Associated Institutions Pension Fund v Le Roux and Others (196/2000) [2001] ZASCA 70; [2002] 1 All SA 261 (A); 2001 (4) SA 262 (SCA) (30 May 2001)

IN THE SUPREME COURT OF APPEAL

OF SOUTH AFRICA

REPORTABLE

CASE NO: 196/2000

In the matter between:THE ASSOCIATED INSTITUTIONS PENSION FUND AppellantandLE ROUX, PETRUS ABRAHAM KRIEL First RespondentBROOKS, PIERRE ELDRID JOSEPH Second RespondentAND 2510 OTHER RESPONDENTS

BEFORE: Vivier, Harms, Zulman, Cameron and Mthiyane JJA

HEARD: 17 May 2001

DELIVERED: 30 May 2001

The transfer regulations promulgated on 22 April 1994 under the Associated Institutions Pension Fund Act 41 of 1963 contemplated the employment of actuarial assumptions in the calculation of pension fund members’ transfer values and the actuary’s determination can therefore not be set aside

JUDGMENT

CAMERON JA:

At stake in this appeal is the magnitude of the pension benefits accorded some 2 500 employees and pensioners of the University of South Africa when in 1994-5, along with about 35 000 others from sixty five government-funded institutions, they elected to leave the central pension fund (“the Fund”) created under the Associated Institutions Pension Fund Act [1] and its regulations (“the general regulations”), [2] and joined autonomous funds established by their own institutions. In terms of regulations promulgated in April 1994 (“the transfer regulations”) [3] each departing member and pensioner was entitled to be credited with an amount “equal to the funding percentage multiplied by the actuarial obligation of the Fund in respect of that member as determined by the actuary [4] on the date on which his membership of the fund is terminated ”. [5] The reason for requiring that a funding percentage be determined was that the Fund, since 1985, had been consistently under-funded, with the result that departing members could not be paid 100% of their pension entitlement. Because the valuation of the Fund fluctuated from time to time, it was in addition necessary to specify that the determination in question be made on a fixed date.

In the court below Southwood J set aside the actuary’s determination of the Fund’s funding percentage and its resulting actuarial obligation to the applicants (respondents on appeal), and granted attendant relief. With his leave the Fund (first respondent in the court below) and the Minister of Finance (second respondent) appeal against that order.

The crux of the dispute is the funding percentage the actuary appointed under the transfer regulations, Mr de Wit, determined for the Fund, since from that figure he calculated the amount transferred on behalf of the applicants to the new University of South Africa pension fund. In April 1995 de Wit determined the funding percentage as at 30 November 1994 (the date agreed for the applicants’ departure from the Fund) at 60,8%, resulting in a transfer to their new scheme of some R459 million. Subsequent documentation showed that, in view of admitted difficulties in ascertaining the exact number of Fund members as at 30 November 1994, de Wit in calculating its aggregate actuarial obligation applied a “data loading factor” of 7,5% to its membership. This entailed increased provision for possible unascertained members, and decreased the Fund’s actuarially calculated value. This resulted in an appreciable reduction of the sum transferred to the benefit of the applicants. The reason for the inaccurate membership data was that the associated institutions were not obliged to render accurate membership returns to the Fund, and obtaining such data had proved intractably difficult for the Fund.

In his subsequent valuation of the Fund as at 30 September 1994 (issued in January 1996), de Wit however reduced this data loading factor in respect of unascertained members by two-thirds to 2,5%. That valuation yielded a funding percentage of 66% — nearly one-tenth higher than the valuation applied at the applicants’ transfer date. A later valuation as at 31 March 1995 (released in July 1996), when the great bulk of the Fund’s members had departed, yielded (on a similar data loading factor of 2,5%) an even higher funding percentage of 84,3% (albeit on a “going-concern” basis of valuation, as opposed to a discontinuance basis, which was used for the transfer value calculations).

On the basis of these figures the applicants mounted a stringent attack on de Wit’s calculations, which, they asserted, had resulted in the transfer of a substantially smaller amount than their entitlement. Southwood J upheld their contentions. Regarding de Wit’s approach to the inaccurate membership figures as at 30 November 1994 and his consequent application of a 7,5% data loading factor, Southwood J concluded that determining the Fund’s calculated aggregate actuarial obligation required a mathematical computation, based on reliable data. De Wit’s approach — that he was faced with avowedly unreliable membership data, but that the regulations authorised the use of estimates and assumptions in regard to such imponderables — was therefore incorrect: he should have waited until accurate membership figures eventually became available (making in the interim a provisional payment to the new fund). His determinations therefore had to be set aside.

