been adopted would not be made.

In any case, Mr Lamb's letter was only the start of a process. The guarantee which was eventually agreed was,presumably, seen by everyone concerned, including Mr Lamb, as the commitment referred to in his letter. It is really by reference to that guarantee and the various statements made about it that the RBs must make their case. Mr Lamb's letter is part of the background, but was in reality overtaken by events.

Mr Lamb's letter attached two letters to him from the Scheme Actuary which "set out two options for such anIBM Corporation Commitment" (that is to say the commitment "to ensuring that ongoing contributions will be met"). The first letter was informed by the discussion which Mr Lamb had previously held with Mr Hirst. The second included an option which had not been discussed but which Mr Lamb thought might be attractive.

The first letter was described by the Scheme Actuary as an outline of what the Trustees should seek if it was toaccept a lower level of contribution than those set out under the "Central Basis" valuation results. The Trustee was to be guaranteed that the agreed funding and statutory payment obligations would be met by IBM

It is apparent from that letter that the Scheme Actuary's concern about adopting the "best estimate" basis wasessentially that Holdings would not be able to meet its obligations. With a guarantee from IBM Corporation, that risk was covered to his satisfaction. Consistently with that, the guarantee would come to an end if either (i) there was no funding shortfall and the solvency position on winding up were satisfactory or (ii) Holdings' own financial position improved to the extent that it could clearly meet any funding or solvency shortfall. I do not consider it is possible to spell out of that letter a requirement, or even understanding, on the part of the Scheme

Mr Lamb himself was well aware of the sensitive position he and the Trustee were in. His view at the time, ashe put it in his second witness statement, was that there was no obvious benefit in insisting on a valuation taking place on such a prudent basis that it would produce a deficit which Holdings might have difficulties financing. And so he considered the best solution was to put in place a guarantee if one could be obtained which would then stand behind the liabilities of the Plans. The balance he perceived was that, although it was necessary to achieve higher company contributions, the Trustee:

I accept this evidence about his and the Trustee's views. Mr Lamb wanted to produce a solution acceptable toIBM which did not provoke IBM Corporation (or indeed Holdings itself) into closing the Plans or scaling back benefits although I shall note that the fear of closure did not derive from any threat made by IBM. Indeed IBM was not considering closure at that time. But that is very different from seeking an assurance (albeit one not contractually binding or one giving rise to an estoppel or actionable misrepresentation) about the future, whatever Mr Lamb's and Holdings' assumptions and intentions might have been.

The effect of these proposals was to reduce the amount which needed to be contributed from that originallyrecommended by the Scheme Actuary. CHQ had previously been contemplating the making of a significant cash injection but the amount could now be reduced.

IBM's choices are nicely reflected in an email to Mr MacDonald from Mr Greene dated 9 October 2004 whereMr Greene noted that any additional cash could be used either to make further contributions and reduce the investment risk in the plans, or to boost EPS. This email shows, I think, that Mr Greene was certainly not contemplating closure of the Plans. He wrote:

IBM's full package of proposals, including both the guarantee and the changes in respect of future accrual, wasset out in an attachment to a letter from Mr Hirst to Mr Lamb dated 15 October 2004, replying to the latter's letter of 4 October 2004. It is important to record what Mr Hirst did and did not say in that letter. First, he referred to the 4 October letter referring to Mr Lamb's indication that the Scheme Actuary would agree to recommend a contribution of £200m pa for the next three years if IBM Corporation was able to provide an acceptable commitment to support the plans. Mr Hirst then stated that there had been extensive consultation within IBM and was able to confirm that "we are willing to agree in principle to such an arrangement in a form which I believe should be acceptable to the Trustee and the Scheme Actuary". He attached a summary of the principles which would be passed to the lawyers to produce an agreed document for signature. What he did not do was go any further than that and in particular he did not say anything about the ongoing accrual rate in the Plans. It seems to me, therefore, that the position at that stage was that Mr Lamb and Mr Hirst had discussed certain proposals; that those proposals had been considered within IBM; and that the proposal on the table for acceptance by the Trustee ("….should be acceptable to the Trustee and the Scheme Actuary") was that contained in the attachment. Mr Lamb could breathe a sigh of relief but he would not, in my view, have been entitled to perceive the arrangements as amounting to any sort of assurance about future benefit accrual other, perhaps, than an implicit assurance that IBM had no current plans to reduce benefits. Nonetheless, he continued to believe – and nothing was said by Mr Hirst or anyone else to disabuse him of the belief – that the basis of their earlier communications and the messages to be taken from them had not in any way altered.

The proposals were then presented to the Trustee by Mr Heath on behalf of Holdings at the TMM held on 21 October 2004. On the guarantee, the Scheme Actuary (who was present) explained its function and that it would allow him to remove the margin of prudence in the valuation assumptions. I interpose here to note that nothing can be inferred from the existence of the margin of prudence about future accrual. That margin was included at the rate actually adopted to reflect the assumed accrual rate. But if the accrual rate reduced or if the Plans were closed, that margin would be reduced (as an absolute number but remaining at the same percentage perhaps) or, on closure, cease to be payable. Accordingly, what the Scheme Actuary said about the margin is not relevant, in my view, to the issue of commitment to the Plans.

According to the minutes, Mr Lamb explained that there was "an implied linkage" between the guarantee and theinvestment strategy "as the Trustee might be better disposed towards certain investment strategies with the Guarantee in place than it would without one." Mr Greene added that "from the Corporation's point of view there was a desire to see improved investment returns and it was hoped that the Guarantee would enable the Trustee to achieve this through its bond strategy and long term approach to its equity strategy". The Trustee resolved that Mr Lamb would work with Nabarro to negotiate with IBM WTC on the final wording of the Guarantee.

On the increase in contribution rates, the TMM was provided with a proposal paper which set out IBM's statedreasons for making the proposals, namely a significant increase in the cost of future accrual as the result of improving life expectancy, lower inflation and lower real rates of investment return. The stated aim, as recorded in the paper, was to strike a "fair balance between the rights and expectations of pension plan members and affordability on the part of the company." Mr Heath emphasised to the Trustee that employees were only "being asked to take a share of additional costs for future service and were not being asked to contribute towards either the deficit or their accrued rights." Mr Lamb explained "the increased contributions levels would have the effect of making it a safer environment for all members."

The proposals to change the contribution rates were approved by the Trustee with the support of 5 boardmembers, with 1 member against, 2 abstentions, and 4 members unable to vote because they were conflicted.

The Trustee resolved to agree to the proposals in principle, but this was subject to employee consultation andexpressly on "the assumption that there are no plans for other major changes to the trustee deed & rules or changes to pension practice". Mr Greene and Mr Heath were present at this meeting. They clearly must have understood the assumption that was being made. Holdings and CHQ must have known the basis of the

I do not think that is disputed by Mr Simmonds. But he submits that at that time, there were no such plans. Iaccept that submission to the extent that there were no positive plans under consideration at the time; and, as I

I note, in this context, that when granting approval on 8 October 2004 to Mr Heath to proceed with the ProjectOcean proposals, Mr MacDonald explained to Mr Heath that:

Mr Tennet placed reliance on what is recorded at page 7 of the minutes of this meeting. He suggested that itdemonstrates that there was already confusion amongst the members of the Trustee board as to what the guarantee meant in practice. The relevant passage is as follows:

Mr Simmonds regards this as a forensic point of no relevance to the issue. I agree, for the reasons which hegives as follows:

The proposals were announced by a combination of (i) a letter from Mr Lamb on behalf of the Trustee sent on 9 November 2004 and (ii) a webcast presented by Mr Heath to all IBM UK employees on the same day ("the Webcast").

The Webcast is an important communication. The RBs rely on it strongly as a representation about IBM's futureconduct in relation to the Plans. IBM contends that there is nothing in it which is inconsistent with what was said and done in its subsequent conduct in formulating and implementing Project Waltz. IBM contends that none of the statements made in the Webcast or in any of the subsequent roadshows carried out by Mr Heath and his colleague, Mr Waller, at various IBM sites across the UK when properly understood in context amounted to a guarantee or commitment that DB accrual would remain open, or that other changes to the Plans' benefits structure would not be made, for any particular period.

It is appropriate for me again to note that, as was made clear in the written opening submissions on behalf of theRBs, it is not, and never has been, their case that IBM guaranteed to members so as to give rise to a legally binding obligation that the Plans would never be closed in the future. The RBs says that that is not the issue. The issue, they say, goes to the reasonable expectations which IBM engendered about whether it would be necessary to revisit the issue of closure or further changes over a long-term time horizon. While not giving an absolute guarantee, it is said that IBM certainly engendered a reasonable expectation that it would not be making further changes to, or closing, the UK DB Plans for the long term.

I set out for convenience the whole of the text of the Webcast in Annex D to this judgment. I have addedparagraph numbers thus: [x] for ease of reference. I refer to the numbered paragraphs of that text as "Webcast [x]". On the RBs' own evidence, the member witnesses' subjective understandings (given such relevance as is due to them in accordance with the discussion I have already conducted) stemmed, in many cases, from the Webcast; and Holdings' conduct in 2006 was relied on as confirming the expectations which had already been engendered in 2004 in relation to Ocean by the Webcast in particular. Accordingly, Mr Simmonds submits that, if the RBs are unable to establish Reasonable Expectations based on the Webcast, then "their whole good-faith case is holed below the waterline".

Although the Webcast is an important document, it does not stand alone even in relation to Project Ocean. Itfollowed on from the agreement of the Trustee to the proposals, as a communication exercise in order to obtain the assent (rather than formal consent which was not required) of the members. In addition, there was the important communication from Mr Lamb to the members already addressed; he wrote as he did knowing that the Trustee had agreed to the proposals only on the assumption that there were no further major changes planned to the rules of the DB plans or changes to pension practice.

As a preliminary point, it is to be noted that the Webcast refers throughout to IBM. It is clear that this is areference to Holdings (although it perhaps includes IBM UKL but nothing turns on that) and not to IBM Corporation or CHQ. Where IBM Corporation or some other corporate entity (eg IBM WTC) is referred to, it is expressly identified: see Webcast [41] and [42].

Mr Heath told the Court that the Webcast was the "primary method of communication" to members, alongsidethe Project Ocean Roadshows (materials from which Mr Tennet points out have never been disclosed to the RBs). He also points out that neither the email announcing the Webcast nor the Webcast itself made any mention of any other materials on the changes (eg Q&A documents). But I attach little significance to that. There is no reason to think that people who knew about the Webcast did not also know about the existence of the Q&As and, in any event, members were given the opportunity to attend roadshows which were referred to in Webcast [50].

I accept, as Mr Tennet submits, that the Webcast was a communication to be listened to and not, as its primarymethod of assimilation, read. Mr Heath's announcement email did not include a transcript but instead just a link to the audio webcast. He could not remember if a transcript was ever made available. Whether it was or not, I do not consider that the hypothetical reasonable member can be expected to have obtained and read a transcript. But I also agree with Mr Simmonds when he says that the reasonable member would have been aware of the following:

It was not possible to listen to the Webcast in the way which it was possible to read a document. With a paperdocument at least the reader can flick backwards and forwards. But he/she could not do that with the Webcast: the most he/she could do was listen to it more than once. It has not been possible for me to listen to the Webcast. It has not proved possible to locate a copy. I must therefore do the best I can to take from the transcript the flavour of the Webcast itself.

Mr Tennet is critical of some parts of Mr Simmonds' cross-examination of the member witnesses in relation tothe Webcast. He says that aspect of that exercise was artificial, with the witness being asked to accept the meaning of one part of the Webcast by reference to another part when there was not in practice the opportunity when simply listening to the Webcast to make the sort of detailed comparison that the cross-examination entailed. Quite reasonably, members instead focused on the overall message, and did not pick through the language as if they were lawyers construing a commercial contract, looking for nuance and inconsistencies. Those are, in principle, perfectly fair submissions. I take account of the point in assessing what the reasonable member would take from the Webcast. I have read and re-read the transcript of the Webcast (it might be said that the re-reading is unfortunate for, in doing so, I have put myself in a very different position from the reasonable member) and been addressed about it. It might have been better if I had been asked to read and record my impressions before being treated to the detailed analysis. I have done my best to imagine what the hearer of the Webcast would have taken away from it and to take account of the fact that the impact of the written text, read and re-read and subjected to analysis, may be different. It is nonetheless inevitable that I must carry out just the sort of close textual analysis that Mr Tennet says is not the correct approach: I need to do so to deal with Mr Simmonds' submissions as well as to see whether Mr Tennet's case might succeed even on this close analysis.

Mr Simmonds put it to the member witnesses (or at least to some of them) that the objective of the proposalswas to secure the sustainability of the Plans. That was the aim of the package taken as a whole. I think that all of the witnesses agreed with that, although they had an idea of what "sustainability" entailed somewhat different from Mr Simmonds' idea. It was, however, common ground that some sort of bargain was struck between Holdings and the members, using the word bargain in a very loose sense. Thus Holdings would take certain actions (increase its contributions, take responsibility for the past service deficit, provide a guarantee (via IBM WTC) under which it would take responsibility for contributions including deficit contributions until 2014); members would pay increased contributions (or, in the case of the N Plan, members could instead suffer a reduced accrual rate) and any change in the cost of future benefits would be shared equally between Holdings and the members (an aspect which became known as "the Evergreen Agreement").

Holdings' case is that since, as is common ground, there is nothing express in the Webcast about future accrual(and certainly no unequivocal express statement or representation that the Plans would remain open to future DB accrual) and since no express commitment is to be found anywhere else (whether in documents or statements on behalf of Holdings), the most that can be derived from the Webcast (and indeed from any other communications) is that Holdings had no present intention to change future benefit accrual.

As to that, I accept that Mr Heath and Mr Hirst honestly believed at the time that there were no further plans tocease DB accrual or to make the other changes which were later proposed as part of Project Waltz, as to which the RBs do not contend otherwise. Mr Simmonds recognises that Mr Heath accepted that he approached the communications process with a view to 'selling' the package of measures to employees; this does not, he says, detract from the facts that (1) he made no representation as to future accrual and (2) he believed that the package could be sold both positively and honestly, and that the package had solved the pension 'problem' which so far as UK management was concerned was a funding problem for the past and a cost (and to that extent funding) problem for the future.

That, on Holdings's case, is enough to dispose of the claim for breach of the Imperial duty insofar as reliance is placed on the Webcast: the Court need not engage at all in any further analysis of the Webcast itself.

Mr Simmonds accepts that it is implicit in the Webcast that Holdings did not have a present intention, which ithad not revealed, to effect further adverse changes to benefits or member contributions in the near future. But he says that, as a matter of fact, Holdings did not have such an intention so that I do not need to be concerned with the consequences of that implicit feature being contrary to Holdings' actual intention. I will need to address in due course what Holdings knew and intended in contrast with what any individuals knew.

Mr Tennet accepts that the Funding Agreement did not, by itself, constitute a promise that DB accrual wouldcontinue (either at all or at any particular rate) in the future. He accepts that Holdings is not precluded from taking steps to close the fund to further accrual as the result of an estoppel or actionable representation. And he accepts, as he must, that Mr Heath did not state expressly that DB accrual would continue in the future or that there would be no further changes to the Plans.

