BT has lost its attempt to convince the Court of Appeal that it should be able to move away from the retail prices index (RPI) for indexation of benefits for one of the sections of its pension scheme. The Court of Appeal upheld the High Court judgment that RPI had not become an inappropriate measure of inflation for indexation of pension benefits due under the BT Pension Scheme.

What did the Court of Appeal hold?

1. BT's appeal is dismissed unanimously and the High Court's decision stands

Asplin LJ said that the High Court judge "was right to come to the conclusions he did" and was entitled to conclude that the RPI has not become inappropriate for the purposes of calculating increases in pensions in payment under the BT Pension Scheme.

2. The question of whether RPI had become inappropriate was not for BT or the Trustee to determine

Asplin LJ said this wasn't a subjective question: "whether RPI has become inappropriate is an objective state of affairs which, if it exists, triggers the obligation to choose another measure of the cost of living."

3. Application of a textual approach following Lord Hodge in Barnardo's v Buckinghamshire and others

Whilst it might have been convenient or better for BT to have a role in determining the appropriateness of RPI, the rule in the BT Pension Scheme did not explicitly provide for this. The 2003 Rules (where BT did have a role) were not relevant in interpreting the 2016 Rules.

4. Evidence from the House of Lords was admitted to Court

Issues regarding parliamentary privilege were considered. However, these did not prevent evidence which post-dated the High Court trial being admitted before the Court of Appeal.

What was the case about?

The proceedings concerned the appropriate cost of living index to be used in calculating increases to pensions in payment under Section C of the BT Pension Scheme (affecting around 83,000 members). Pension increases for these members are currently calculated and paid on an RPI basis. A change in the rate indexation for Section C member could reduce liabilities in the scheme by c. £1.4 billion (with an average impact of £15,000 per member, according to Unite). Sections A and B have already moved to consumer prices index (CPI).

BT, the sponsor of the pension scheme, sought directions from the Court as to whether the CPI (which generally produces a lower measure of inflation than RPI) could be applied in place of RPI. Arguments were put forward by BT ,acting in its own capacity and representing all members whose interest it was that increases should no longer be linked to RPI, and by a Representative Beneficiary acting on behalf of those members whose interest it was that increases should remain linked to RPI. The Trustee adopted a neutral role, though it did make some submissions.

What were the rules to be construed?

There were various iterations of the rules before the Court, of which the Rules from 1993 and 2016 were most relevant (the 1993 Rules and 2016 Rules respectively).

The 1993 Rules prescribed that pension increases would be calculated by reference to the General Index, defined as the General Index of Retail Prices, save that:

"[i]f the General Index ceases to be published, or is so amended as to invalidate it in the view of the Principal Company as a continuous basis for purposes of calculating increases, the Principal Company shall substitute such other index or appropriate basis of comparison as it shall in consultation with the Trustees decide." (Rule 25(3))

The 2016 Rules contained a trigger mechanism, which when activated would mean that an alternative inflation measure could be adopted in place of the RPI:

"[t]he cost of living will be measured by the Government's published General (All Items) Index of Retail Prices or if this ceases to be published or becomes inappropriate, such other measure as the Principal Company, in consultation with the Trustees, decides". (Rule 10.2)

The 2016 Rules did not state who should determine whether RPI had become inappropriate. However, it was common ground that, once the trigger had been activated, BT and the Trustee would need to decide on any alternative inflation measure to be adopted in place of RPI.

What did the High Court decide?

In the High Court judgment, Zacaroli J held that the relevant trigger event for a move away from RPI had not occurred.

He held that, in deciding whether RPI had become inappropriate for the purposes of the BT Pension Scheme, it was necessary to apply an objective test. It was not for BT to decide whether it had become inappropriate. Even though the 1993 Rules gave BT a role in determining whether the trigger had occurred, that was irrelevant in construing the 2016 Rules. In any event, until such time as RPI had been substantively amended, the 1993 Rules would not have been engaged.

Significant expert evidence was adduced by experts in the field of inflation measurement; Paul Johnson of the Institute of Fiscal Studies and Simon Briscoe of the Royal Statistical Society (the Society). Zacaroli J had found this to be of limited assistance to the question of the appropriateness of RPI for this particular scheme. The Judge considered various events in the history of RPI (e.g. the clothing change in 2010, the freezing of RPI in 2013 and the de-designation of RPI as a national statistic) but concluded that none of these, individually or cumulatively, meant that the index had become inappropriate. In doing so, he rejected the contention that because there was a new definitive and replacing deed in 2016, that 'reset the clock' and restricted the Court to considering events after that date in deciding whether RPI had become inappropriate.

The Judge accepted that, in view of these events, it may be inappropriate to use RPI if one were founding a new pension scheme today. However, that did not mean that RPI was inappropriate for the purposes of the BT Pension Scheme, with its long history. He also noted that many of the alleged flaws in RPI were known in 2002 when the 2016 Rules were introduced.