Southwood J also found that de Wit had failed to include in his valuation certain special governmental capital contributions to the Fund. But Mr Wallis, who appeared for the applicants at the appeal, on an analysis of the Fund’s depositions in my view correctly disavowed reliance on this. He also abandoned the contention, advanced in the applicants’ heads of argument on appeal, that the actuary had not been entitled to base his calculation on the figures derived from the 1991 valuation of the Fund (updated according to de Wit’s evidence with the latest available information), and confined his argument instead to de Wit’s application of the 7,5% data loading factor to the unreliable membership data. This Mr Wallis argued conflicted with his powers under the transfer regulations.

Although the transfer regulations were promulgated before the interim Constitution [6] came into effect on 27 April 1994, it is clear that in regard to their interpretation and application the applicants were entitled to administrative justice under the Fundamental Rights Chapter of that Constitution. The regulations must therefore be interpreted through the prism of the interim Constitution, [7] and de Wit’s determination of the funding percentage had to be lawful and procedurally fair as well as justifiable in relation to the reasons he gave for it. [8] It is also clear (though the applicants’ attack, and the basis on which they succeeded in the court below, was that de Wit had acted on an improper understanding of his powers under the transfer regulations) that a determination infringing any of the applicants’ other fundamental rights could also have been impugned as in conflict with the interim Constitution.

Against this background the critical question is the proper interpretation of the transfer regulations, and the nub of the appeal is whether so interpreted the regulations entitled de Wit, given the unsatisfactory membership figures, to apply the data loading factor of 7,5% so as to reach the funding percentage of 60,8% applied to the applicants. The pivotal concepts are “funding percentage” and “actuarial obligation”. The regulations define “funding percentage” as “the market value of the net assets of the Fund on a fixed date expressed as a percentage of the calculated aggregate actuarial obligation of the Fund on that date, as determined by the actuary”. “Actuarial obligation” is defined, “with regard to a particular member, pensioner or dormant member of the Fund”, as “the actuarial obligation of the Fund with regard to that member, pensioner or dormant member on a fixed date, calculated by the actuary”.

The transfer regulations themselves in my view contain significant textual pointers to their proper interpretation. The most striking feature of the definitions [9] is their insistent allusion to the actuarial function. The definition of “actuarial obligation” is, apart from the reference to a fixed date, a meaningless repetition unless the words “calculated by the actuary” are acknowledged to be significant. The definition of “actuary” (“means the actuary appointed to evaluate the Fund actuarially as contemplated in regulation 24A of the Regulations”) likewise contains surplusage (“actuary ... actuarially”) unless the adverb is held to contribute to their sense. [10] The definition of “funding percentage”, similarly, in which a cross-allusion to “actuary” and “actuarial obligation” is already embedded, raises the pitch by adding “as determined by the actuary”. The executive provision, Reg 2(4)(a) (quoted in the opening paragraph of this judgment), which is similarly replete with already defined terms (“funding percentage”; “actuarial obligation”; “actuary”), itself for good measure adds “as determined by the actuary”. If the cross-allusions in that provision are disaggregated, as Mr Solomon for the Fund and the Minister correctly pointed out, the words “actuary”, “actuarial” and “actuarially” obtrude repeatedly and insistently.

Given this linguistic accumulation, the phrase “as determined by the actuary” can hardly have been intended, as Mr Wallis suggested, only to identify the actuary in whom the regulations vest the power to perform the calculations they enjoin. That the instrument attains with economy and clarity by a separate definition of “actuary”. The repetition in my view points not only to functionary, but to function, and it must have been intended to imbue the latter with attributes of professionalism and skill peculiar to the field of expertise they name.

There can in short be no doubt that invocation of the actuarial function was fundamental to a proper understanding and application of the regulations, and that they contemplated, authorised and required the employment of actuarial expertise and skill in the calculation of the transfer values applicable to the applicants. I agree with Southwood J that this entailed calculations “in accordance with the principles of actuarial theory and practice”. What is significant, however, is that the only evidence before the court of the methodology applicable to the actuarial function was that of de Wit himself, supported by the depositions of Professor Asher, incumbent of the chair of actuarial sciences at the University of the Witwatersrand (who was closely involved in the work leading up to the “emancipation” of the associated institutions’ pension funds), and Mr Milburn-Pyle, an actuary employed in a managing capacity by the firm in which de Wit at material times was an executive director (which in the court below as third respondent opposed the relief the applicants sought).