Mr Tennet contends, however, that the Webcast gave a number of assurances as to what the bargain which Ihave just described would achieve; and that these assurances and commitments are enough to give rise to Reasonable Expectations, the disappointment of which were capable of giving rise to a breach of the Imperial duty (and which in fact did give rise to such a breach). In that context, he submits that the Webcast "repeatedly states Holdings' "commitment" to the UK DB Plans including their "sustainability" for ongoing accrual".

As to that last point, the Webcast in fact uses the word "commitment" twice. Although Mr Tennet's statement isliterally correct, the expectation (legitimate perhaps) of the reasonable reader and hearer (me) of his submissions would be to expect to find many more references. Be that as it may, the two references are found in Webcast [43] and [49]. The first refers to Holdings' "continuing commitment to the Plans and their members" being demonstrated by the guarantee. The second refers to Holdings' "commitment to underpin the sustainability of the defined benefit plans in the UK" being demonstrated by "this series of initiatives", that is to say the change to employee contributions, the significant cash contributions Holdings was to make (largely directed at making good the deficit) and the guarantee.

As to "sustainability", this word is used three times (in Webcast [10], [14] and [49]).

It is worth noting Mr Heath's description at Webcast [9] of Holdings as a "strong and responsible company" andhis belief that the proposals "offer a positive, measured response to a number of fundamental changes which are impacting employers across the UK".

At Webcast [41] Mr Heath starts his explanation of how it was proposed "to further enhance the security of theplans", a reference back, no doubt, to the "further measures to underpin the security of the plans" referred to in Webcast [2]. The explanation relates to the guarantee and nothing else. The detail of the guarantee was not given, nor were its terms (let alone a draft) available. All the members were told was that Holdings had agreed in principle to guarantee company contributions to the fund. I see no reason why anyone would have taken this to be a guarantee only of the deficit contributions; and the eventual guarantee in the Funding Agreement was not in fact limited in that way.

Mr Simmonds relies on an answer given by Mr Heath in answer to a question from me whether or not heexpected that Holdings might not later turn around and seek to reduce the benefits (thereby reducing the size of the deficit) the response was:

And so it is said that Mr Heath's purpose was to reassure people that Holdings was recognising its

For my part, I do not consider that the references to the guarantee indicate that it is to apply only in respect ofthe past service deficit and the contributions referable to that deficit. It is true that it follows on immediately from the section of the Webcast dealing with past service, but earlier in the Webcast (see Webcast [17] and [18]) reference is made to a sharing of the increased cost (so that Holdings will bear a share) and to the fact that Holdings will continue to bear the lion's share of the contributions. That is the conclusion which I reached after having taken the opportunity to read, several times, the hard copy of the Webcast. It is also the impression which I think I would have taken from listening to the Webcast on a single or number of occasions (had it been possible to do so) without the benefit of a hard copy. My assessment is that the overwhelming overall impression is that the reasonable listener to the Webcast would understand that the guarantee would relate to all contributions payable by Holdings to the Plans. I reject Mr Simmonds' submission that the guarantee as described in the Webcast (in contrast with how it actually appeared in the Funding Agreement) was restricted to the same territory as the increased employer contributions referred to in Webcast [39] which is what I understand him to submit.

Mr Simmonds suggests that it is not clear how any member would have gleaned from the Webcast anunderstanding that the guarantee covered future service. To my mind, it is the natural reading, at least of the transcript of the Webcast, that the guarantee would relate to all contributions falling due during the course of the guarantee (until 2014 by which time Holdings hoped the deficit would have been eliminated). But it does not necessarily follow that any particular level of further benefit accrual (or any accrual at all) is guaranteed: all that is guaranteed is the payment of such contributions as actually fall due in accordance with whatever accrual rate is in force from time to time. I agree with Mr Simmonds that there was no guarantee of any particular rate of future accrual. However, the RBs do not assert, as I have said, that the Webcast amounted to a promise or legally binding guarantee. Their case, rather, is that a departure from the "commitment" displayed by the Webcast resulted, in all the circumstances, in a breach of the Imperial duty which is something very different.

Past service benefits are mentioned in Webcast [12] and addressed more fully separately in Webcast [34]ff. As to those benefits, the clear message is that Holdings is taking responsibility for the deficit: see Webcast [12] and [34] in particular. In saying what he did in those paragraphs, there is no reason to think that Mr Heath (or Holdings) had failed to recognise that the actual benefits in respect of past service included the final salary link and nothing to suggest that he considered that existing pension practice (eg early retirement policy) was to be left out of account when ascertaining the past service liabilities.

At Webcast [47], Mr Heath adds that other alternative proposals had been considered including changes toretirement age and more widespread changes in benefits design. He does not say, although it is the case, that closure of the Plans to future accrual had also been considered. These alternatives were not adopted because, at least in Mr Heath's view, the proposals actually put forward represented the right initiatives which "balance the need to ask you to take a share of increased future costs with long term security". The balance he is identifying in that statement is, it seems to me, the balance in what it is that the members and Holdings each bring to the overall package: members pay increased contributions (or in the N Plan receive less benefits) whilst Holdings makes its cash injection and procures the guarantee (thus increasing the security of benefit).

As to the references to putting the Plans "on a firm footing for the future" and to the "commitment" in Webcast[48] and [49], Mr Simmonds suggests that they are a perfectly accurate description of what Holdings was proposing to do and the objective, which he has identified, of the package, especially given that Mr Heath had identified the large deficit as the principal threat. But that, in my view, is to diminish inappropriately the message which was also being conveyed in relation to future service by those same paragraphs. Thus Webcast [48] refers to "the changes we are proposing to make" which can only be a reference to increased contributions in the C and I Plans and the reduced accrual rate in the N Plan. And the "commitment" in Webcast [49] is a commitment to sustainability which must include looking at the Plans as ongoing Plans. The natural assumption would be that the Plans would continue to provide further benefits: Holdings could not have intended, and a reasonable member would certainly not have understood, the Webcast to be contemplating sustainability as drawing a distinction simply between an unsustainable scheme (which would have to be put into winding-up) and a sustainable scheme (which could continue as a closed scheme with no further accrual). Such a sophisticated distinction is not one which I would expect the reasonable member – even as rational and intelligent as the Holdings workforce – not skilled in arcane matters of pensions law and practice to understand from listening to the Webcast.

Part of the message in Webcast [47] to [49] may, in reality, have been the most important part of the Webcast:

This approach is not inconsistent with Mr Heath's own evidence in cross-examination where he said this:

Mr Simmonds submits that Mr Heath's answer accurately summarises not only the intention of Holdings at thetime but also what a reasonable employee would have taken from the Webcast. I do not disagree to this extent: the Webcast was not giving a guarantee about future accrual either indefinitely or for any particular period and a reasonable member would not have understood that it was doing so. But, as I have explained, that is not the RBs' case.

Mr Simmonds goes as far as to suggest that a reasonable member might not even have drawn the inference fromthe Webcast that Holdings hoped to keep DB accrual going given these factors: (i) that there is no clear reference to keeping open DB accrual for the future (ii) that the guarantee itself is described by Mr Heath solely in terms of funding and (iii) that benefits under both the Main Plan and the I Plan were variable for future service. So, he says, a commitment to fund did not necessarily presuppose any particular level of benefits for future service. It is, however, clear that the proposals were proceeding on the basis that the Plans would continue unaltered: or to put it simply another way (as is Holdings' own case) no further changes to the Plans were being considered by Holdings. This is the only conclusion consistent with the basis on which the Trustee agreed to the proposals and with Mr Heath's own observation that he was not expecting "a further change 2 years down the line". I regard that (inevitable) conclusion as consistent only with the further conclusion that Holdings hoped that there would be no further changes in the foreseeable future. I accordingly reject the suggestion identified at the beginning of this paragraph.

There is more substance in Mr Simmonds' argument that a guarantee of future funding does not imply aguarantee of future accrual. Rather, there is simply a guarantee of contributions in relation to whatever benefits actually accrue. On this approach, the existence of the guarantee from IBM WTC does not constrain Holdings in making changes to the Plans: those changes stand or fall on their own merits.

I agree that a close analysis of the transcript of the Webcast shows that the guarantee is concerned only to securethe payment by Holdings of its contributions under the Trust Deed and Rules and the cash contribution. I do not consider that the terms of the guarantee, as explained in the Webcast, have anything to say about a commitment to keep the DB plans open to further accrual or to preserve the pension practices then current. But that is not to say that references to the guarantee are irrelevant to further accrual and pensions practice. The presence of the guarantee, in contrast with its terms, is a demonstration of Holdings' commitment to the Plans (I actually wonder whether Mr Heath intended to refer to IBM Corporation's commitment when he referred to IBM's commitment). To the extent that the references to commitment and sustainability give rise to Reasonable Expectations, those

Mr Tennet suggests that the reasonable member would not appreciate that the guarantee of "companycontributions to the fund" (see Webcast [41] and [42]) up to 2014, did not necessarily mean that accrual would actually continue. The understandable expectation, he says, was that if contributions in respect of future accrual to 2014 were guaranteed, there would be no reason to stop accrual. He suggests (although I do not know the evidential basis for this, but shall assume for the moment that it is correct) that many (although not all) members did not differentiate between different IBM entities and assumed that IBM as a global business would stand and fall together. Many never expected that Holdings would ever be permitted by CHQ to become insolvent. The value of a funding guarantee would not necessarily have been apparent to them. It was therefore perfectly reasonable for members to think that the guarantee did something beyond merely providing security for Holdings existing funding obligations in the event that it was unable to pay.

I do not agree with that line of argument. First, I do not agree that the reasonable member would have consideredthat there would be no reason to stop accrual if the guarantee was in place. The reasonable member may not be an expert in the law of guarantees, but he might well have known – I have not been addressed with argument either way on this point – that a principal debtor will usually remain responsible for indemnifying his guarantor; thus Holdings would end up with an inter-company obligation if IBM WTC were called on under its guarantee. Even Mr Tennet does not appear to suggest that the reasonable member (in contrast with many although not all members) saw IBM globally as a business which would stand or fall together. It is hard to think that that was the view of UK management who were acutely aware of the internal competition for investment from Armonk. And even if it was reasonable for some members to think that the guarantee did more than simply guarantee Holdings' contribution obligation, Mr Tennet's argument does not establish that the reasonable member would have thought this way. Nor does the argument lead to the conclusion that the "something beyond" went all the way to an expectation that the Plans would remain open until 2014.

I conclude therefore that however one approaches the Webcast, the references to the guarantee, standing alone,do not justify the Reasonable Expectations on which the members rely in relation to future service accrual. I should add this: It does not follow from this conclusion that, when one comes to consider whether Project Waltz gave rise to a breach of the Imperial duty, responsibility for the deficit is irrelevant: any business justification for the changes will have to explain why the particular changes were ones which were appropriate. The statements concerning responsibility for the deficit in the Webcast and elsewhere are factors then to be taken into account in answering the question whether any reasonable employer would act in the way that Holdings acted.

My conclusion is not affected by Mr Heath's statement in Webcast [42] and [43] that the guarantee is "excellentnews" for members. I do not consider that the description of the guarantee as excellent news gives any support to the establishment the alleged Reasonable Expectations. The guarantee clearly was excellent news for anyone who was concerned about Holdings ability to meet its funding commitment. As to Mr Tennet's argument about that being a fanciful concern, see paragraph 705 below. Mr Heath was also entitled to see it as excellent news (although he did not articulate this reason) because without it the Trustee might well have proceeded with its switch to bonds which might (Mr Lamb feared as much) provoke IBM into closing the Plans or even winding them up.

As to "sustainability", Mr Tennet must be right to say that this means that the DB Plans would be sustainable intheir current form. I consider that to be the only sensible interpretation of the word "sustainable" in context. Mr Heath agreed that he intended to demonstrate the company's intention "to continue to support the plan in its current form", saying that "There was no doubt that at that time we believed that…what we were doing was securing the sustainability of the plan for the longer-term". Indeed, it was precisely to ensure the sustainability of that which existed already that the proposals were put forward.

Mr Tennet identifies two implications to be taken from the Webcast: the first is that, following adoption of theproposals, Holdings saw the DB Plans as sustainable for the foreseeable future. The consequence of that would then be that the members had a Reasonable Expectation that benefit accrual would continue for the foreseeable future. The second implication is that, in making these assurances as to sustainability and statements of commitment, Holdings implied that it had reasonable grounds to believe that the assurances were true.

Before addressing those two implications, I make this observation. If it were established that either Holdings didnot believe, or is to be treated as not believing that the DB Plans were sustainable (eg it was reckless as to whether that was true or not) that cannot have any impact in relation to the Reasonable Expectations engendered by the Webcast. A Reasonable Expectation, as I have coined that expression for the purposes of this judgment, is an expectation as to future conduct. An expectation that someone is telling the truth is of a different type. An incorrect statement may give rise to a remedy, for instance based on misrepresentation or estoppel, but that is because it is incorrect and was acted upon, not because the recipient believed it was true. An incorrect statement, for instance one made dishonestly or recklessly, may give rise to a breach of the Imperial duty but that is not because it gives rise to a Reasonable Expectation.

In the present case, it is not alleged that the statement as to sustainability was itself a breach of Holdings'Imperial duty. If it had been, and if a breach were established, the consequences of such a breach in terms of any remedy to be granted today are terrible to contemplate. For the reasons just given, those statements are not, in my view, relevant to Reasonable Expectations. The accuracy of the statements about sustainability and putting the Plans on a firm footing do not, therefore, have the importance which Mr Tennet attaches to them.

I nonetheless need to say more about Mr Tennet's two implications: in relation to the first implication, I have,unfortunately, no clear idea what Mr Tennet means by the words "foreseeable future". If, as I consider it does, sustainability refers to sustainability of the Plans in their current form, the Webcast cannot reasonably be read as looking further ahead than, at longest, the period to 2014. There is some merit in taking that period because, according to the Webcast, it was anticipated that the deficit would, at the contribution rates then proposed, be eliminated by 2014 which would suggest that Holdings at least had some material to suggest that the Plans would remain viable with their current structure until then.

But that is to read a lot into both those words ("sustainability" and "commitment") and into the whole phrases("on a firm footing for the future" and "commitment to underpin the sustainability") as they would be read by the reasonable member rather than by advocates and Judge in the courtroom. At most, in my view, a member would be able to take away from the Webcast the message that Holdings considered that continued accrual at the current rate would be affordable by Holdings in accordance with current cost projections and in the light of its own current financial circumstances. I use the word affordable rather than sustainable deliberately. From Holdings' perspective, once the members' contributions were increased, there would be no difference between sustainability and affordability by Holdings of its obligations.

If I am correct in my conclusion that the duration of the guarantee does not provide a useful indicator of theperiod of the foreseeable future, the Webcast contains, expressly, nothing to tell us how long that period lasts.

Some help can be found in Mr Tennet's own submissions that it is to be implied that Holdings had reasonablegrounds to believe that the assurances as to sustainability which it gave were true. One of his submissions is that the members were entitled to think that IBM had taken reasonable steps to understand the present circumstances of the UK DB Plans and what might happen in the future and had reasonably assessed the possibility, and possible consequences, of unfavourable market performance in the future: see paragraphs 661ff below. The reasonable steps, assuming the submission to be correct, are essentially obtaining financial information and advice and turning those into projections. There is a limit about the distance into the future for which it is reasonable to expect projections to be made. Even if Holdings is to be taken as making an implied commitment into the future, I do not think that the implication can reasonably be expected to last longer than the period during which reasonable projections can be made.