A letter from the Society was put before the Court which, it was said, indicated a change in position on the part of the Society as to the future role of RPI. Zacaroli J was not convinced that there had been such a change [190] of his judgment.

BT appealed the judgment to the Court of Appeal on the basis that (amongst other things) Zacaroli J erred on his construction of the 2016 Rules ;on his approach to the evidence; and that he had reached incorrect conclusions as to the position of the Society and the Office for National Statistics on the appropriateness of RPI. There was also a cross-appeal brought by the Representative Beneficiary, which included whether any power to move away from RPI would lapse if not exercised within a reasonable time.

What did the Court of Appeal say?

Asplin LJ gave the leading judgment, with which Patten LJ and Richards LJ agreed, dismissing BT's appeal.

Applying Barnardo's textual analysis

Asplin LJ referred to the recent Supreme Court judgment in Barnardo's (another 'RPI/CPI case') which addressed the issue of the construction of pension scheme documents. With that in mind, Asplin LJ noted that she would "give due weight to textual analysis by concentrating on the words that the draftsman chose to use" [19].

Asplin LJ agreed with Zacaroli J that simply because it might be convenient for BT to be the party to determine the "appropriateness" of RPI (indeed, the 1993 Rules had given BT such a role) did not negate the fact that Rule 10.2 did not give BT (or the Trustee) the power to decide whether RPI had become inappropriate for the purposes of the BT Pension Scheme. Asplin LJ also noted (obiter) that, despite it being common ground between the parties that upon the occurrence of a trigger event BT and the Trustee then held a joint power to move away from RPI, she interpreted the rule as being a "substitution obligation" ([30]). This dispensed with the Representative Beneficiary's cross-appeal on the lapsing of an unexercised power ([121]).

Applying Barnado's, the Court of Appeal rejected an argument that the 1993 Rules was an aid to construing the 2016 Rules, noting that "pension scheme archaeology is unlikely to be of much assistance" ([32]). Asplin LJ also noted Lord Hodge's comments that background facts and circumstances which existed at the time the BT Pension Scheme was set up wouldn't be available to members reading the rules later and so should not be considered an aid to construction.

High Court judge did not err in principle on the evidence

Asplin LJ found that Zacaroli J was correct to find that pre-2016 events in the history of RPI could be taken into account in deciding whether the index had become inappropriate.

Further, she could find no error of principle in the Judge's approach to the evidence before him, which had led him to conclude that the events in the history of RPI (singly or cumulatively) had not caused it to become inappropriate for the purposes of the BT Pension Scheme. It was not the role of the Court of Appeal to conduct its own review.

Fresh evidence and the issue of Parliamentary privilege

An application was brought by the Representative Beneficiary to admit evidence given to the House of Lords Economic Affairs Committee as part of the inquiry into RPI.

Asplin LJ concluded that, as this evidence post-dated the High Court hearing; the evidence would have had an important influence on the outcome of the judgment; and no question arose as to whether the information was credible, the evidence should be admitted ([90]). There was no reason that the Court could not refer to evidence of proceedings of Parliament when they are simply "relevant historical facts or events", so parliamentary privilege was not an issue. ([87]). Asplin LJ concluded that the evidence supported Zacaroli J's conclusion that the Society, in summary, still considered that RPI should be maintained and managed rather than shelved or discontinued.

What does this mean for other schemes?

To a certain extent, this case was no different from other 'RPI/CPI' pension cases the courts have considered up until now, which have turned on the particular wording of the relevant rule. However, there are a few aspects which do make this case of interest.

This case is the first to apply the textual analysis principles set out by Lord Hodge in the Supreme Court judgment in Barnardo's.

Although it is unlikely that the outcome would have been any different had the case been decided before Barnardo's, in her judgment Asplin LJ did make specific reference to the principle that members reading a scheme document years after a deed was executed would have no knowledge of the facts and matters available to the parties to the deed, so such facts and matters should not be an aide to construction. Time will tell how the continued application of this principle will affect construction cases in the future.

The use of the words "becomes inappropriate" in the Section C Rules meant that the Court of Appeal (and the High Court before it) had to look closer at the RPI itself and whether it had become inappropriate. As with the Warren J's judgment in Thales UK Ltd v Thales Pension Trustees Ltd and others [2017], significant expert evidence in relation to RPI was considered, and the views of the Office for National Statistics and the Society were considered in detail. However, the Court of Appeal endorsed the High Court's approach to that evidence and did not make a fresh determination that RPI was not inappropriate. That being the case, it would be an overstatement to consider this case an endorsement by the Court of Appeal to the continued use of RPI for indexation in all schemes.

The expert field of inflation measurement is moving rapidly. The continued prevalence of RPI in various fields has led to enhanced scrutiny of it as a suitable inflation index and this case will only have increased that scrutiny.