In his answering affidavit de Wit testified, and the applicants in reply admitted, that the allusion to the “market value” of the Fund’s assets necessarily entailed that the valuation for 30 November 1994 be performed on a “discontinuance” rather than “going-concern” basis, that is, to determine not the long-term financial soundness of the Fund, but the actual benefits imminently payable over the short term.

What an actuarial valuation of a pension fund’s funding level entails de Wit described thus: “The determination of a pension fund’s funding level is not an exact exercise. It can only be described as the actuary’s best estimate of the funding level at the time. It is also correct to conclude that two actuaries will seldom determine the same funding level for a particular fund at a particular date. There are simply too many imponderables and discretionary matters involved in such an assessment. The actuary must do the best he can with the information available to him at the time and apply whatever provisions are necessary in the circumstances.”

And:

“There is no such thing as one ‘correct’ funding level. The applicants’ contentions in this regard are simplistic and demonstrate a lack of understanding of the function which the transfer regulations require the actuary to perform. The applicants appear to labour under the misapprehension that the funding level of a retirement fund is capable of a simple mathematical calculation based upon known facts. It is not such a calculation.”

The applicants in reply took issue with de Wit but, significantly, they treated these passages as his factual justification for his approach to the determination. Their deponent asserted that “the ex post facto determination of the funding percentage on a discontinuance basis is, by definition, an exact exercise” — but tendered no evidence (definitional or otherwise) to support this averment. They also claimed that “no determination of a funding percentage, and especially not one made in terms of the transfer regulations, can be legally acceptable if its factual basis, in the form of the data concerned, is clearly deficient”. That restates their fundamental complaint, but it does not meet the bite of de Wit’s evidence, which entailed more than a factual averment. It contained an exposition of the professional methodology the transfer regulations contemplated for the performance of the statutory duty they created, and to that the applicants had no answer, since they put forward no expert evidence of their own and their principal deponent, a practising attorney, rightly professed no expertise in the field.

With great respect to the care and thought that inform the reasons of Southwood J, I am unable to agree with the meaning and weight he assigned to “calculate” in the definitions, and Mr Wallis for the applicants did not attempt to support that meaning. In any event, de Wit’s critical evidence in this regard runs counter to Southwood J’s finding that the statutory duty entrusted to the actuary could be performed with mathematical precision, bereft of assumptions, allowances or margins in regard to uncertain facts and figures.

During argument Mr Wallis shifted the focus of the applicants’ attack from a complaint that de Wit botched his brief by using inaccurate figures to arrive, wrongly, at the 7,5% loading, to the proposition that, on de Wit’s own evidence, the discontinuance basis did not permit any data loading to be applied at all. This argument cannot in my view be sustained. On a true construction the transfer regulations required the invocation and application of actuarial expertise and that, on the uncontested evidence before the court as to the professional methodology involved, necessarily entailed that assumptions would be made to allow for contingencies and imponderables. That is the nature of the actuary’s job, and it was a job the regulations required de Wit to perform. In a different context, but one not inapposite to the present, Marais JA pointed out in Tek Corporation Provident Fund and Others v Lorentz [11] “In assessing the financial health of a pension fund an actuary is gazing into the proverbial crystal ball to see what the future will hold. The use of the metaphor is not intended to demean the exercise; it is highly sophisticated and requires considerable training and skill, yet it remains, when all is said and done, an exercise in prophecy. Some of the data available may be relatively immutable and provide a secure foundation for predictions. Much of it is not. There are a host of factors about which assumptions have to be made because they lie in the future. Examples are rates of return upon different categories of investment, the rate of inflation, governmental fiscal policy, increases in salary, mortality rates for active and retired members, the rate of employee turnover, the incidence of disability and the extent to which early retirement options may be exercised. The list is not exhaustive but it suffices to show the very considerable role that assumption plays in the assessment of the financial soundness of a pension fund and explains why even the most meticulously assessed valuation may be confounded by subsequent experience.”

De Wit was not, of course, gazing into the future, but attempting to establish the present. He did so on the basis of avowedly inaccurate membership data, but having regard to this fact, as he had to, he made an adjustment to allow for possible contingencies. These were not so much unforeseen as unknown. At the time he made his determination they were nevertheless ineluctable realities and in my view he rightly took them into account in performing his statutory duty.