As to that, I consider a much shorter horizon than 2014 to be appropriate. To suggest that Holdings could in

As to the second implication which Mr Tennet seeks to draw, it is certainly the case that Mr Heath must be takenas believing the truth of what he said. What is more, Holdings itself cannot be heard to say – and does not say – that what Mr Heath said was unauthorised and not to be taken as a statement by Holdings. Further, although Mr Heath did not speak on behalf of CHQ, what he did say was known to and approved by CHQ. I make no judgment about what, as a matter of either law or morality, CHQ ought to have said if they knew that Mr Heath's remarks about sustainability and commitment were not thought by CHQ to be accurate. But what I do say is that, if CHQ did not think them to be accurate and remained silent, I do not think that Holdings can be in any better position insofar as breach of the Imperial duty is concerned than if Holdings itself had thought them to be inaccurate.

On the basis of that second implication, Mr Tennet submits that members could reasonably expect that:

Without such an implication, he suggests that the assurances and statements of commitment would be of littlevalue. Consistently with that submission, it is pointed out that Mr Heath confirmed that Holdings own intention was to explain that the UK DB Plans would be sustainable even if things changed. The RBs' written closing relies on this part of what he said, "we believed that what we were putting in place again was a proposal which would give greater sustainability, notwithstanding other things might happen". The RBs' written closing does not, however, include the next part of his answer: "In the circumstances at the time". That is an important part of the answer: it makes clear that Mr Heath saw the proposals as providing the necessary stability in the circumstances at the time. Significant changes in financial and economic conditions would not, therefore, preclude a subsequent change in the terms of the Plans. It is also important to note that the part of his answer which was quoted did not suggest that Holdings was making a statement about absolute sustainability. The aim, he said, was to give greater sustainability, that is to say as compared with the situation prior to the amendments.

If one asks whether the proposals produced greater sustainability making that comparison, it is clear to me thatthe answer is that they did so. So far as concerns past service benefits, the capital injection coupled with the IBM WTC guarantee was clearly an improvement (subject to the next paragraph of this judgment); so far as the future was concerned, the increased member contributions and the terms of the Evergreen Agreement (ie the agreement to share any increased cost between Holdings and members) self-evidently made future benefits more affordable from Holdings' perspective and, again, the guarantee gave an added element of comfort. In any case, Mr Heath was alive to the possibility of termination of further accrual at that time and that acceptance of the proposals would avoid the risk of immediate termination. I do not, therefore, consider that Mr Heath's answer assists the RBs' case and I do not agree with the submission that, without such implications, the assurances and statements of commitment would be of little value, although they would certainly, I accept, be of less value.

I do, however, accept that the reasonable member was entitled to think that Holdings had made an assessment ofthe current circumstances of the DB Plans. I consider also that the reasonable member was also entitled to think that the representations about sustainability and IBM's commitment would hold good in the light of (i) the investment policy to be adopted as Holdings understood it would be and (ii) the actuarial valuation assumptions to be adopted as Holdings understood them. Although the members did not know the details of either of (i) or (ii), Holdings certainly did and, indeed, the bargain with the Trustee was that the shift to bonds should be slowed down and that actuarial valuations should no longer be on the "central basis" but on the less conservative "best estimate basis".

I do not accept, assuming that I have understood it correctly, Mr Tennet's further submission, namely that "it isto be implied that Holdings was prepared to tolerate the consequences of such unfavourable market performance if it occurred and to stand by the DB Plans". I understand this to mean that even if markets moved adversely, the reasonable member could expect that benefits would remain unchanged so that pension practice would continue and the DB Plans would remain open to future accrual at the current rate.

The submission which I have recorded – it is how it is put in the RBs' written closing submissions – is anextreme one. In fact, even Mr Tennet accepts that a really serious, unexpected and unpredictable change in financial and economic circumstances which put the very existence of Holdings at risk might justify changes to the DB Plans even in the short term. But suppose that the Trustee had been allowed to continue with its planned shift to bonds. The funds would still have remained invested significantly in equities and other return-seeking assets. I do not consider that the reasonable member could understand the statements in the Webcast about sustainability and commitment as creating an expectation that, even if the value of the funds fell significantly, the DB Plans would continue unchanged whether in relation to future service benefits or in relation to pension practice.

I do not consider that conclusion should be any different even though part of the bargain between Holdings andthe Trustee resulted in a slowing down in the rate of shift into bonds and a change to the "best estimate basis". The resulting investment spread, although different from that which the Trustee would otherwise have adopted, was within the range of reasonable and responsible investments given the existence of the guarantee (and perhaps even without it). There is no justifiable criticism of the Trustee for having adopted the policy which it did given the existence of the guarantee.

At the other end of the spectrum, I do consider, for what it is worth, that in the short term the members couldreasonably expect that no further changes would be made in the absence of a significant change in financial and economic circumstances. For instance, the members would justifiably be shocked if, the day after the Ocean changes had been implemented, Holdings had announced that it was reviewing the level of future benefit accrual and its early retirement policy. They would be justified in feeling shocked because the Webcast certainly gives the impression that there is a problem to be addressed and that the initiatives which it explains provide at least some fix to that problem. If Holdings had embarked on a review of DB Pension provision very soon after the implementation of the Ocean changes, this would require an explanation, otherwise its good faith in presenting the changes may be brought into question. It is not as though the Webcast had said: here is a problem which requires immediate treatment; here is an immediate fix to that problem; but it cannot be seen as a final solution and we will be embarking on a review of the DB Plans in the near future. Quite the reverse: it stated that the DB Plans were sustainable and that Holdings was committed to underpinning their security.

There is another indicator of the duration of Holdings' commitment to keep the DB Plans open. The £200m pacontribution was assessed by Watson Wyatt as a single figure including a deficit contribution and a future service contribution. The latter element was assessed on the basis of the current benefit structure. In my judgement, the reasonable member would understand that this annual contribution would be made for 3 years (see Webcast [39]) and that the DB Plans would therefore remain open for at least that period.

But this is all hypothetical: relevant changes were not in fact made within the short term. [The Project Sotoproposals were made within the 3 year period but these are not challenged by the RBs]. The Project Waltz proposals did not start to be formulated until 2008. That, in my view, is far beyond the short-term horizon within which it is obvious that further changes would not, absent a significant change of circumstance, be made.

The next stage of the RBs' argument is this. It is said that, by these assurances and statements of commitment,Holdings implied that it would not change its mind and renege on its commitment in future without good reason. Mr Tennet relies on two factors which I have already mentioned:

As to the first of those factors, I do not dissent from the proposition that the impact, such as it is, of the Webcastis not restricted to that very moment, and that is reflected in what I have said in the preceding two paragraphs.

As to the second of those factors again I have accepted that, in the short-term, the members could reasonablyexpect that no further changes would be made in the absence of a significant change in financial and economic circumstances. However, I disagree with the submission insofar as it relates to a period beyond which reasonable projections could be made, for reasons already given. Further, the assurances and expressions of commitment are not worthless even if Holdings could, other than in the short term, change its mind and act contrary to them. As well as the factors identified at paragraph 671 above, there are these additional factors. First, they gave the members comfort that the DB Plans were not about to be wound up or closed to further accrual. Secondly, they could be taken as an indication that there were no current plans (or even discussions on foot) to take such action. That would be so, it seems to me, even if Mr Heath had added, at the end of the Webcast, that although Holdings had no current intention to close the DB Plans or make any changes to accrual or pension practice, it reserved its right to do so: the addition of those words would not bring about an internal inconsistency.

Next it is said that the Webcast made clear that Holdings was taking sole responsibility for any past servicedeficit and explained the steps which would be taken to deal with it, over the period of the next three actuarial valuations. In contrast, there was to be an immediate increase in contributions to reflect increased costs due to factors such as increased longevity and any further increase in the costs of future service would be shared. It is not possible, subject to one qualification, to quarrel with either of those propositions. The acceptance of responsibility for past service benefits can be seen in Webcast [12], [18], [34]; and this is emphasised by the action Holdings was taking to deal with the deficit as set out in Webcast [39] (the £200m pa cash injection), [40] (later cash contributions to be based on subsequent valuations), and [42] (the guarantee until 2014).

The qualification is this. Mr Heath referred to Holdings' "responsibilities"; he was also referring to the deficit asdisclosed by actuarial valuations. He said nothing about the basis on which past service liabilities were to be ascertained. There are two related points to make here:

It might therefore be said that the "responsibility" to which Mr Heath was referring was for the deficit in respectof past service which would ultimately depend on the benefits which had to be provided in accordance with the provisions of the DB Plans and pension practice from time to time. On that approach, there is nothing in the argument that responsibility for the past included a commitment to future accrual on the basis that that is how the final salary link would be preserved. And nothing in the argument, either, that responsibility for the deficit in relation to past service benefit included a commitment not to change pension practice.

In that last context, I draw attention to what I have already said in paragraph 638 above. What I said there wasthere was no reason to think that Mr Heath was not aware of the final salary link or pension practice when it came to calculating the deficit. It does not follow from that that Mr Heath considered that pension practice could not be changed consistently with what he said in the Webcast.

Mr Tennet nonetheless submits that the statements about past service benefits have two consequences. Iparaphrase his two submissions to express them in the language of the expectations which the statements would engender in the reasonable employee.

The first submission is that the reasonable member would have the expectation, arising out of Holdings takingresponsibility for any deficit, that it would not change benefits applicable to past service in order to reduce the UK DB Plans' liabilities and thereby the very deficit for which it was taking responsibility. Cutting benefits to reduce any deficit would hardly be taking responsibility for it. Included in the benefits which are thus protected would be what Mr Tennet refers to as the "automatic entitlement" to early retirement under Holdings' longstanding practices or the linkage of emerging pension to final salary. Automatic entitlement to an early retirement pension

I need to disgress for a moment to consider this "automatic entitlement". There can be no doubt, in my view,that there was a widely held view that there was a right to early retirement. The RBs' opening submissions presented a number of factors leading to the conclusion that by 2009, at the time of Project Waltz, there was a cultural expectation of such a right resulting from a practice established over many, many years. The material does indeed show, I consider, that that expectation existed well before that and certainly before the time of the Ocean changes.

The first thing to note is that the evidence shows that, in practice, early retirement was commonplace. Theactuarial valuation reports up to 1997 show the average retirement age within the Main Plan during the 1980s and 1990s. It was never above 60.42, was once as low as 54.47 and was more often than not in the range 55 to 59.

The cultural expectation of early retirement was reinforced by the separation packages and enhanced earlyretirement terms under which members routinely left IBM UK, permitting early retirement from age 50 (even from deferment) without any requirement for consent or, from age 53, known as the Career Transition Programme, the Skills Rebalancing Offers and the Individual Separation Plan programmes. Holdings itself recognised, in its answers to questions presented at the Pensions Consultation Committee ("PCC") meeting on 12 August 2009, that historically most people retired in their late 50s and that retirement at NRA (63) was under 10%. There is no reason to think that that was not a long-standing situation.

The DB members' benefit statements reflected this culture. The statements contained tables showing the pensionavailable at each year from age 50 (later, 55) onwards, on the basis of the ERDF prescribed by the applicable plan rules as applicable on retirement from active service. Before 1997, the statements did not contain anything to suggest that such early retirement was at Holdings' discretion. After 1997, the statement continued to set out the amounts available at each age from 50 (or in due course 55).

On 13 July 2009, IBM presented some charts as part of the UKI Transformation summarising Holdings' earlyretirement practice and procedures in operation prior to the 2009 Project Waltz proposals. One slide described the essentially administrative process for obtaining approval from relevant line managers. It was noted (in bold typeface):

Mr Riley stated in his email dated 31 July 2010 to Mr Murphy that retirement at 57 was "obviously ingrained inthe mindset of our people". And he had stated at the final PCC meeting on 8 September 2009 that early retirement had become an automatic entitlement. As to UK management's approach to recording this in the minutes, see paragraph 1299 below. In his oral testimony, Mr Riley accepted that a member receiving benefit statements in the form sent out by IBM's Pension Trust department would have thought that the pension shown at age 55, 56, 57 etc was what they had earned to date as a pension and that this was so notwithstanding the small print on a different page about early retirement being "at the discretion of the Company". He was taken to a typical statement for 2008 but his answer could not have been any different had he been shown the

Although the material which I have just addressed post-dates the Webcast by a few years, I have no reason tothink that the culture at the time of the Webcast was any different and every reason to think that it was the same.

Whether or not the members had an enforceable entitlement based on contract or estoppel at the date of theWebcast to benefits based on the then current early retirement policy is not an issue which has been raised. But what is clear, I think, is that even in 2004 there was a widely-held perception on the part of members that they would be able to retire before NRD if they wished to do so and that in practice, even if consent was formally required, it would ordinarily not be refused; that was a perception engendered by Holdings which knew that the perception was widely held. It was a Reasonable Expectation. Action by Holdings contrary to that Reasonable Expectation is at least capable of engaging the Imperial duty. The Expectation was, however, the result of a policy and a practice. Policy and practice can obviously be changed for the future in respect of future service. Whether it can be changed in respect of past service is a different question. I do not think that it could be said that the expectation in respect of past service can never be changed without a breach of the Imperial duty. For instance, it would be surprising if Holdings was unable to revise its policy and practice on early retirement in relation to a 35-year-old man with 4 years' pensionable service in the DB Plans. In contrast, it could well give rise to a breach of the Imperial duty to change the practice with only a month's warning in relation to a 58-yearold man with 20 years' pensionable service in the DB Plans. Communication with members (continued)

After that digression, I return to Mr Tennet's submission about the reasonable member's expectation thatHoldings would not change benefits applicable to past service in order to reduce the DB Plans' liabilities. I am concerned here, in reality, only with the change to early retirement policy. So far as benefits which are laid down in the Trust Deed and Rules (in contrast with a discretionary benefit such as an early retirement pension), there is no need to invoke expectations at all. What can and cannot be done is laid down by the Plans' amendment powers under which accrued rights are protected to the extent set out in the fetters which I have discussed in relation to Issues 1 to 3. So far as the final salary link is concerned, nothing now turns on that in the light of my decision on Issues 1 and 2 (although an issue remains, in the context of Project Waltz, about the validity of the 2009 Non-Pensionability Agreements).

So far as early retirement practice is concerned, I have concluded in my digression into "automatic entitlement"first that the reasonable member had a Reasonable Expectation that the pension policy and practice would apply to him if he wished so far as concerned his past service benefits at least but secondly that it is not necessarily the case that a change in policy and practice would amount to a breach of the Imperial duty. A reasonable member hearing the Webcast would, I have no doubt, see confirmation and reinforcement of that Reasonable Expectation. In particular, although he may have appreciated before the Webcast that policy and practice might change, he would, in my judgement, take from the Webcast the message that the deficit attributable to all the benefits (in respect of past service) provided by the DB Plans including those arising under the early retirement policy and practice was guaranteed up to the first quarter of 2014. It follows, in my judgement, that the reasonable member would have had an expectation that the benefits to which that guarantee relates would continue to be provided. I

Reasonable Expectations are, however, no more than expectations. To disappoint those expectations does notnecessarily lead to a breach of the Imperial duty, even in a matter as important to an employee as pensions. Thus, even in the context of the early retirement policy and practice, a significant change in financial and economic circumstances might be such as to justify a departure from the Reasonable Expectation without the Imperial duty being engaged or, if engaged, being breached. Whether the Imperial duty was engaged by the Project Waltz changes and, if it was, whether there was a breach of the Imperial duty are matters to which I

In reaching these conclusions, I do not overlook Mr Tennet's submission (the second submission referred to atparagraph 678 above) that members could reasonably expect that benefit changes in the future would not be justified on the basis that the past service deficit had increased. As a starting point, that may be right, and the onus will be on Holdings to demonstrate why a departure from the Reasonable Expectation would not give rise to a breach of duty. But it is not possible to rule out, a priori, an increase in the past service deficit as a sufficient demonstration of that result. That, as I have said, will need to be assessed in the light of all the circumstances, including the extent to which Holdings can be said to be responsible for any increase in the deficit.

These conclusions are not, I think, in any way inconsistent on the evidence which Mr Heath gave and which Mr

The question he was asked related to past service benefits and his answer was given in relation to past servicebenefits. I do not think that it was anywhere made clear, if indeed this was the questioner's (Mr Tennet) intention, that past service benefits included the "automatic entitlement" to early retirement benefits alleged or that Mr Heath understood the question that way. Even if Mr Heath did understand it that way, I do not consider that his answer was directed at a situation where a significant, unforeseen, change in financial and economic

I do not consider that the conclusions are affected, either, by the awareness on the part of the reasonable member(see paragraph 620 above) of the contents of the 2003 members' report and the explanatory literature issued to members. The members' report tells the reader nothing relevant to the proposed changes; quite clearly, the Webcast and other communications supersede anything said in that report. The explanatory material does refer to Holdings' power to amend the Trust Deed and Rules, but the hearer of the Webcast was not reminded of that and is entitled to take the message about the future of the DB Plans from the Webcast. In any case, so far as early retirement policy and practice is concerned, Holdings must, in relying on the reference to the Trust Deed and Rules, take the rough with the smooth; if a change to policy and practice is to be seen as an amendment within the scope of the reservation in the explanatory material, it should be seen as subject to the same fetters as an actual amendment, in which case the practice and policy could not be changed even in the face of a significant change of financial and economic circumstances.

A separate, but related issue is this. Mr Tennet submits that Holdings' acceptance of responsibility for any deficitalso necessarily meant that the reasonable member could expect that Holdings was doing more than taking responsibility for the funding of any deficit: it was also taking responsibility for any increase in the accounting cost of the UK DB Plans (ie NPPC) consequent on an increase in the deficit (to the extent that they even contemplated the possibility of a difference between the two). Accounting and funding were in effect two sides of the same coin; it would be artificial to suggest that Holdings accepted responsibility for funding, but not for the effects of the DB Plans on its I&E account. I do not agree with that submission. The Webcast said nothing about accounting aspects or the consequences of a continuing, perhaps increasing, deficit. Everything in the Webcast was, as I read the transcript, directed at English concepts and consequences. In my view, NPPC does not come into the picture in relation to expectations at all. Rather, it comes into play as one factor in the assessment of whether a change in benefits (eg the change in pension policy and practice) gives rise to a breach of the Imperial duty. The issue may turn on whether producing a particular NPPC result is a justification for the change. Applying the test which I have identified, the question is whether achieving that result by making the relevant change is something which no reasonable employer would do.

In the above discussion, I have carried out a close textual analysis of the Webcast. I have also attempted tointerpret the various statements in it as the reasonable member might interpret them. I need now to carry out a reality check by standing back to consider the Webcast as a whole through the ears of the reasonable member hearing it on one or two occasions. The conclusions which I reach are, inevitably, impressionistic. But my conclusion is the same as that which I have already reached. I can assure the parties that this is not a "he would say that wouldn't he" assessment. I have given this the most anxious consideration. My assessment is that the central message in relation to past service is that nothing is changing or will change without justification at least prior to 2014: I do not need to say anything about changes after that. The central message in relation to future service is that benefits will continue to accrue. But the reasonable member could not properly hear a message that Holdings had committed itself (albeit in a way not amounting to a legal promise) to future accrual until 2014 although I consider that the reasonable member, reflecting on what he had heard, would be justified in reaching the conclusion that benefits would continue to accrue for at least 3 years. I do not propose to say where the cut-off date for the continuing commitment is, at least at this stage.

Further, I do not consider that the reasonable member would form an expectation that further accrual wouldcontinue for any particular period (except perhaps in the very short term) regardless of changes in financial and economic circumstances (falling short of the changes which even Mr Tennet accepts would justify a change in future accrual). His expectation would be qualified by an acceptance that such changes in financial and economic circumstances could justify a change. Whether one categorises that situation as one where the extent of the Reasonable Expectation is circumscribed by the qualification or as one where a disappointment of that Expectation has no consequence does not matter much, if at all.

Moving on from the contents of the Webcast and its impact, it is, however, convenient to record at this point thatCHQ saw the communications sent to members and approved of them. Thus Ms Gherson was sent the Webcast on 16 November 2004. Her response to David Heath was:

Mr Lamb's letter to members dated 9 November 2004 explained the proposals. He made some of the pointswhich Mr Heath had made in the Webcast. Mr Lamb identified the large deficit (£900m) which had been determined by the Scheme actuary as at the end of 2003. He noted that the deficit was large both in absolute terms and in terms of the resources of the employer. As to that he said:

The figure of £181 m represents the C Plan's share of the £200 m referred to in the Webcast as the Q&As laterexplained. Although the letter does not expressly state this, the £181 m for each of the three years 2005, 2006 and 2007 covers the cost, as then estimated, of future service benefits for that year and the contribution towards the deficit. A member reading this carefully would see that this "compares to £121 million this year" and would know that during "this year" accrual was at the current rate (to state the obvious) and would have no reason for thinking that the three subsequent years would be treated any differently and every reason for thinking that they would be treated the same.

The letter also included the following passage:

Mr Simmonds observes that the latter advice was correct. I agree since these proposals did not break the finalsalary link. Mr Simmonds also adds that the description of the Trustee's role, which was to protect accrued rights rather than to be concerned with future accruals, was entirely consistent with the approach adopted by Mr Lamb throughout the discussions relating to the Project Ocean Changes. Although what Mr Lamb said reflected how he saw the Trustee's role, that is not to say that his letter did not go further: it clearly did. The question is what message his letter sent over and above protecting accrued interests.

Mr Tennet has focused some attention on Mr Lamb's statement that "the deficit is large in absolute terms andlarge in relation to the resources of the plan sponsor" remarking that IBM has relied on those words as evidence of a fuller explanation to members as to why a guarantee only of funding would be such "excellent news" for members; namely that there was in fact some prospect of IBM UK not being able to meet the deficit. In my judgement, the reasonable reader of Mr Lamb's letter would not have taken the guarantee as relating to anything other than funding. Whether IBM's later invocation of the guarantee as being part of what made the Plans sustainable is another matter. But in the context of Mr Lamb's letter, it is reading far too much into it to treat it as a commitment that future accrual will continue.

What is one to derive from Mr Lamb's statement that the guarantee was good news for members? One reason forit being good news was that it allayed any fears there might be about Holdings being unable to meet its commitments.

The RBs, however, say that it was fanciful to think that IBM Corporation would allow Holdings to enter into aninsolvency process and that in practice Holdings would meet its contribution liability. There was no need, therefore, for a purely financial guarantee, so that the guarantee must have carried with it an implication that there would be future accrual at the current rate. One difficulty with that argument is that the major financial problem with the Plans related to the past service deficit. If, as is the RBs' case, Holdings would not have been allowed to default then the past service hole would have been plugged without a guarantee. Why then, I ask, would not exactly the same occur if Holdings had agreed to commit itself to future accrual without a guarantee? Surely, in practice, IBM Corporation would be expected to stand behind Holdings as much in the latter situation as in the former. The reality, it seems to me, is that the guarantee was put in place because this was a requirement of the Scheme Actuary if he was to advise the Trustee to agree to IBM's proposed contribution rate. The guarantee gave legal effect to what the RBs say would have happened in any case. But it was a real benefit to them, as Mr Lamb communicated. And he was able to communicate that message because he knew that the guarantee was essential to IBM's funding commitment and that without the guarantee, there was a fear that IBM would take unpalatable action, even closing the scheme altogether or placing it into winding-up. In other words, the real benefit was precisely to ensure that the Plans did continue albeit without a commitment that they would do so for any particular period.

There is one further point to make in relation to Mr Lamb's letter. In it, he explained the main reasons for thedeficit. The first and second reasons given were (i) returns on plan assets were lower than assumed; and (ii) the assumptions about future investment returns in the 2003 valuation were overall lower than assumed in the 2000 valuation. The third reason was the assumption made for life expectancy which had "a significant effect" on the financial position of the Plans. Mr Tennet relies on the first two reasons and submits that, since those were the present causes of the deficit, members could expect that if further asset falls or changes in asset assumptions increased the past service deficit in future, IBM would continue to meet its responsibilities for any such increase in the deficit. This is the same submission as made in relation to the Webcast; I have dealt with it already.

As part of the consultation exercise, IBM produced a Q&A document dated 8 December 2004. Whether or not itreceived as much attention as the Webcast (Mr Tennet submits it did not) it was part of the communication exercise and cannot simply be sidelined. I have explained what I consider the hypothetical reasonable member would have done with this material above. To some extent, it simply repeated the essence of what Mr Heath had said in the Webcast including the following:

The Q&A document explained the difference between the figure of £181m pa referred to in Mr Lamb's letterand the figure of £200m pa referred to in the Webcast (as to this see paragraph 700 above).

The Q&A document also contained an albeit limited explanation of the guarantee. It raised and answered anumber of questions including the following:

The Q&As confirm, if confirmation is needed, that the guarantee referred to in the Webcast and Mr Lamb's letterto be given by IBM WTC is precisely that, a financial or funding guarantee. I do not consider that it provides a commitment in relation to future service.

However, what the Q&As make clear is that the £200 million contribution (split between the Main Plan and the IPlan) for each of the years 2005, 2006 and 2007, is a contribution in respect of not only the deficit but also in respect of future accrual. This contribution is a commitment by Holdings. It is true that it falls within the scope of the guarantee but that is not relevant to the critical point which I draw attention to, namely that the funding commitment relates to future service accrual for those three years.

Mr Tennet relies on the statement in section 11.1 set out above. He acknowledges that to a pensions lawyer, thatexplanation indicates that the guarantee was merely a funding guarantee. However, he says, it would be perfectly reasonable to assume that, as "future funding payments" covered funding for any future accrual, such future accrual would in fact occur. Again, this is a point made in relation to the Webcast and I have, again, already dealt with it. Taking section 11.1 in isolation, I would reject Mr Tennet's submission for the same reason as I have rejected it in relation to the Webcast and Mr Lamb's letter. But the point is now made in a different context, a context where the very same document contains the explanation in sections 11.2 and 11.4 stating clearly that the contribution promised by Holdings, and guaranteed by IBM WTC, includes the cost of future service benefits. In context, there can be no doubt that the future service benefits are those accruing under the then current provisions of the Plans.

Accordingly, even if the Q&As do not give rise to a promise or legally enforceable guarantee of future accrual atcurrent rates, I have no doubt that a reader of the Q&As would be entitled to take them as an assurance or commitment that future accrual would continue at the current rate for the years 2005, 2006 and 2007.

But equally, I consider that the members are not able to sideline the Q&As. They can take advantage of theQ&As in their favour but they are also to be taken as having a general understanding of the Q&As and in particular of the explanation of the guarantee. I do not detect in the Q&As any support for the view that there is an assurance or commitment of any sort concerning benefit accrual after 2007. It might be said that the fact that the Q&As clearly contemplate further accrual up to 2007 is an indicator that nothing is to change thereafter. If it said that accrual is assured after 2007, the immediate question is, for how long? There is no reason to select 2014. 2014 is relevant to the period of the guarantee but not to the period of a commitment to further accrual on the basis that nothing is to change. The logic of the argument is that accrual is to continue unchanged indefinitely absent a serious change of circumstances: but that is a conclusion which cannot be right. In any case, the contrary argument is that a clear commitment to continued accrual until 2007 is an indication of precisely the opposite result: an express – or at least necessarily implicit – conclusion that accrual is to continue through to 2007 can be said to be an indication that, after that time, there is to be no commitment and that there is certainly no guarantee or assurance.

My conclusion, subject to paragraph 717 below, is that the reasonable member is not entitled to read theWebcast, Mr Lamb's letter and the Q&A's taken together as giving an assurance or commitment that benefit accrual will continue after 2007 at the current rate. This goes further than my conclusion in relation to the Webcast. Even in relation to that, my inclination was to say that the reasonable member would have not seen a commitment to continue accrual after 2007, although he/she (and indeed Holdings) might have hoped that it would. If it did then the Evergreen Agreement would take effect. But I do not see the Evergreen Agreement as inconsistent with termination of future accrual. That inclination is confirmed by Mr Lamb's letter and the Q&A's which make no reference to the Evergreen Agreement.

I do not consider that the contents of sections 17.4 and 17.5 of the Q&As which have been relied on affect myconclusions. Those sections repeat previous comments about commitment and sustainability but do not, in my judgement, take the debate further.

There is one caveat which I must express to these conclusions. Unbeknownst to the members who read MrLamb's letter, the Trustee had approved the Ocean changes only on the express assumption (known to Holdings by the time of the member consultation) that there were no plans for other major changes to the Trust Deed and Rules or to pension practice. The reasonable member's expectation from the Webcast and reading Mr Heath's letter cannot, therefore, have been seen as influenced by the Trustee's assumption. But this factor is one to be brought into account in assessing whether Project Waltz gives rise to a breach of the Imperial duty.

The next documents to be considered are the 2004 Members' Reports for the Main Plan and the I Plan, publishedin July 2005. Both of these contained further information about the guarantee. Mr Lamb's letter of November 2004 was reproduced.

IBM relies on those Reports as reinforcing the message which it submits the previous documents gave. I agreethat the Reports reinforce the previous message, although as already explained, the message which I consider those documents gave is not quite the same as that for which either Mr Simmonds or Mr Tennet contends.

A brief summary of the key features of the Funding Agreement was reproduced in those reports from the Plansvaluation reports under the heading "Key Features of the Funding Agreement contained in the 2003 Valuation Report". After that, there followed an outline of the termination provisions. That outline uses some fairly dense and technical language. The relevant passage is set out in the RBs' closing submissions: this was done, it is said, merely to show the difficulty members would have faced in understanding what it meant.

Whatever difficulty there may have been in understanding that outline, I agree with Mr Simmonds that thereasonable members interested in the nature and content of the guarantee (which our hypothetical reasonable member would have been) would have read the section of the Report under the heading I have just mentioned. His submission is that it provided a clear and accurate summary of the principal terms of the Funding Agreement (this section of the Report does not actually use the word "guarantee" at all although the previous section, replicating the November letter, did of course do so). A fair reading of this summary in full would, he suggests, have informed the reasonable member of the following:

Mr Simmonds states, correctly, that the Report describes the circumstances in which the Funding Agreementwould be terminated prior to 31 March 2014, including where the Plans were no longer in deficit on a prescribed basis, or where the UK employers had sufficient financial resources to meet any funding shortfall on a prescribed basis. It is clear to me, notwithstanding Mr Tennet's submissions, that the outline could not be read by the reasonable member as supporting the idea that the Funding Agreement (or the guarantee which formed part of it) was a guarantee of future benefit accrual or, indeed, that the Funding Agreement was concerned with future accrual at all.