Mr Wallis contended that the Fund’s aggregate actuarial obligations could not be calculated by assuming obligations that did not in fact exist, and that the regulations did not permit de Wit, for prudential reasons, to make contingency allowances for what Mr Wallis called “potentially non-existent obligations”. This may be seen to reveal the weakness at the core of the applicants’ argument, since it was precisely the potentiality in the situation that de Wit was obliged to take into account, and in the circumstances he faced he could do so only by making provision for all contingencies that might reasonably affect his calculation.

His duty in this regard included assessing the Fund’s total membership on the information available to him at the time. He owed this duty as much to those who chose to stay in the Fund as to those who chose to go; and the fact that the result proved in the longer run to the advantage of those who stayed and to the disadvantage of those who left cannot invalidate his assumptions at the time they were made. The applicants attacked de Wit for adopting an unjustifiably “conservative” approach to the determination of their entitlement. De Wit denied that his approach was conservative in this sense, but admitted that in determining the applicants’ funding percentage, because of the inadequate membership data, he adopted a “cautious and professionally prudent approach” to the Fund’s liabilities. Later-acquired wisdom showed that a higher percentage, calculated with perhaps less prudence and less caution, would have matched the facts as subsequently revealed. This does not mean that he erred. By the methodology appropriate to what the regulations required of him, de Wit acted properly and lawfully at the time he made his determination. There is no suggestion that the assumptions he employed were inappropriate or unreasonable. The applicants’ case as developed by Mr Wallis was that the regulations permitted him to make no assumptions at all; and for the reasons I have given this contention does not withstand scrutiny.

The transfer regulations do not specify when the actuarial determination must be made. Though there was much debate about when the actuary was permitted or required to act, the starting point must self-evidently be that he was required to perform his statutory duty within a reasonable time. De Wit’s affidavit convincingly itemised the circumstances that impelled him to act in April 1995 rather than waiting for another eighteen months — or longer — before more accurate membership figures might have become available. (It was not in fact clear when those figures became available, if they ever did.) Those circumstances cannot, as Southwood J pointed out, dictate the proper construction of the transfer regulations. [12] But if their true construction did not require de Wit to wait for “accurate” figures, as I have held, then the difficulties that waiting would have produced bore most materially on de Wit’s decision to act when he did.

To summarise: The transfer regulations contemplated the employment of actuarial methods in the determination of the benefits to be credited to the applicants on their departure from the Fund. Those calculations obliged the actuary to make assumptions in respect of contingencies. These included the fact that at the time he performed the calculations he was confronted with uncertain and unreliable membership data. The actuary acted reasonably in making the determination when he did, and the data loading factor of 7,5% that he applied to the applicants cannot be faulted. His determination therefore fulfilled the requirements of the statutory provision under which he was acting, and it cannot be set aside.

The appeal is accordingly upheld with costs, including the costs of two counsel. The order of the Court below is set aside. In its place there is substituted: ‘The application is dismissed with costs, including the costs of two counsel.’

E CAMERON JUDGE OF APPEAL

VIVIER JA )HARMS JA )ZULMAN JA ) CONCURMTHIYANE JA )

[1]Act 41 of 1963

[2]GN R1653, GG 5285 of 10 September 1976

[3]GN R821, GG 15665 of 22 April 1994

[4]Appointed under Regulation 24A of the general regulations, inserted by GN R191, GG 11133 of 12 February 1988, to provide for regular valuation of the Fund by an “actuary”.

[5]Reg 2(4)(b), of the transfer regulations, read with reg 2(1)(c) and reg 3(1)(b)

[6]Constitution of the Republic of South Africa, Act 200 of 1993

[7]See Investigating Directorate: Serious Economic Offences and Others v Hyundai Motor Distributors (Pty) Ltd and Others: In re Hyundai Motor Distributors (Pty) Ltd and Others v Smit NO and Others [2000] ZACC 12; 2001 (1) SA 545 (CC) par 21

[8]Section 24 (a), (b) and (d) of the interim Constitution, which before the Promotion of Administrative Justice Act 2 of 2000 came into operation on 9 March 2001 remained applicable also under the final Constitution.

[9]Reg 1(2)

[10]The Shorter Oxford English Dictionary gives the relevant meaning of “actuary” as “one whose profession it is to solve monetary problems depending on Interest and Probability, in connection with life, fire, or other accidents, etc.”

[11] 1999 (4) SA 884 (SCA) par 16

[12]Amalgamated Packaging Industries Ltd v Hutt and Another 1975 (4) SA 943 (A) 951C-D