Both sides are correct in thinking that the Report is consistent with their positions. In that sense, if the previousdocuments showed that the RBs are correct or that IBM is correct, then the Reports would underline the correct position. Equally, the Reports are entirely consistent with the approach which I have concluded is actually correct in relation to the earlier documents. In the end, I do not think that the Reports assist.

It is fair to say, as Mr Simmonds does, that there was no significant opposition to the proposals amongst themembership during the course of the consultation process. However, as a result of the consultation, one amendment was accepted by Holdings. That amendment (which was notified to the Trustee at a TMM held on 9 December 2004) was to give N Plan members a one-off chance to elect to make contributions of 2% in respect of the increased cost of future accruals and, in return, to continue to accrue benefits at the existing rates, rather than have reduced accrual for future service. Subject to that amendment, IBM decided that it wished to proceed with the proposals. They were approved by the Trustee at the TMM held on 9 December 2004 just referred to and were implemented with effect from 6 April 2005. At that same meeting, the Trustee approved the wording of the relevant documentation effecting the changes and including in particular the Funding Agreement which contained the guarantee.

Mr Tennet submitted in his written opening submissions that IBM was undoubtedly aware that members wereleft with the impression that the plans would stay open in future for the long term. Thus he refers to Mr Newman's report to Mr Heath on 9 November 2004:

I do not consider that one can take from this the message that members thought the Plans would remain open forthe long term. The fact is that there was, as I have said, a concern at the time, and there had been press speculation, that IBM was actually considering closing the Plans. The Webcast and Mr Lamb's letter made clear that that was not going to happen. The members were able, like Mr Lamb had been, to breathe a sigh of relief. But they were not entitled on the basis of what they had been told to conclude that the issue would not be revisited. They were not entitled to think that the Plans would remain open for the long term and not even, in my view, in the medium term. members may well have hoped that what they were told would turn out to be a longterm solution and that the Plans would remain open for the foreseeable future. And they may well have felt disappointed (many certainly were) and even angry when the Project Waltz proposals were presented. They may well have felt equally disappointed and angry if the Ocean proposals had never seen the light of day and Project Waltz had been the first change. The issue is not whether they were disappointed and angry; nor, indeed, is it whether they believed, rather than merely hoped, that the Plans would remain open, although that is a prerequisite to their claim. The issue is whether Reasonable Expectations were engendered by the communications from Holdings on the basis of which they were entitled to form such a belief.

At this stage, I wish to refer to one point on which Mr Tennet relies and to flag one general point raised by MrSimmonds. Mr Tennet relies on what Mr Heath told me, as Mr Tennet puts it, that he had received "a lot of feedback" from many members over a long period of time after the Webcast, and that he was well aware, even at the time of the 2005 Ocean changes, that a number of members (even if not, in his opinion, the majority of members) were under the impression that IBM had guaranteed that future benefit accrual would continue unchanged. That is a slightly tendentious way of putting it. The exchange went like this:

And in relation to what members thought the formal guarantee referred to in the Webcast meant, there was thisexchange:

That is not a very strong base, I venture to suggest, from which to assert that IBM knew that "many members"were under a misapprehension, let alone that the hypothetical reasonable member would have held such a belief. Nonetheless, Mr Tennet goes on to observe that when Mr Heath was asked at the Project Ocean and Soto roadshows whether the guarantee provided such an absolute promise, he always put members right. Again, it is worth setting out the exchange:

Mr Heath did not, there at least, answer the question about how the Webcast could reasonably have beeninterpreted. That, in any case, is a matter for me. Mr Tennet refers to Mr Heath's answer, however, as showing support for his submission that Holdings knew that many members were under a misapprehension. I do not understand why it does so. The fact that a member asks a specific question does not mean that he had misunderstood the message which Mr Heath was intending the Webcast to convey. The questioner might have been unclear about the message and to have been simply seeking clarification, thus demonstrating at most ambiguity in the Webcast. Or he might have understood the message as being that which Mr Heath intended and

The general point which Mr Simmonds makes (although it is jumping ahead a bit) is this. If and insofar as themembers had gained the impression that there would be no further changes to benefit structures for a particular period, it is clear that the announcement of the Project Soto changes ought to have led them to question that impression or, at the very least, to reflect on the fact that IBM did not share that impression as to what had been agreed or committed to by it in 2004. He says that the possibility of future changes was then clearly communicated. Having flagged the point, I park it for now.

Just to complete the narrative of the Ocean proposals, after completion of the employee consultation, the Trusteeresolved, at a TMM meeting on 9 December 2004, to implement the changes, and approved the wording of the necessary Deeds. At the same meeting, the Trustee then considered the investment strategy in the light of the guarantee, with the assistance of the Scheme investment adviser Mr Pardoe. The minutes record that Mr Pardoe explained that the investment risk within the Main Plan was high, relative to the current liability profile and the size of IBM UK, and also record that:

The Trustee ultimately resolved to move 10% from equities to bonds over the next 3 years. As Mr Lambexplained and I accept, that decision was on the basis of the Trustee's understanding that it "had reached a bargain with IBM that would secure the sustainability of the DB pension schemes".

The Funding Agreement was executed on the same day as the TMM, 9 December 2004. I have already set out itsessential provisions: see paragraph 111 above. Accordingly, the Funding Agreement had achieved CHQ's twin objectives of removing the margins for prudence in the funding assumptions and of persuading the Trustee to

It is to be noted that the Funding Agreement does not refer expressly to a commitment to make contributions of£181m/19m to the Plans over each of the years 2005, 2006 and 2007. Those figures arise out of the actuary's advice that, if the guarantee were provided, he would be able to recommend contributions at that level.

I have described the Project Soto changes in paragraph 112-115 above; they are to be found at Annex C to thisjudgment. The changes need to be considered in the context of the Trustee's investment policy which, in turn, reflected Holdings' strategy as a result of the adoption of the Ocean changes. It will be recalled that the Trustee had wanted to switch gradually from what it saw as overexposure to equities, particularly UK equities, into bonds (of which there were, of course, a variety of types including index linked bonds, UK fixed interest bonds and global bonds). Consideration was given to the type of bonds best suited to the Plan at an Investment

IBM operated a Global Bond Fund ("GBF"). As appears from the minutes of the Investment Committee meeting on 17 February 2005, of the £150m received from Holdings in December 2004, £114m had been invested in the GBF. A further £200m, representing the proceeds of other investments which were being realised, was due to be invested in it. The result would be that the Plans would hold 25% of the GBF.

The subject of the bond allocation was raised again at the Investment Committee meeting on 20 October 2005. Watson Wyatt advised that the "duration of the bond assets in the Fund…is significantly lower than the duration of the liabilities". The duration of the total bond portfolio was about 8 years, whereas the duration of the liabilities was around 16 years: if interest rates were to fall by 1%, the Plans' liabilities would increase by around 16%, whereas the value of the bonds (which were under 30% of the total assets) would rise only by around 8%. In other words, there was not a matching of assets and liabilities. CHQ were, I have no doubt, aware of this issue. Watson Wyatt also advised that a significant proportion of the liabilities were inflation linked, meaning that, since much of the income from the bond portfolio was fixed, the Plans were exposed not only to interest rate changes (via the mismatch in duration), but also to inflation rate changes.

Mr Tennet submits that the obvious solution to the interest rate exposure identified was to switch into longerbonds. However, the Trustee by then had no control over the duration of that part of its bond allocation invested in the GBF so long as it remained invested in that fund. And, as Mr Martin Jack (Director, IBM Retirement Funds, EMEA) told the Trustee at the 20 October 2005 meeting: "the purpose of the GBF is not intended to match liabilities but is to maximize alpha." As Mr Tennet correctly observes: "In other words, IBM did not view the GBF as designed to match the Plans' liabilities, but instead it viewed it as an opportunity to seek outperformance".

The strategy which Mr Tennet suggests may be one way of dealing with the mismatch. But it could only ever bepart of the strategy. In any case, the appropriate investment policy for a scheme depends on a number of factors of which matching is only one. Moreover, it is accepted that the Trustee's actual investment policy was within

The Trustee's view was that no immediate action was required: as Mr Lamb observed, "the guarantee has giventhe Trustee a period of time in which to put in place a structure to achieve better liability matching." According to Mr Tennet, it is plain from that that the Trustee's approach was entirely founded upon its assumption that if the risk did not pay off IBM would not withdraw its support of the DB Plans by ceasing future accrual. That goes too far: I agree with what he says but only with the omission of the words "by ceasing future accrual". The benefit of the guarantee was not dependent in any way upon future benefit accrual. It was not even dependent on the Plans staying out of wind-up and continuing as closed schemes since the Funding Agreement provided protection in that event.

The Soto changes also need to be considered in the context of the actual and projected NPPC figures whichwere of importance to IBM Corporation (and CHQ) in meeting its targets for EPS. The RBs' written opening (which I do not think is challenged in this respect) contains a table (at paragraph 184) which shows that UK

Further projections by CHQ in early January 2005 were that 2005 NPPC would reach $110m. By 17 January

It was this increase which, according to the RBs, was the trigger for the Project Soto changes. IBM has adifferent perspective, suggesting that the underlying causes of Project Soto were twofold. First, it was a reaction to the impact of the working into the NPPC figures of the bursting of the 'dot com' bubble in 2002 and 2003. Secondly, it was a reaction to the fall in bond yields experienced in 2005. On either view, however, it was the increase in the NPPC figure which was the catalyst for the change.

From CHQ's perspective, the objectives of Project Soto were threefold:

It was not suggested at the time, and I do not believe it is suggested now, that operational performance of thebusiness within the UK was a driver of the Soto proposals. There is no evidence to suggest that such

There is no doubt that the impetus for the Project Soto changes came from CHQ rather than the UKmanagement. In March 2005, Mr Koppl gave a presentation (his paper "Retirement Related Expense – February

Mr Tennet also makes reference to the fact of CHQ's awareness of the potential problems with DB schemes atthis time. As it was put by Mary Barton in an email to Mr Koppl dated 31 October 2005:

It was in response to these projected cost increases that Project Soto was initiated. Its purpose was to analyse foreach relevant country (including the UK) the potential benefit redesign options that might be available and the financial, HR and legal implications of those options. Work on Project Soto began with a meeting at CHQ on 14

Now, it might be said that CHQ had no business opening up a review of the UK DB Plans in the light of whathad been communicated to members at the time of the Ocean changes. I would not agree with that because CHQ was, quite reasonably, investigating all options in the light of the circumstances (in a rapidly changing environment) in March 2005. The review included consideration of the legal implications; there could be no suggestion at that stage that CHQ was set on a course of action even if it might be of doubtful legal validity.

Slide 8 also shows that the proposals would be devised by CHQ and then IBM would work with countrymanagement to implement them. Objective advice based on external legal expertise was to be obtained. The process, which would involve consideration of the HR and legal position, is fully recorded in a paper prepared at the end of July 2005 by Towers Perrin who were engaged by CHQ to undertake analysis of the options.

Work was done on a further presentation by Mr Koppl (given on 12 August 2005) in which cessation of DBaccrual was one of the options for consideration, as was reducing or eliminating early retirement "subsidies". CHQ had been heavily involved in the Ocean project less than a year before and yet there was no mention of the selling points which had been made of sustainability, security or the firm footing on which the Plans had allegedly been placed. Nor was there any mention of the "automatic entitlement" to early retirement which many members reasonably believed they had (which belief Holdings certainly knew about and which is difficult to imagine CHQ did not know about). The aim was stated as simply being "to reduce 2006 expense"; reference to volatility had disappeared. Perhaps Mr Koppl did not know what had been said to employees in the UK. Nor was any of this mentioned in Towers Perrin's paper "Initial Findings" dated 16 August 2008.

Before local teams were informed about Soto, CHQ carried out further analysis, and considered how it wasgoing to present the Soto initiative to local management.

I ought at this stage to mention some of the communications leading up to the first meeting with local UKmanagement.

On 17 August 2005, Ms Karen Salinaro sent an email to Mr MacDonald. She explained that she had spoken toMr Michael Burkhardt (HR leader for Northeast Europe), Mr Joe Severi (VP Human Resources, South West

Mr Koppl gave his "Work Plan" presentation on 18 August 2005 which showed among other things that CHQstill had work to do on its analysis of the savings from freezing future accrual and other design options. The cessation of accrual was thus an option which had not, at this stage, been ruled out.

On 1 September 2005, Towers Perrin produced their draft work in progress report in support of the designdiscussion. For each plan, the covering email explained that they had assessed, among other things, the possible employee relations implications. The column which had appeared in a previous version quantifying the size of the potential saving had been removed since it had been agreed (presumably with CHQ) that "this information will not be shared, at this stage, with the local IBM teams". Towers Perrin identified one problem: cessation of DB accrual "Might be viewed as a breach of trust by employees".

Further, Towers Perrin were working on the assumption (since shown by my decision on the RectificationAction to be incorrect) that there was no right to retirement with an unreduced pension from age 60. Nothing was said, so far as I can see, about changing early retirement policy and practice although it must have been implicit in the option to restrict post-60 early retirement subsidies.

The possibility of avoiding the perception of breach of trust would not have been helped if the UK Team hadknown of an email sent on 30 August 2005 by Ms Salinaro to Mr Koppl commenting on a briefing paper prepared by him for Mr Greene:

As I understand, Mr MacDonald had communicated the message in the first sentence just quoted. But theremainder of the passage is Ms Salinaro's explanation to Mr Koppl of CHQ's thinking which the UK Team would not have known about. On that basis, Mr Tennet submits with some force that this was a:

One might add that it does not seem consistent either with the still recent communications in the context ofOcean including not only "commitment" and "sustainability" but also an apparent undertaking to make contributions for 2005, 2006 and 2007 in respect of future service (as already explained).

Coming to communication with the UK team, on 31 August 2005, Ms Salinaro emailed Mr Burkardt, Mr Severiand Mr Slocum with an outline of the next steps for Project Soto. She asked that her email be forwarded to the key people in the UK who would be working on the project. The email set out a tight timeline (a "very rapid process" as she described it), which envisaged approvals, including the approval of the Trustee, would be obtained in the week of 28 November 2005 so that the proposals would be ready for implementation by the end

Following that, the UK team was informed of the Soto project, Mr Heath being forwarded that email on 7 September 2005. There can be no doubt that the UK team's initial reaction was one of great concern especially, as Mr Heath describes and I accept, because these changes came so soon after the Ocean changes. Mr Wilson himself shared that concern as he explained and I accept. He had been sent an email on 31 August 2005 by Mr Jack with his preliminary thoughts and containing NPPC projections and an analysis of the effect of certain changes in the design of pension provision.

It was intended that the meetings with local management, the so-called 'kick-off' sessions, in the severalcountries would take place in mid-September 2005. The first kick-off session for the UK was booked for 13 September 2005 at Bedfont Lakes. The proposed attendees included the UK team (which included Mr Heath, Mr

Before that meeting, an interim conference call was planned for 9 September 2005 to take place between the UKteam, personnel from CHQ, Mr Stark, Towers Perrin and Jones Day (IBM's US lawyers). Prior to that call, Mr Stark had forwarded to the UK team a one-page summary of possible design options from Towers Perrin, which listed some options and the possible US GAAP (ie NPPC) effect of them. Mr Stark did not send, and nor did the UK end receive, Towers Perrin's full report. Mr Heath's response (to Mr Wilson, Ms Middleton (a member of the UK legal team) and Mr Hirst) was: "I think we have already looked at most of these but ought to include our view of doability in our pitch for Tuesday [13.09.05] on any we have not already considered".

The UK team was not happy and continued to express concern about cessation of DB accrual. Thus on the sameday, 9 September 2005, and before the call took place, Mr Waller sent CHQ the UK team's initial "UKI Soto Initiatives" charts, which broadly corresponded to the list of options in Towers Perrin's one-page slide to which I have just referred. The slides reminded CHQ of the actions that had been taken during Ocean. In particular, one slide emphasised that closing the plans to future accrual or compulsory transfer to DC "Would be viewed as breach of trust. Contrary to message delivered end 2004/early 2005" and that "Benefit reduction not consistent with earlier message".

These matters were raised in the UK Workshop on 13 September 2005. When cessation of DB accrual wasraised, Mr Heath said that he was worried about the employee relations impact and it would be a "basic breach of trust given l.y." which I take to mean "last year" and to be a reference to the Ocean changes.

It is not without significance that Ms Salinaro's note records two items. The first, at the beginning of the note, isthis:

The second item appears later in the note:

This is confirmation, if it is needed, that CHQ's starting point was the delivery of NPPC savings; and that wasnecessary in order to maintain EPS. The real reason for the changes was driven by that consideration and the Soto proposals were a solution to that problem of maintaining EPS and, possibly, a desire to get rid of DB Plans as a matter of policy. The search was then for a justification for the changes which could then be presented as reasons for the changes.

The upshot of the meeting was that the possibility of cessation of DB accrual was not abandonednotwithstanding the UK team's concerns and its presentation; and it was agreed that both voluntary and compulsory changes from DB to DC for future benefit accrual would be analysed. It was also agreed that consideration would be given to future COLA.

Following the launch of Soto, with its objective to reduce the year-on-year increase in NPPC for 2006, MrWilson started work on establishing what that year on year increase was. He started with an estimate of $199m.

The figure of $82m is found in IBM's paper dated 23 September 2005 – "Project Soto – Redesign of PensionPlans" (as to which see paragraph 776 below).

Mr Tennet says that one might have expected IBM to see this as good news. But Mr Stark had a different takeon Mr Wilson's email in his response to Ms Salinaro on 17 September 2005: "You were right when you said that setting a # could distract focus… we have a unique opportunity now to address the costs of pensions – both the level and the volatility – to solve this for the long term once and for all…..". Further, at this time, there was a communication, for the first time to anyone in the UK, that the Soto target was not simply the year-on-year increase for 2006 but a different objective, the long-term overall cost of the Plan: see Mr DeAngelo's email to Mr Wilson on 19 September 2005.

The paper "Project Soto – Redesign of Pension Plan" to which I have just referred is marked as "Prepared forIBM Attorney". The slides describe an "Extremely aggressive timeline" and stated that [we] "Must take into consideration country-specific challenges but we must recognise cost decreases for I&E purposes by beginning of 2006", no doubt in order to meet NPPC targets. The slides also state that a communication plan to members would be developed. It would convey a key message that the changes would "Allow continued necessary investment in other compensation programs, salary, variable pay". One might think that the message here was to be that changes to the Plans would be followed by more money being made available for pay increases, something which did not happen.

Slide 31 of the draft presentation for Mr MacDonald and Mr Loughridge made the UK's concern about the

At this stage of the development of the proposals (mid-September 2005), CHQ was not entirely happy with theapproach of UK management which it did not regard as sufficiently enthusiastic about the proposals. CHQ therefore wanted to keep control of the Project Soto Pension Briefing presentation: as Ms Salinaro put it, it was better if CHQ were controlling the document so that it could be directed appropriately. By 25 September 2005, Mr Greene and Ms Salinaro were discussing whether to express their "concerns about UK management" to other managers. There was frustration on the part of Ms Salinaro, expressed in her email to Mr MacDonald dated 28 September 2005, that "the UK team does not seem to feel the same sense of urgency" [the comparator country has been redacted from the trial bundle] and in her email of the same date to Mr Burkhardt where she complained that:

Mr MacDonald distanced himself from Ms Salinaro when giving evidence. Perhaps her thinking is not to beattributed to CHQ in all its starkness. But it is the case that, at this stage, she at least was not just considering cessation of DB accrual: unknown to the UK team, Ms Salinaro and Towers Perrin had also been considering winding up the plan. This was less than one year after Ocean, and demonstrated some insouciance with regard to the concerns about breach of trust of which she must have known.

During the late summer and autumn of 2005, various discussions were held. In particular, two review sessions,involving Mr Loughridge and Mr MacDonald, were scheduled for 28 September and 25 October 2005.

The first of these sessions was attended by 25 people from CHQ, the UK, NE IOT, and Towers Perrin. The UKmanagement's recommendation was that further work be done on two of the identified options, namely (1) voluntary transition to DC with compulsory capping of pensionable DB pay growth and (2) voluntary transition to DC. There was also a third option, reduction in discretionary COLA.

Although the UK management did not close its mind to compulsory cessation of DB accrual, it subsequentlyconcluded that it would be the wrong thing to do, certainly from a commercial perspective. For that reason, it recommended the two voluntary transition options. It also recommended that further work be done on COLA: the UK management's then recommendation was to restrict such adjustments in respect of pre-April 1997 service to 50% of UK RPI (subject to a cap of 5%). In an email dated 7 October 2005 following that meeting to Mr Castellanos, VP CHQ HR, concerned in large part with the time by which the proposed changes could be implemented, Mr Heath again explained the difficulties that the Ocean changes had created, but that he thought the UK's proposal was achievable:

Ms Salinaro then asked Mr Heath a number of questions about the proposals. His response made it clear that theUK team thought that the proposals would address what they were told were CHQ's two key aims of NPPC reduction, and NPPC volatility reduction:

According to IBM's written opening submissions, none of these options would have achieved savings sufficientto offset the estimated increase in NPPC of $92m. Accordingly, the UK management was asked to continue its work in order to identify other options that would save $92m, to consider alternatives that would reduce longand short-term cost and volatility, and to consider an acceleration of the timetable for implementation. The RBs' written opening submissions contend that the effect of such changes would have been to reduce NPPC by $100m for a full year, and up to $88.5m for 2006, depending on when the changes were implemented, and thus it would have been enough to meet IBM's $92m requirement.

It is the case, however, that Watson Wyatt's projections dated 21 October 2005 sent by the UK team to CHQshowed that the current proposals would make substantial reductions in a full year NPPC, that NPPC would be far less than IBM's service cost in 2008 and that NPPC would turn negative in 2010. There is some reason to think therefore the UK team thought that the concerns about NPPC were being met.

However, unknown to the UK team, CHQ had been considering other options. Ms Salinaro and Towers Perrinhad even been considering winding up the plan with serious consequences for members in many ways including their loss of COLA and early retirement entitlements. Towers Perrin prepared presentations "Project Soto – Creative Strategies" for 23 and 25 September 2005 identifying those two serious disadvantages and noting the substantial reduction in benefits (and thus cost, I would add) in the event of winding-up.

Significantly, Mr Potgieter of Towers Perrin wrote to Ms Salinaro and Mr Koppl on 27 September 2005. Therelevant part of his email and her reply are set out at paragraphs 970ff below.

In those emails we see an acknowledgment from IBM's own advisers that the Soto proposals would not solve thefuture volatility problems. The type of volatility being mentioned here was volatility which has an impact of IBM's P&L account, that is to say NPPC volatility. It is also recognition that such volatility was related to the past. Mr Koppl agreed with that, adding that the service cost only added about 1% each year to the liabilities. His evidence was that NPPC volatility was the result of the large size of the pension funds and the large amount of the PBO. And yet we have Mr Heath saying that the proposals following on from the first review session were seen as reducing volatility and thus meeting one of CHQ's objectives. I am sure that he genuinely did perceive the proposals as having that result, but perhaps he was focusing less on past service benefits and more on reducing any further volatility arising from yet more defined benefits accruing in the future.

There is some further email correspondence which throws some light on IBM's approach. On 16 October 2005,Mr Stark (Europe HR) raised a concern that DB transfer to DC would expose IBM to funding costs, since the DB contribution schedule would remain in place. Mr Wilson's response was to agree that "shifting people to DC will incur additional cash funding costs over the next 2 years" but warned that if cash funding was a concern the whole proposal would have to be rethought. He explained that he thought cash funding was not a concern:

Ms Salinaro considered that "real long-term savings" were the "first priority" and that "volatility is next and thatwe have some amount of flexibility to use cash to help us get to the optimal solution". I am unclear how this expression of view fits with her response to Towers Perrin (see paragraph 971 below) that volatility was not really addressed by the proposals. As to savings, Watson Wyatt produced some projections which followed shortly on 21 October 2005. I have found the tables very difficult to follow. But the position was, happily for me, summarised in IBM's own graph produced shortly afterwards, using very similar figures. The graph shows NPPC falling from 2006, and turning negative in 2010. That chart is reproduced here, with a line showing IBM's Service Cost after the changes if 35% of DB members decided to move to DC:

The second review session took place on 25 October 2005. The recommendation of the UK team was toincentivise members to transfer voluntarily from DB to DC accrual by providing enhanced DC terms (in the form of increased employer contributions) and by reducing the attractiveness of DB (by restricting pensionable earnings increases).

In more detail the proposals were:

It was estimated that there would be a 35% take up of the incentivised switch. This would give rise to total costsavings on a full-year basis of $96m but, because the changes would not be implemented until part way through 2006, the actual saving for that calendar year was projected to be $61m. It has been repeatedly stated on behalf of IBM that this was not a target. But both CHQ and UK management hoped for a high take-up rate. In the event, the transfer rate was only in the region of 20%.

On 14 November 2005 there was a third review meeting at which a finalised package of changes was presented.Mr MacDonald came to the UK for that meeting. He gave his approval for the proposals to be put to the Trustee, but these were subject to a number of matters which can be found under the heading "Rules" in Ms Salinaro's manuscript notes on the front page of her copy of the pensions briefing paper for that meeting:

The specific proposals to be put to the Trustee were:

The revised cost savings were $100m on a full-year basis but only $70m for 2006 because the Project Sotochanges were to be introduced part way through 2006. It was at this session that approval of the package was given by CHQ and the UK management was given the go-ahead to approach the Trustee. The board of Holdings does not appear to have been involved in any of this. This is surprising even given the powers reserved structure.

I turn next to the communication and discussion with the Trustee.

On 15 November 2005, the day after approval by CHQ and the UK team of the proposals, Mr Wilson and MrHirst met with Mr Lamb and Mr Newman to discuss those proposals. On 16 November 2005, Mr Wilson emailed a note of the meeting to CHQ. He recorded, among other things, that Mr Lamb was "very concerned about how this could be articulated against backdrop of last year's actions"; and that he, Mr Wilson, had confirmed that "accounting of pensions costs under US FAS was an important consideration for the company". Mr Lamb has confirmed in his own evidence, which I accept, that he was "upset at the time that the Company was proposing yet further changes so soon after the 2005 changes had been implemented" and "said to Mr Wilson that if the Company was intent on pushing ahead with the changes I felt that my position on the Trustee board might become untenable and I would need seriously to consider resigning".

It seems that CHQ was surprised by Mr Lamb's reaction ("stunned by the severity of the position taken by theTrustee Chairman" according to Mr Hirst in an email from him to Colleen Arnold) although UK management considered him to have reacted reasonably.

The next event was the production by the UK team of a briefing paper for the Trustee to be presented at theforthcoming TMM meeting on 8 December 2005. The paper focused on NPPC savings which Project Soto was intended to achieve. The introduction and background records:

The paper went on to explain that the increase in NPPC came at a time when competitive pressures on the UKbusiness were increasing, and that:

The TMM met on 8 December 2005 when the proposals were presented by Mr Hirst and Mr Heath. It is fair tosay that the board members of the Trustee were fairly hostile to the proposals. Mr Hirst explained (and it is accepted that he was acting in a straightforward and honest way) that the changes were to address not just cost but volatility. And Mr Heath said the same thing. They did not, it is fair to infer, have the benefit of Ms Salinaro's thinking.

Mr Heath admitted that IBM had no new information which it did not have at the time of Ocean except that themarket place was even tougher than had been anticipated, saying that with the benefit of hindsight, it would have been better to have brought forward the Soto proposals at the time of the Ocean proposals. More significantly, Mr Greene suggested that the proposals were being put forward because the fundamental problem was not an accounting one but was related to other factors such as longevity: but, as Mr Lamb pointed out, longevity had been addressed under the Ocean proposals by increased contribution rates and the sharing of future increased costs 50/50 between IBM and members. Mr Greene also argued that there had been a cost which had been built up over time, but here Mr Lamb pointed out that Ocean had been explicit about IBM's commitment to take responsibility for the past service deficit. I agree with Mr Lamb's ripostes. I also agree to this extent with Mr Tennet when he says that CHQ was attempting to resile from statements made just one year before: the purported justification given by Mr Greene was no justification (if one were needed) at all.

The Trustee's response to the proposals was set out in a letter dated 13 December 2005 from Mr Newman. Theletter was written to ensure that there was no misunderstanding about the oral response which Mr Hirst and the four conflicted directors had been given at the TMM on 8 December 2005. Mr Simmonds notes that Mr

The letter went on to explain the advice which the Trustee had received from leading counsel to the effect thatIBM had no power to break the link, for past service, with final salary and that, ultimately, the question could only be resolved by the court. Mr Newman's letter was constructive, however. After noting the Trustee's careful hearing of Mr Hirst's explanation of the business challenges facing IBM, he wrote this:

Mr Simmonds remarks that "financial distress" was not an expression that had been used by IBM to justify itsproposals. That, however, seems to me to miss the point being made which was that a bargain had been made and that IBM was being required to jump a high hurdle to get the Trustee on side.

A further letter was sent by Mr Newman on 22 December 2005 that referred to the elements of the bargain thathad been reached in 2004. Mr Newman identified the elements of the bargain as follows:

These are all matters I have considered at length in addressing the Ocean changes.

The UK team reacted to the Trustee's responses to the proposals presented on 8 December 2005 by makingcertain revisions. The team reformulated the restriction on pensionable salary increases from a cap at the rate of inflation up to a maximum of 5% to two thirds of salary increases granted. The UK team was, by now, projecting savings of $110m for a full year and of $75m for the part-year 2006.

This revised proposal was put to Mr MacDonald on 31 December 2005; he approved it. On 5 January 2006, theteam took what Mr Wilson referred to as "our senior leadership team" through the changes in "a long and tough meeting". On 6 January 2006, Mr Wilson sent the revised proposal to Mr Lamb.

The longer matters dragged on, the smaller the savings would be for 2006 so that, from IBM's perspective, therewas a significant degree of urgency. IBM needed to obtain the approval of the Trustee to the changes and so far as possible to get the DB members on side. From what I have said already, it is apparent that the UK team knew they had a difficult task to persuade the Trustee and the scale of that task can only have been underlined by the TMM meeting on 15 December 2005 and the subsequent letters from Mr Newman.

In order to resolve the apparent differences between IBM and the Trustee, it was agreed that there should be ameeting between, among others, Mr MacDonald and Mr Lamb. Mr MacDonald was, of course, very senior and no doubt Mr Lamb considered that if Mr MacDonald agreed something on the pensions front, that is what CHQ would adhere to.

The meeting, attended also by Mr Hirst, Mr Wilson and Mr Newman among others, took place at the MarriottHotel in Portsmouth on 8 January 2006 (a Sunday). At that meeting, Mr MacDonald agreed on behalf of IBM that revised proposals would be put to the Trustee at its next meeting (on 19 January). The two revisions offered were:

Mr Newman made a note of the meeting. In that note he recorded that:

Mr Tennet does not challenge this statement of Mr MacDonald insofar as it relates to his understanding at thetime, although he submits that there were people in CHQ who knew it was not the case because the proposals would not solve the volatility problem. What is more contentious about what Mr MacDonald is alleged to have said relates to his use of the phrase "push back" and references to "sustainable". In a later email to Mr Heath, Mr Lamb described Mr MacDonald as telling him, at the Marriott Hotel meeting, that "he had no plans for change and would push back on anyone who wanted to revisit this topic during his watch".

Mr Tennet has spent a considerable amount of energy in addressing what Mr MacDonald said at this meeting,and in particular, the use of the words "sustainability" and "push back" and what they mean.

Mr MacDonald's understanding, as he explained in his oral evidence, of the word "sustainability" was that itmeant something that is ongoing and consistent. He seemed to accept that that would be so not only if "things happen to go well or if things don't change", but also "even if things do change or things don't go quite so well". He later qualified that answer saying:

Mr MacDonald described what he was doing was to give Mr Lamb some assurance but not any sort of

Nonetheless, he acknowledged that the assurance Mr Lamb was seeking was to give the Trustee confidence that,once the Soto changes were made, the Plans would remain open for the long term. That is a rather different assurance from the one that Mr MacDonald now perceives himself as intending to give. In the light of that difference, I asked Mr MacDonald a question which was essentially why he thought Mr Lamb was happy with what he had been told if he had not been given the assurance which he was seeking. The answer was this:

He did not challenge the suggestion that Mr Lamb in fact took away a very different impression but did notthink so at the time: "Now I would, but I didn't at the time".

Mr MacDonald says that he does not think he used the word "sustainable" and that Mr Newman's note was aparaphrase of what he said. I do not think, in the end, that he challenged the conclusion that what he actually said conveyed the same message as sustainability. Similarly, he acknowledges that he told the Trustee that although he could not guarantee anything for the foreseeable future, "things were going to be okay".

What I found very surprising about Mr MacDonald's evidence in relation to sustainability is that, so far as hewas concerned, volatility did not come into the picture at all. The whole discussion, in his mind, was about funding; and reference to a "firm footing" was reference to a firm financial footing. What is clear, however, is that Mr Lamb (and indeed the UK team) thought that volatility was an issue with CHQ (as it was) and that the discussion they were having with IBM through Mr MacDonald related to the issues which had been the subject of previous communications and presentations, including volatility. That was a not an unreasonable basis for them to proceed on.

As to "push back", Mr MacDonald accepts that he used those words as well as "on my watch". And althoughthis does not appear in Mr Newman's note, there is no dispute that Mr MacDonald told Mr Lamb directly "if Sam [Palmisano] or anyone else asks me to revisit this topic during my watch I will push back on them". Mr Palmisano was, at the time, the man who Mr MacDonald described as the "top man" in IBM.

There is, therefore, a large amount of common ground. The dispute is about what Mr MacDonald meant when heused the words "push back". The RBs says that, in context, it can only have meant "resist". Mr MacDonald did not accept that the ordinary meaning of the words was "resist"; still less did he accept that he used them in that sense. What he told me was this:

I am bound to say that, in England at least, the ordinary use of those words as a noun is "resistance" and as averb "resist". I do not perceive anything in the context in which Mr MacDonald used the words that it should be given a different meaning. It is not unhelpful to consider what those present at the meeting actually understood Mr MacDonald to be saying when he used those words. The evidence of all of them other than Mr MacDonald himself, namely Mr Lamb, Mr Newman and Mr Wilson (as well as Mr Hirst in his witness statement), was that

In any case, if the words mean what Mr MacDonald says he intended them to mean, they really add nothing towhat he would, or ought to have done, anyway. Suppose he had said to Mr Lamb that he would test and challenge any suggestion to make further adverse changes before making a decision. Mr Lamb could properly have thought to himself that that is precisely what Mr MacDonald ought to do and would have derived no additional assurance from the words at all. I do not need to decide what it was that Mr MacDonald actually meant at the time. What is clear, as I have said, is that the impression which he gave and which the other people at the meeting were entitled to gain was that Mr MacDonald would resist any future changes which were inconsistent with his understanding of sustainability. Mr Lamb and the others present were entitled to understand Mr MacDonald as giving an assurance – not a guarantee – that the Plans would remain open in the current form for longer than merely in the short term. He clearly did not give a guarantee: the most that he could have been projecting was a level of confidence that things would not change.

That conclusion about Mr MacDonald's assurance must be qualified, however. He was not giving an assurancethat, if the external financial or trading environment changed significantly, the level of pension provision would not be revisited. What he was saying was that if the environment did not change significantly, nor would the Plans. My use of the word "significantly" is deliberate. It is intended to reflect the test which I have explained in relation to the Imperial duty. A minor change in the business environment could not be relied on as justification for a change to the pension provision in the short or even medium term. To react to such a change by reducing benefits could be "perverse or irrational" in the sense explained.

The extent to which any of this is relevant is a matter for debate. I only note for the moment that Mr Simmondssubmits it is all totally irrelevant because there is no suggestion that the members ever knew anything about what Mr MacDonald had said. How then, it is asked, can any expectations, reasonable or otherwise, have been engendered capable of forming the foundation of a claim based on breach of the Imperial duty? The answer to that is that it cannot have been a factor in engendering any Reasonable Expectation any more than the

At the conclusion of the Marriott Hotel meeting, Mr Lamb said that he would need to seek the approval of theTrustee but that he would recommend that such approval be given. That took place at the TMM on 19 January 2006 when Mr Heath presented the revised proposals. members would be asked to consent to the change, and any DB member not consenting would only receive a salary increase of two thirds of what they would otherwise be given. As later explained in a Q&A document prepared for members, decisions on an employee's salary increase would be made by his or her manager, who did not know what kind of pension the employee had, or what election he or she made. Once the manager had decided on the salary increase, one third of it, described as a "non pensionable salary supplement", would be taken away if the employee did not consent to the change in pensionability. As the RBs correctly observe, the award of a full salary increase was expressly unrelated to the performance of employees, but instead to their agreement to have their pensionable salary restricted.

I need to digress for a moment to consider IBM's attitude to a cash injection into the DB Plans which, it will beremembered, was in a state of significant deficit related to past service. CHQ was attracted by the idea of making a cash contribution:

As well as being attracted by the idea, consideration was given to the possibility from time to time:

So it would seem that internally, at least, CHQ was committed or resigned, depending on your point of view, tothe making of a significant contribution to the Plans by the time it had reached its agreement with the Trustee about the proposals to be presented to the membership. But the Trustee board knew nothing of that.

After that digression, I return to the TMM on 19 January 2006. The Trustee was told that, if it rejected theproposals, the offer of a lump sum payment would be withdrawn as it was only available that day. Mr Lamb reiterated the deep disappointment of the Trustee that the proposals had been brought forward "so soon after having reached an agreement with the Company in December 2004 which the Trustee had been led to believe was both affordable and would sustain the defined benefit plans." Mr Lamb said he undersood the accounting issues faced by IBM and the impact of welfare costs on those issues. He asked what action Holdings would take "to ensure that this position did not happen again". It is worth setting out Mr Hirst's response as recorded in the minutes:

The Trustee then agreed in principle to the proposals. But, as the minutes record, it would be wrong to concludethat anyone was happy with the outcome. A communication exercise with the members would follow.

The submissions from each side in relation to the Soto changes reflect very much the submissions made inrelation to the Ocean changes. Both Mr Newman and Mr Bridges, who was at the time of Project Soto an independent Trustee director, stated in their witness statements that no assurances were given by IBM at the time

Mr Hirst reported the Trustee's agreement to the proposals to Mr MacDonald on 20 January 2006. MrMacDonald's response to Mr Heath, Mr Wilson, Mr Hirst, and copying in Sam Palmisano, was "you did a terrific job. We teamed well too. The three of you made a huge difference to IBM UK and IBM too. It was a total pleasure working with you and your leadership was quite evident every moment." CHQ, at least, was apparently very pleased with the outcome.

In the meantime, beginning on 11 January 2006, IBM had consulted on the proposals with the so-called UKForum, an elected representative employee body used by IBM for the purpose of consulting on employee relations issues. The consultation continued through to April 2006. A total of eight meetings took place.

In preparation for the first meeting, Paul Simmons, UK Employee Relations Manager, emailed Declan Murphy(UK HR) with draft UK Forum Charts. He was clearly concerned with the issue of member trust as appears from his covering email: "we should also have a couple of charts addressing the issue of trust and whether this goes against what we communicated last year".

The presentation itself gave the same message as that given to the Trustee. The following are extracted from theslides:

In relation to the pensions actions taken in 2005, the presentation explained that the Soto proposals tackledvolatility and cost, as opposed to dealing with the past service deficit and increases in longevity as had been the principal focus of Ocean:

Another slide asked why the proposals were going to be any more successful. A number of factors would, IBMbelieved, "mitigate the current volatility" (namely transfers to the DC plans, a one-off cash injection further securing the Plans and a reduced, but guaranteed COLA commitment). The proposals formed part of a broader set of changes "developed by IBM Corporation to address pension costs and volatility across a number of countries". There one sees confirmation, again if it were needed, that it was CHQ which was driving the reforms and which was giving assurances. The proposals were seen by IBM as having an enduring effect:

The minutes record Mr Hirst as having been asked whether "IBM was clear about the goals so that we do nothave to return to this next year". He replied "that there was no guarantee, but that the company was doing everything possible to prevent that and further issues". Mr Tennet submits that this statement, taken with the other material which I have just referred to gives the clear impression to the representatives that Soto would tackle volatility and that IBM was doing everything it could to address that issue.

Mr Simmonds has a different take on this exchange. He submits that what Mr Hirst actually made clear was thatIBM was not precluding itself from making further changes even as early as the next year; after all, his statement that there was no guarantee was in response to a question about the need to "return to this next year". He goes further, saying this: not only does this mean that it did not preclude itself at the time of Project Soto from announcing the Project Waltz changes in 2009 but also it indicates that IBM did not consider itself to have precluded itself from making such changes at the time of Project Ocean. And he points out that there is nothing in the minutes to indicate that anyone raised the suggestion that IBM was in 2006 precluded from doing what it then proposed to do by reason of anything it had said or done in 2004 or 2005. In particular, nobody is recorded as relying on the guarantee (and what the members had been told about it) as an objection to the proposed Soto changes. Mr Simmonds goes as far as to submit that the very fact that the question was asked suggests that the questioner, for her part, did not consider that the Project Ocean precluded the proposals or further changes in the future.

Mr Hirst's answer was, of course, given in the context of the presentation as a whole, where it is interesting, Ithink, to note the use of the words "sustainable" and "volatility". In the context of this presentation and meeting, "sustainability" can surely only mean an ongoing scheme for future service; and that one problem in relation to sustainability, namely "volatility" had, been addressed (in spite of what Mr MacDonald says about what he thought the proposals were about). In saying that, I do not suggest that Mr Hirst's caveat about nothing being guaranteed has no content: a significant change in the financial and trading environment might lead to further changes having to be made.

Although Mr Hirst was making absolutely clear that IBM was not giving a guarantee, there was, to my mind, aclear representation about IBM's belief, namely that the Plans were thought to be sustainable and that volatility had been addressed. There was, at least, an implied representation that IBM considered that the Plans would not represent an unacceptable drain on IBM's resources in cash terms or give rise to unacceptable consequences for accountancy purposes. The message, it seems to me, is "We can afford to keep these Plans going" but no guarantee could be given because in reality all sorts of things might change. It was also implicit in the message that IBM was not, at the time, considering yet further changes in the future.

On 26 and 27 January 2006, the proposals were announced by Mr Hirst at three all-managers meetings in

The RBs' written opening submissions and their written closing submissions both drew attention to certainaspects of the presentation, and more was said by Mr Tennet in his oral closing submissions. Drawing the threads together, I summarise the submissions in the following paragraphs. I start with what was said in the meetings:

Mr Hirst spoke about the bargain which had been made the year before. The effect of what he said is not entirelyclear. I think it is helpful to include the whole of the relevant part of the transcript of the meeting in order to attempt to understand the message which he was conveying. It is important to appreciate that Mr Hirst started this part of his presentation by reference to press reports:

It is to be noted also that Mr Hirst stated that Mr Heath would talk about it (ie the bargain) later in the presentation so that, in understanding the message which was being conveyed by the presentation overall, it is necessary to see what Mr Heath said too. Let me record what he did say:

Then a little later, speaking to the slide which I have already mentioned at paragraph 846(vi) above - "ABalanced Package", Mr Heath said this:

In outlining the proposals, Mr Heath explained how the restriction in pensionable salary increases might affect aDB member over a 10-year period. Mr Tennet submits that this necessarily implied that members were entitled to expect that the way their pensionable salary increases counted towards their pension would remain unchanged for at least 10 years. I do not agree that it necessarily follows in the sense that a change would be inconsistent with the statement. This is particularly so given that the example was given to effect a comparison with what the position would be if no changes were made: the result would be that the member would receive something like 89% of his previous entitlement.

Mr Tennet submits that, from the explanation above, those present at the presentation could reasonably expect(and communicate onwards the message) that:

Mr Heath himself, in his oral evidence, told me that in his view, the presentation gave the impression that theproposals:

Mr Tennet perceives two different bargains being identified by Mr Hirst. The first, found at the beginning of thequotation above, was "increase your contributions" and then IBM "will guarantee the pension fund". The second was that the bargain involved merely an increase in contribution from members in exchange for IBM agreeing to fund the deficit over a 9-year period: it was in relation to that that he said the "bargain is still in place" or "is still on the table". Mr Heath referred principally to the funding guarantee, although what he said about review of contributions on each triennial review only makes sense on the hypothesis that benefit accrual is still taking place.

In the light of all this, Mr Tennet says that, reading the transcript and watching the video carefully today, onegets an entirely confusing picture of what message Mr Hirst was attempting to give and submits that it was doubtless at least as confusing for members. He concludes that one thing is obvious: this was no way to correct the misapprehension that IBM had given a legally binding guarantee that DB accrual would continue unchanged

He also prays in aid what Mr Heath agreed with in Court, namely that if members were labouring under such amisapprehension, this presentation would not have corrected them. In fact, that is not what Mr Heath agreed with: he agreed that the slides would not have corrected any misapprehension, but that was not an admission that what Mr Hirst said about the bargain from IBM's perspective was unclear, and in any case he did not accept that members were under any such misapprehension. He would regard what he himself said as entirely consistent with the absence of any sort of guarantee about future accrual or about the absence of any future changes to the Plans at least until 2014.

There is one further point to make arising out of the Q&As session following the presentation. One questionerasked about how many DB members IBM expected to move to the Enhanced M Plan. Mr Wilson told the gathering that "we constructed this package very carefully to be quite robust in terms of the financial dynamics and quite neutral to the number of people that actually transfer…so that we wouldn't be overly concerned about having to hit a particular number or not". From that answer, Mr Tennet submits that the listener could reasonably expect that IBM was content to allow DB members to continue with DB accrual on the post-Soto modified basis. That is almost certainly true. But it does not answer the question about how long a member could expect accrual to continue on that basis or in what circumstances it would be proper for IBM to change the accrual rate or even close the DB Plans to further accrual altogether.

There is also a move away in the Q&As from the language of "addressing" cost and volatility to "mitigating" or"reducing" or "seeking to address".

Mr Simmonds has a different take on this presentation. He recognizes that Mr Hirst addressed the perceptionwhich might have been held by some people that IBM was about to break the "bargain" or "deal" that had been reached in 2004. But he says it is clear, and I agree, that Mr Hirst was saying that the perception which he had identified was wrong and that he was suggesting that the bargain had not been broken. Mr Simmonds goes further and submits that it was clear from Mr Hirst's description of the bargain that it consisted, on IBM's part, of a commitment to eliminate the Plans' funding deficits – or, as Mr Hirst put it, to "fill the hole". Regardless of what impression members of the audience might previously have had, Mr Hirst explained that the statement (in the first paragraph quoted above) that "for nine years the IBM Company will guarantee the pension fund" meant (as is made clear in the second paragraph) that the IBM Company over a nine-year period would fill the hole that currently sits in the pension plan. Nothing in Mr Hirst's comments says anything about a commitment to maintain benefit levels until 2014. In this regard, Mr Hirst's comments must have served to clarify in the minds of the audience the true nature of the guarantee and disabused them of the impression (if they had held such an impression) that the maintenance of future benefit levels formed part of the Project Ocean bargain.

Similarly, Mr Simmonds says that Mr Heath's comments at the meeting would have cast further light on – andwould have served to make clear to the audience if they had been under any misunderstanding – the nature of the guarantee and the commitments into which IBM had (and, more importantly, had not) entered in 2004. His description of the guarantee – and, in particular, his observation that people might ask: "but wasn't one of the criteria in terms of setting up the guarantee over 9 years last year that we were in deficit and if the company ever did balance the plan that the company could then pull out of that guarantee?" – indicates that he considered that his audience would have understood a provision for early termination to have been one of the terms of the guarantee, showing that they would have had a fairly detailed understanding of the nature of the guarantee. It is said, in any event, that after listening to the passages of Mr Heath's presentation which I have set out, nobody in the audience could have been under any misconception as to the nature of the guarantee, ie that it was a funding guarantee and moreover that it did not amount to a commitment to maintain future benefit levels through to 2014.

Mr Simmonds makes the point that, contrary to what has been suggested on behalf of the RBs, it was perfectlyclear from Mr Heath's words that the reference in the slide to "protecting the pension fund" referred only to

It is of course the case that the membership as a whole did not have the advantage of the presentations from MrHirst and Mr Heath. But it was the role of managers (as acknowledged by Mr Johnson, a manager who did attend the presentation) to pass on faithfully the information that they had received at this briefing to their employees. Mr Simmonds says that it is to be expected that, when explaining the proposals to the workforce at large, the attendees at the all-managers meetings would have done so on the basis of an accurate understanding as to the nature and effect of the guarantee and, in particular, would have said nothing to suggest that the guarantee amounted to a commitment to maintain the existing benefit structure until 2014. I consider that that is putting matters too high. What managers ought to have explained to the workforce was precisely that which had been presented to them at the presentations. To the extent that there was ambiguity in what was presented, managers could not be blamed if they then failed to convey the message which Mr Simmonds now submits they should have conveyed.

It has been important for me to address the presentation and the submissions in relation to it in some detailbecause it is heavily relied on by both sides. But, in my judgement, it is actually of very little assistance. Of itself, I do not consider that the presentation makes clear one way or the other what IBM's commitment to future benefit accrual was: subject to two related caveats, I agree with Mr Heath who accepted, in his oral evidence, that for a person who mistakenly, as he would say, believed as a result of the Ocean communications that IBM was guaranteeing further benefit accrual to 2014 or some other date, this presentation would have done nothing to dispel that misunderstanding. So one is thrown back to the Ocean communications to discover what it is that the reasonable manager attending this presentation would have understood.

The first caveat is this. There is some support for IBM's position to be derived from the reaction (or rather lackof reaction) from those attending the presentation. As with the Webcast, the extended guarantee and the alleged commitment to continue to allow benefit accrual are separate matters. Even if it is correct that the guarantee does not of itself have anything to say about accrual, that does not mean that the presentation given by Mr Hirst and Mr Heath taken as a whole had nothing to say about accrual. In fact, it did say something, namely that, for those making the relevant election, only part of future salary increase would be pensionable. Although that is not inconsistent with a commitment to future accrual per se, it is inconsistent with accrual at the then current rates. Yet no-one at the presentation to managers appears to have suggested that the proposal was inconsistent with what IBM had, as is now alleged by the RBs, committed to such a short time before. If the message given by the Ocean proposals had been so strong – giving rise to the Reasonable Expectations alleged by the RBs and forming the central plank of their claim in these proceedings – it is remarkable that the presentation to managers provoked not a single question on the topic.

The second caveat is this, and lends some support to the RBs' position at least to some extent. IBM wasexpecting, indeed requiring, members to make elections about whether to transfer to the DC Plan on the basis of the material contained in the presentation (to be communicated by managers) and the subsequent

The proposals were announced to members and employees generally by Mr Hirst and Mr Lamb respectively on27 and 30 January 2006. Mr Hirst's email making his announcement contains the following:

A number of points need to be made about this email.

Mr Hirst stated clearly that IBM was not proposing to close any of the DB Plans. I accept, as Mr Simmondssubmits, that this was a response to rumours (to which Mr Hirst had alluded in his presentation to managers) circulating amongst the workforce and the press following the announcement earlier in the month of the closure of the DB plan in the United States. I also accept that the statement was entirely correct as a matter of fact: when Mr Hirst sent his e-mail, Holdings was not proposing to close the DB sections of the Plans and such closure formed no part of the Project Soto proposals. Moreover, closure was not even on the discussion agenda within the UK management team, although what CHQ were thinking about at this time is more open to question. The statement cannot, in my view, and here I agree with Mr Simmonds, on any reasonable reading of the e-mail be construed as a promise by Mr Hirst that IBM would never close the Plans to DB accrual or even as a statement that IBM's current intention was never to end DB accrual.

Not all of the RBs' witnesses accepted the proposition, when put to them in cross-examination, that Mr Hirst wasnot promising that the Plans would not close. For my part, I do not see how a statement that there is no proposal to close the Plans can be taken as a promise never to do so, or even in context not to do so before 2014. The reasonable reader of the email would not draw that conclusion although he certainly would draw the conclusion that IBM had no undisclosed agenda to make further changes in the future.

As can be seen from the first paragraph of his email, Mr Hirst explained that he was outlining the proposals being put forward by IBM. It was obviously not a detailed exposition of the changes. It was recognised that members would be faced with difficult choices. This was why members were to be assisted in making their choices; and so Mr Hirst informed members as follows:

The reasonable reader would not have thought that the whole picture could be seen from the email; and mostcertainty the reasonable member would not have made his choices based only on what he was told by Mr Hirst in

At least one of the RBs' witnesses said she did not read the Q&As. Mr Tennet submits that I should be verycautious about placing any substantial weight on them principally because it is very difficult to know which members read which Q&As (if at all) and when. That is not the correct approach. In addressing whether the communications gave rise to Reasonable Expectations, it is the conduct of the reasonable employee/member which is relevant and what individual members actually did or did not do is not to the point.

I consider that Mr Simmonds is correct when he submits that any reasonable employee would have consideredall of the further information and would have taken advantage of at least some, if not all, of these further resources. A reasonable reader may not have read every document in detail; but he would certainly have looked at them and focused on questions of particular personal interest. If pensions are to be seen as a very important aspect of the employees' terms and conditions of employment (which they are), it is not sensible to think that the reasonable would be satisfied by this communication that there would be no change in the future.

In any event, there are certain key documents: the presentation, Mr Hirst's announcement and the announcementto members in a letter from Mr Lamb to which I now turn.

Mr Lamb wrote to members in the week of 30 January 2006, explaining that the Trustee had agreed in principleto the changes IBM was intending to make. Near the beginning of the letter he explained, in particular that:

There then follows text under a number of headings – Changes for Employees, Changes for Retirees, Changesfor Deferred Members, Funding, Conclusion.

Under the "Funding" heading, Mr Lamb wrote:

That reference to the support agreement can only be to the guarantee in the Funding Agreement. It providesfurther confirmation that the guarantee is to do only with funding. This paragraph of the letter gives no support at all to a suggestion that IBM was guaranteeing future accrual. Indeed, it lends support to IBM's case. This is because the earlier part of the letter, in accurately explaining the proposals, discloses that only base salary increase will be pensionable, and that the so-called salary supplement will not be. That is clearly something different from the then current position. If it is said that there was a guarantee of future accrual under Ocean (ie at the then current accrual rate), then that guarantee is clearly not being continued under the Soto proposals. But the paragraph quoted above records that "the support agreement", that is to say, the Funding Agreement including the guarantee, will remain in place. This results in an inconsistency: the guarantee which the RBs assert in relation to Ocean (and which they say was to continue until 2014) related to future accrual at the then current rate; but the ongoing guarantee clearly does not do so and could at most relate to a guarantee at the reduced accrual rate with only part of salary increases being pensionable. And yet Mr Lamb's letter stated that the support agreement "will remain in place".

That is, I appreciate, at a detailed level quite a sophisticated point and the reasonable reader of the email wouldnot have been likely to appreciate it or have been able to articulate it quite as I have done. But the reasonable reader would have appreciated that in the future not all of his salary would be pensionable and that would, I suggest, have raised a question in his mind about whether IBM could properly make the change in the light of what, on the RBs' submission, would have been appreciated as the effect of the Ocean commitment – namely

It is also to be noted that, under the heading "Changes for Retirees", pension increases are dealt with. For thefirst 5 years, increases were to be awarded at a specified rate; for the next 10 years, increases were to be at a lower specified rate. The period to 2014 includes, of course, the first 5 years and also 3 of the next 10 years. Mr Lamb wrote: "If the plan is wound up the pension in payment guarantee will fall way". That, it seems to me, is a possibility which is entirely inconsistent with the plan guaranteeing future accrual at the then current rate or at all and is not easy to reconcile with a long term commitment to keep the Plans open. The reasonable active member reading the letter would not focus only on the material under the heading "Changes for Employees" but, as a future pensioner, would have the material under the heading "Changes for Retirees". In any case, there is a limit to the selective reading of this letter for which the RBs could contend.

The section of Mr Lamb's letter under the heading "Conclusions" should be set out in full:

What is clear from that is that there was to be no guarantee going forward of further accrual of benefits at thepreviously current rate. It is also clear that there was to be no guarantee going forward that there would be no further changes (or indeed a winding-up of the Plans). As to the Ocean commitment, it was recognised by Mr Lamb that the Trustee had not been able secure the position where the members' "pension benefits remain at least in line with their expectations" but that the best deal possible in the circumstances had been obtained. He did not suggest that IBM was acting in breach of any agreement which it had made or even of a commitment falling short of a promise or guarantee. I agree to some extent with Mr Simmonds when he says that any reasonable member reading this letter who was under the impression that IBM was committed to maintaining the benefit structure would have been caused, at the very least, to question that impression in view of the fact that the Trustee was accepting (albeit reluctantly) a change to the benefit structure: I say "to some extent" because, in my view, it is only in relation to future benefit accrual that he is right. Nothing was said one way or the other about other aspects such as ER policy.

Further, Mr Lamb recognised that IBM had made it clear that it was unwilling to give even a commitment that there would be no further changes. It did, however, view the changes as "long term" and had "no plans for further change".

This was a statement of Holdings' view. But there can be no doubt that some, at least, of those in Armonk knewthat this view was being expressed. CHQ must be taken as approving of, and adopting, that view. Or if that is not so, it may nonetheless be highly relevant when considering a breach of its Imperial duties by Holdings.

Apart from that, the same points arise in relation to these statements as I have already discussed in relation tosimilar statements in the context of Project Ocean, but subject to the points made in the preceding few paragraphs.

The reference to "long term" and "no plans for future change" are however stronger than the statements made inrelation to Ocean. A reasonable member reading those words in isolation would, in my view, be entitled to form a Reasonable Expectation that, in the long term (whatever that may mean) there would be no changes absent a significant change in economic and financial circumstances.

At this stage, and before coming on to further communication with the membership, I wish to say a bit moreabout what was happening between IBM Corporation and CHQ on the one hand, and those in the UK – the UK pensions team and the Trustee – on the other. I have dealt with some matters in the course of the narrative so far, but there are two particular matters on which I want to say a bit more.

Since Mr Lamb's letter is most recently in the mind of the reader, I record that drafts of that letter had been thesubject of comment from and discussion with CHQ as well as other members of the Trustee board. Mr Lamb had sent a first draft to the Trustee directors on 25 January 2006, inviting comment. It prompted a number of responses. I do not propose to go into the detail of the drafting changes and the different attitude adopted by the conflicted and the non-conflicted directors, with the conflicted directors taking what, on one view, was a particularly company-friendly approach. The fact is that the board agreed the final draft. I should, however, mention the response made by Mr Lamb to a suggestion by Mr Heath which would have resulted in the omission of the words "long term" to describe the changes. Mr Lamb wrote this to Mr Heath:

Mr Heath did not respond, even to confirm the long-term nature of the changes, nor did anyone from IBM. Thewords remained in the final letter. Although members would not have seen this letter, it informs what the Trustee and Holdings meant by use of the word; and it goes to emphasise the importance to the Trustee's decision of the fact that there were no plans to alter the DB Plans in the future. Mr Lamb clearly did not distinguish between Holdings and IBM Corporation/CHQ in this thinking not least because the message was coming from Mr MacDonald. Mr Hirst cannot have made such a distinction either.

The other matter I want to mention is the suggestion made by the RBs that pressure was placed on the UK teamby CHQ when driving through the Soto changes. There are three aspects: personal pressure on senior UK executives, a threat to withhold SIP from the UK if it did not make the changes required and a confrontational stance adopted by CHQ. As to the first of those there is nothing wrong with putting pressure on people providing that it is not improper pressure. As Mr MacDonald said:

It would not be right, as I think Mr Tennet hints would be right, to take from Mr MacDonald's answer anacceptance that improper pressure had been brought to bear. But whatever Mr MacDonald did or did not accept, the conclusion that there was personal pressure on senior executives is said by Mr Tennet to flow from the following:

With the possible exception of the interchange with Mr Hirst, there is insufficient in this evidence (and nothingmore came out of the oral evidence) to show that CHQ placed any improper pressure on the UK team. I must be cautious in placing reliance on Mr Hirst's evidence since (i) he could not be cross-examined and (ii) he seems to be suggesting greater, and possibly improper, threats to Mr Wilson and Mr Heath when their own evidence does not establish that. In any case, I see no reason to think that the actual decision of the UK team was in any way influenced by considerations of the personal position of any of the individuals concerned. Rather, they agreed to the Soto proposals because they were persuaded that the alternative of rejection would be even worse for employees and members.

As to the threat to withhold SIP, Mr Tennet relies on the following: