THE SUPREME COURT OF APPEAL
OF SOUTH AFRICA
REPORTABLE
Case number: 268/03
In the matter between:
THE ASSOCIATED INSTITUTIONS
PENSION FUND FIRST APPELLANT
THE MINISTER OF FINANCE SECOND APPELLANT
THE DIRECTOR GENERAL OF
FINANCE THIRD APPELLANT
LEON DE WIT FOURTH APPELLANT
and
JOHAN VAN ZYL & 1 699 OTHERS RESPONDENTS
CORAM: HARMS, MTHIYANE, BRAND, CLOETE
JJA et COMRIE AJA
HEARD: 26 AUGUST 2004
DELIVERED: 17 MAY 2004
Summary: Pension funds – review of actuarial determination of 'funding
percentage' in terms of regulations promulgated under Act 41 of 1963 –
interpretation of regulations – review of determination on substantive grounds
– undue delay in launching review application.
_____________________________________________________
JUDGMENT
BRAND JA/
2
BRAND JA:
[1] The first appellant ('the AIPF') was established in terms of s
2 of the Associated Institutions Pe nsion Fund Act 41 of 1963 ('the
Act') to provide pension funds for associated institutions. The latter
were mainly universities, technic ons and other institutions of a
similar kind. The second and third appellants are the Minister and
the Director General of Finance. T heir involvement is said to have
arisen from the fact that t he AIPF was administered by the
National Department of Finance. The appeal has its origin in the
large scale withdrawal by employees of associated institutions of
their pension interest from the AIPF to the newly established own
pension funds of the individual institutions during 1994 and 1995.
[2] One of the associated institutions was the University of
Pretoria. All the respondents w ere employed by that university,
and until 31 December 1994 they were members of the AIPF. On
that date their individual pension interests were transferred to the
Universiteit van Pretoria Voorso rgfonds ('the Voorsorgfonds').
These transfers took place pursua nt to regulations ('the transfer
regulations') that were promulgated in terms of the Act in
Government Notice 821 of 22 Apri l 1994. According to the transfer
regulations, the transfer value of each member's pension interest
3
as at the transfer date was to be determined by the actuary of the
AIPF. He is the fourth appellant in this matter, Mr Leon de Wit ('De
Wit'). Broadly stated , the respondents' ca se is that the
determination by De Wit was not properly made, with a resultant
shortfall in the amounts that were transferred to the
Voorsorgfonds.
[3] The determination of the tr ansfer values was finalised by
August 1995. Nearly four years lat er, at the end of June 1999, the
respondents launched an application in the Pretoria High Court for
De Wit's determination to be re viewed and set aside. They also
sought an order that the transfer values of their pension interest be
recalculated and that the reca lculated amounts, together with
interest, be paid over to th e Voorsorgfonds by the AIPF,
alternatively by the Minister. The court a quo (Botha J) held, upon
the strength of its inte rpretation of the trans fer regulations, that De
Wit's determination was not in a ccordance with these regulations.
Consequently the orders sought were granted. The appeal against
that judgment is with the leave of the court a quo.
[4] On appeal it was contended by the appellants that the court
a quo misconstrued the transfer regul ations. They also contended
that, in all the circumstan ces of the ca se, the court a quo erred in
4
not dismissing the review application on the ground of
respondents' unreasonable delay in bringing the application. The
issues arising from these c ontentions can best be understood
against the factual background that follows.
FACTUAL BACKGROUND
[5] The AIPF was governed by the pr ovisions of the Act and the
regulations ('the general regulatio ns') that were promulgated by
Government Notice 1653 of 10 Se ptember 1976. Pursuant to the
general regulations, the AIPF was s ubject to a statutory triennial
actuarial valuation. These valu ations were done on the basis of
audited figures. During the 1980 s and the 1990s the triennial
actuarial valuations revealed a substantial funding deficit. This
gave rise to concern on the part of AIPF members and the
associated institutions that were ultimately responsible for funding
their pension benefits. As a result, by 1993 many of the associated
institutions had for several ye ars been exerting pressure on the
government to allow members to transfer their pension benefits out
of the AIPF into privately adm inistered pension funds to be
established by each of the inst itutions for that purpose. This
pressure increased with t he anxiety of the inst itutions to have the
5
AIPF removed from government control before any political
changeover after the democratic elections in April 1994.
[6] The transfer regulations as eventually promulgated were
preceded by several drafts which were prepared under the
direction of a workgroup consisting of representatives of the
government and the associated institutions. The workgroup
included De Wit as well as two p rofessors of actuarial science,
Prof Anthony Asher of the University of the Witwatersrand and Prof
George Marx of the University of Pretoria. Subsequent to the
promulgation of the transfer regul ations, the Minister appointed an
advisory board of experts to oversee the transfer process. Different
members of the board were appoint ed to protect the interests of
the government, the associated institutions and the membership of
the AIPF. Professors Asher and Ma rx were also members of the
advisory board.
[7] In terms of the transfer r egulations, new pension funds,
called 'own established funds', c ould be established by associated
institutions for the benefit of th eir employees and members could
elect between keeping their pensi on benefits in the AIPF or to
transfer them to own established funds.
6
[8] According to the transfer regul ations, the transfer dates were
to be agreed upon by the Direct or General and ea ch associated
institution concerned, provided that such date coul d not be later
than 31 March 1995. At the ti me there were 211 associated
institutions. Many of them es tablished their own funds. Various
transfer dates were agreed upo n over the 13 month period
between April 1994 and March 1995. The transfer date for the
University of Pretoria was 31 De cember 1994. Eventually, over 80
per cent of the original 50 000 me mbers of the AIPF decided to
join the autonomous funds establis hed by their own institutions.
Included in their number were the 1 700 respondents.
[9] In terms of regulation 2(4)(b) of the transfer regulations the
AIPF had to make available to a member who elected to terminate
his membership, an amount ' equal to the funding percentage
multiplied by the actuarial obligation of the [AIPF] in respect of that
member as determined by the act uary on the date on which his
membership of the fund is term inated, with interest thereon
calculated at the bank rate from that date to the date on which the
amount is paid …' The member was then obliged to pay the
amount thus made available into the own established fund of his or
her institution.
7
[10] The 'actuary' referred to in regulation 2(4)(b) was De Wit.
There is no dispute with regard to his determination of the actuarial
obligations of the AIPF in res pect of each of the individual
respondents on 31 December 1994. It is the other factor in the
calculation provided for in regul ation 2(4)(b), ie the 'funding
percentage' of the AIPF, which f orms the subject of the dispute
between the parties. 'Funding percentage' is defined in the transfer
regulations. This definition, which is pivotal to the appeal, reads as
follows:
'"funding percentage", means the market value of the net assets of the [AIPF]
on a fixed date [ie 31 December 1994], expressed as a percentage of the
calculated aggregate actuaria l obligation of the [A IPF] on that date, as
determined by the actuary.'
Actuarial obligation' is in turn defined as signifying, 'with regard to
a particular member … of [the AIPF], … the actuarial obligation of
[the AIPF] with regard to that member … on a fixed date,
calculated by the actuary'. T he term 'actuary' is defined
independently as a reference to the actuary of the AIPF, appointed
in terms of the general regulations.
[11] De Wit determined the funding percentage of the AIPF for 31
December 1994 at 60 per cent . This determination was
communicated to the University of Pretoria on 6 February 1995
8
and on 23 August 1995 the individual tran sfer amounts pertaining
to those who elected to mo ve to the Voorsorgfonds were
calculated by him on that basis. These totalled some R286,5m
which was paid over to the Voorsorgfonds.
[12] In January 1996 the triennial statutory valuation of the AIPF
as at 30 September 1994, was co mpleted by De Wit and became
generally available. According to this valuation the funding
percentage as at the latter da te was fixed at 66 per cent.
Subsequently De Wit performed a f urther valuation of the rump of
the AIPF as at 31 March 1995 when those members who had
elected to leave the AIPF were fi nally identified. This valuation
yielded a funding percentage of 84,3 per cent as at that date.
These significant variations in the funding percentage over a
relatively short period of time led to discontent among those
members whose parting benefits were calculated on substantially
lower funding percentages.
[13] First to act were some 2 500 employees a nd pensioners of
the University of South Africa (' Unisa'). Their transfer date was 30
November 1994 and the fundi ng percentage of the AIPF
determined by De Wit as at that date was 60,8 per cent. In October
1998 they instituted review proceedi ngs similar to these in the
9
Pretoria High Court against the AIPF ('the Unisa case'). The initial
outcome of those proceedings favoured the Unisa employees. In a
judgment handed down by Southw ood J in February 2000, De
Wit's determination of the fundi ng percentage as at 30 November
1994 was set aside and the atten dant relief sought was granted.
On appeal to this court that dec ision was, however, overturned.
The judgment of this court, which was handed down on 31 May
2001, has subsequently been reported as Associated Institutions
Pension Fund and Another v Le Roux and Others 2001 (4) SA 262
(SCA).
[14] The applicants in the Unisa case contended that the root of
De Wit's inaccurate determinati on of the funding percentage at
their transfer date was to be fo und in his adoption of a 'data
loading factor' of 7,5 per cent to the actuarial ob ligations which
appeared from the records of t he AIPF in establishing the
aggregate actuarial obligations of the fund, thereby increasing
these obligations by a notional 7, 5 per cent. De Wit's explanation
for adopting this loading factor was that it was done in an attempt
to compensate for a known understa tement of liabilities in the
records of the AIPF.
10
[15] The application in the Unisa case was upheld by the Pretoria
High Court essentially on the basis that the transfer regulations did
not allow for the adoption of a lo ading factor in calculating the
obligations of the AIPF. What these regulations required of De Wit,
so the court found, was to calculate the aggregate actuarial
obligations of the AIPF on the ba sis of reliable data. According to
this interpretation, De Wit' s approach, which was to rely on
estimates and assumptions when faced with unreliable
membership data, was ultra vires the regulations. The reasons
why this court upheld the appeal against that judgment appear, in
short, from the following dicta by Cameron JA (in para 16):
'On a true construction t he transfer regulations r equired the invocation and
application of actuaria l expertise and that, on the uncontested evidence
before the Court as to the professional methodolog y involved, necessarily
entailed that assumptions would be m ade to allow for contingencies and
imponderables. That is the nature of the actuary's j ob, and it was a job the
regulations required De Wit to perform.'
And (in para 19):
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 [De Wit] erred. By the
methodology appropriate to what the r egulations required of him, De Wit
acted properly and lawfully at the time he made his determination. There is no
11
suggestion that the assumptions he employed were inappropriate or
unreasonable. The applicants' case … 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.'
[16] The respondents in this case did not contend that De Wit
was not entitled to apply any data loading at all in calculating the
obligations of the fund. In the light of this court's decision in the
Unisa case, such contention woul d have had little prospect of
success. Instead their first contention was that he should have
applied a much lower data loading percentage than 7,5 per cent. In
support of this contention they relied on the expert opinion
expressed by an actuary, Mr Michael Lowther, in an affidavit filed
on their behalf. This contention di d not find favour with the court a
quo and on appeal it was expressly abandone d on behalf of the
respondents. It accordingly requires no further consideration.
DETERMINATION OF ASSETS
[17] The only remaining objecti on by the respondents in this
case, which was not raised in the Unisa case at all, was that De
Wit had failed to determine the mark et value of the net assets of
the AIPF in accordance with the de finition of 'funding percentage'.
12
It was on the basis of this objecti on that the matter was decided in
favour of respondents in the court a quo.
[18] This objection goes to the heart of De Wit's whole approach
to his brief, as it appears from his answering affidavit. According to
De Wit it was known at an early stage that different transfer dates
would be chosen by the various inst itutions involved. One of the
crucial decisions De Wit was th erefore required to take at the
outset was whether the transfer valuations should be done by
using the latest available actuarial valuation of the AIPF, ie the
valuation at 30 September 1991, with appropriate adjustments to
take account of new data, or whether they should rather be held
over until fully audited data f or each transfer date became
available. By September 1993, De Wit, with the approval of the
work group, including Professo rs Asher and Marx, decided to
adopt the former appro ach. This decision was subsequently
endorsed by the advisory board. According to De Wit, the decision
was motivated by the consideratio n that the latt er approach would
have occasioned an initial delay of about 18 months coupled with a
further delay of about four months for the determ ination at each
individual transfer date. The pr ospect of such delays was
unacceptable to the associated in stitutions and their members
13
because they were, for the reas ons already stated, anxious to
finalise the transfers as soon as possible. Delays thus caused
would also have result ed in numerous practica l difficulties for the
newly established individual fund s and their members. Moreover,
the cost of every determination on the basis of fully audited data at
each transfer date would run into thousands, if not millions, of
rands. Although Lowther suggested that he would have awaited
fully audited data for each calc ulation, it was not contended that
De Wit's decision to the contrary was ultra vires the transfer
regulations or even unreasonable in all the circumstances.
[19] Once the selected appr oach had been adopted, the
methodology applied by De Wit was, broadly stated, to use the
September 1991 actuarial valuat ion, appropriately adjusted to
reflect all data available to him, to determine a 'base funding
percentage' for September 1993. As is apparent from the definition
of 'funding percentage', the concept is essentially comprised of two
elements, ie the market value of th e net assets of the fund, on the
one hand, and the aggreg ate actuarial obligation of the fund, on
the other. In order to establish the base funding percentage as at
September 1993, De Wit obtained complete information regarding
the assets of the AIPF from t he Public Investment Commissioners.
14
For the corresponding liability fi gure, he applied the actuarially
projected liability figures as at 30 September 1993, which included
the data loading of 7,5 per cent. Once that base funding level had
been determined for September 1993, it was then adjusted by De
Wit on a monthly basis to reflect ad hoc changes to the liability
profile of the fund and sh ifts in the market values of the assets of
the fund. In this manner a set of adjusted transfer values was
obtained for each month after S eptember 1993 unt il the transfer
process was completed.
[20] The assets of the AIPF consisted mainly of government
stock. Adjustments to the market value of these assets subsequent
to September 1993 were based on a financial model involving a
'basket' of government stock as at the end of each month. The
result of this methodology was that the only actual determination of
assets that took place was at 30 September 1993. Subsequent
adjustments to the assets were done on a notional basis. This of
course also happened on the transfer date pertaining to
respondents, ie 31 December 1994. It is common cause that as an
inevitable consequence of De Wi t's methodology, he took the
assets of the AIPF on 30 Sept ember 1993 and calculated their
value as at 31 December 1994. As a result, no account was taken
15
of any change in the number of AIPF's assets, due, say, to
increased contributions made from September 1993 to 31
December 1994.
[21] The court a quo found that on a proper construction of the
definition of 'funding percentage', De Wit's use of estimates and
projections in determining the assets of the AIPF as at the transfer
date was ultra vires the transfer regul ations. Though the
regulations permitted the use of actuarial estimates in the
determination of liabilities, so the court found, it was not allowed
in relation to assets. Since the court's reasons for this finding were
directly linked to the definition of 'funding percentage', I will for
convenience quote that definition again. It reads as follows:
'"funding percentage", means 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.'
[22] The court a quo's reasoning, with reference to the wording of
the definition, appears from the following part of its judgment:
'In my view the expression " market value" connotes the actual value and not
a notional value established by projections.
The value obtained by [De Wit] from t he Public Investment Commissioners at
the commencement of his exercise was such a value. The subsequent values
that he obtained by making projections cannot be described as market values.
16
It was argued that the phrase " as determined by the actuary " refers to the
determination of the market value. I cannot agree.
In my view it refers to the " calculated aggregate actuarial obligation of the
Fund". It tells one who is to calculate the aggregate actuarial obligation of the
fund, namely the actuary appointed in te rms of the general regulations. I
cannot see any licence in this phrase to the actuary to determine asset values
that are not market values.
The phrase " as determined by the actuary " was mentioned in the Unisa
case, but apparently as qualifying " actuarial obligation". See the Unisa case
supra, paragraph 9 at 268.
For all these reasons I have come to t he conclusion that the approach of [De
Wit] to determine the fundi ng level as at the 31
st December 1994 not with
reference to the market value of the assets, but with reference to a notional
value, was ultra vires the transfer regulations.'
[23] I agree with counsel for the app ellants that purely as a matter
of literal construction, the c ourt's reasoning c annot be sustained.
Interpreted along ordinary lit eral lines, the two commas in the
definition create a parent hetical clause so that the phrase 'as
determined by the actuary' qualifies the words 'market value of the
net assets of the fund on a fixed date'. If the phrase 'as determined
by the actuary' is to be linked ex clusively to the words 'calculated
aggregate actuarial obligation of t he fund', as sugg ested by the
17
court a quo, there is no explanation for the second comma in the
definition.
[24] Moreover, if the phrase aft er the second comma is to be
understood as applying solely to t he calculation of the actuarial
obligation of the fund, it will hardly have any independent meaning.
After all, 'actuarial obligation' is already defined in the regulations
as pertaining to calculations don e by 'the actuary' . Added to this,
the latter term is independently defined as a reference to the
actuary of the AIPF appointed in t erms of the general regulations.
The suggestion by the court a quo, that the phrase 'determined by
the actuary' was intended to id entify the person who was to
calculate the aggregate actuarial obligation of the fund, namely the
actuary appointed in the terms of the general regulations' therefore
cannot be accepted for this reason as well.
[25] Even more significant than t he literal construction, however,
is the consideration that the subs tantive result of the interpretation
adopted by the court a quo could not, in my view, have been
intended. Since this considerat ion also has a bearing on the new
interpretation of the definition contended for by counsel for the
respondents in this court, I shall elaborate upon it after considering
18
the arguments advanced in support of what I shall call 'the new
interpretation'.
[26] The new interpretation was advanced for the first time by
counsel for the respondents duri ng oral argument in this court.
Though it was presented as an al ternative to the interpretation
adopted by the court a quo , it is clear that the two constructions
are in fact mutually destruct ive. According to the new
interpretation, the phrase 'as de termined by the actuary' does not
relate to the words 'aggregate actu arial obligation of the fund' at
all. It pertains only to 'the market value of the net assets of the fund
on a fixed date'. But, so the argument went, 'determination' is to be
understood as requiring an actual , empirical determination as
opposed to the actuarial determinat ion. Accordingly, the reference
to 'the actuary' in the phrase 'as determination by the actuary' was
intended only to identify the functionary and not the function.
[27] The substructure for the new interpretation was exclusively
founded on the difference between the terms 'calculate' and
'determine' as used in the tr ansfer regulations. On a proper
analysis of the regulations, counsel for the respondents
contended, it is apparent that 'calculate' was consistently used with
reference to the 'actuarial obligat ion of the fund'. 'Calculate' must
19
therefore be understood as a reference to an actuarial function.
Since a change in wording must be construed to indicate a
different intent, 'determination' must be underst ood to mean the
opposite, ie a non-actuarial function.
[28] I do not consider that the tenuous foundation upon which the
argument rests justifies its adoption. In the Unisa case Cameron
JA referred at para 9 to the repeat ed use of the words 'actuary',
'actuarial' and 'actuarially' in th e transfer regulations as well as
'their insistent allusion to the ac tuarial function'. The conclusion to
be drawn from this is formulated as follows (para 10):
'Given the linguistic accumulation, t he phrase 'as determi ned by the actuary'
can hardly have been intended … only to i dentify the actuar y in whom the
regulations vest the power to perform t he calculations they enjoin. That the
instrument attains with economy and cl arity by a separ ate 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 it names.'
[29] I respectfully agree with t he aforegoing reasoning. Given the
emphasis on actuarial function, to which reference was repeatedly
made, I find it unlikely that t he legislature would have intended a
fundamental change in the nature of the functi on conferred upon
the same functionary within the scope of a single definition and
20
while prescribing what is in esse nce a single process, namely the
determination of a funding percentage . I find it even more unlikely
that the legislature would have i ndicated such a dramatic change
in intent through an almost imperc eptible change in expression. I
say almost imperceptible, because it is obvious that in the present
context 'calculation' and 'determ ination' are linguistically
interchangeable. Both terms can legitimately be used with
reference to either an actuarial or a non-actuarial function,
depending on the functionary entrusted with its performance.
[30] The starting point in de termining the meaning of the
definition is the decision of this court in the Unisa case, as is
accepted by the respondent s, that the actuarial obligations of the
fund were to be determined in a ccordance with ac tuarial practice.
In Tek Corporation Provident Fund and others v Lorentz 1999 (4)
SA 884 (SCA) 894G-895B, para 16, M arais JA explained what is
normally understood by 'actuarial practice'. It en tails, he said, a
highly sophisticated process requiring considerable training,
expertise and skill. But it remains, he said, an exercise in prophecy
involving a host of factors abou t which assumptions have to be
made. In the Unisa case it was accepted (at 270F-H, para 16) that
a determination in accordance with actuarial practice necessarily
21
involved the making of assumption s and predictions to allow for
contingencies and imponderables. It was not suggested by the
respondents that a determinatio n of assets cannot be done
actuarially, or even that it invo lves an exercise which an actuary
could or should not perform.
[31] Against this backgroun d, I can think of no reason in logic why
the legislature would have intended that, in determining the
funding percentage of the AIPF, the actuary should apply his
training, skills and expertise in establishing the aggregate actuarial
liabilities of the fund in accordance with act uarial practice, but that
he could not use the same attribut es in establishing the net assets
of the fund. According to De Wit' s testimony, it is a known fact of
actuarial practice and experience that assumptions which later
prove to be inaccurate often tend to cancel each other out. So, for
example, an underestimate of cont ributions to the fund, which will
affect the projected assets, ma y be matched by a parallel
underestimate of pensions, whic h will affect the actuarial
obligations, because increases in contributions usually have to
keep up with increases in pensions. If we accept this testimony, as
we are bound to by the well esta blished rules pertaining to motion
proceedings, acceptance of the interpretation contended for by the
22
respondents is so divorc ed from the reality of actuarial practice
and experience that it can be just ified on the acceptance of only
one of two possible assumptions. It must be assumed either that
the legislature had no conception of these actuarial realities, or
that it chose to ignore them. I have no reason to think that either of
these rather starling propositi ons should be accepted and none
was advanced by counsel. It follow s that the interpretation of the
definition contended for by the respondents cannot be sustained.
[32] On a proper interpretation t he transfer regulations, in my
view, enjoined the actuary to ac t, throughout the en tire process of
determining the funding percentage, in accordance with actuarial
practice. On this basis it was accepted in the Unisa case (in para
16), with reference to the determin ation of actuarial liabilities that
this necessarily entailed the applic ation of professional actuarial
methodology. By the same token, the regulations, in my view,
prescribed the selfsame actu arial methodology, involving
assumptions and predictions to allow for contingencies and
imponderables when it came to th e determination of assets. It
follows that De Wit's methodology per se was not ultra vires the
regulations.
23
[33] The respondents' alternative c ontention was that, even if De
Wit was in principle entitled to act on the basis of assumptions, he
erred in relying on assumptions re garding the assets of the AIPF
which were known to be invali d when the 31 December 1994
determination was made. In support of this contention it was
pointed out by the re spondents' actuary, Lowth er, that as a result
of De Wit's methodology, he did not establish what assets were in
fact held by the AIPF as at the transfer date of 31 December 1994.
Instead, he took the assets of the AIPF on 30 September 1993 and
calculated their value with referen ce to the prices of those types of
assets as at the transfer date. As a result of this methodology, it
was stated by Lowther in resp ondents' founding papers, De Wit
intentionally ignored additional a ssets accumulated by the AIPF
through an increase in employer contributions between 30
September 1993 and 31 Decem ber 1994, which assets it
appeared subsequently, had a value of approximately R160m.
[34] All this was conceded by De Wit in his answering affidavit.
He pointed out, however, that he had assumed that this increase in
assets would have been cancell ed out by an increase in greater
pension liabilities. In the event, De Wit stated, his actuarial
assumptions were borne out by subsequent events, because the
24
R160m was 'cancelled out' by an increase in lia bilities of about
R180m. In reply Lowther denied that the assets of R160m could be
regarded as 'cancelled out' by the liabilities of R180m. Moreover,
he contended that, on reflection, t he value of the assets excluded
by De Wit was not R160m but approximately R450m.
[35] I do not find it necessary to get involved in this debate
between the two actuaries. First, to the extent that it resulted in a
dispute of fact, we must prefer the version of De Wit (see Plascon-
Evans Paints Ltd v Van Riebeeck Paints (Pty) Ltd 1984 (3) SA 623
(A) 634G-635D). Second, insofar as Lowther's contentions in reply
constitute a new case, we are precluded from taking it into account
(see eg Director of Hospital Services v Mistry 1979 (1) SA 626 (A)
635H-636F). The most im portant consideration, however, is that
the debate takes the matter no furt her. Respondents' case is not
that De Wit's exclusion of a ssets, whatever their value may be,
was due to any ignorance or oversi ght on his part. He knew about
the existence of these assets a nd their exclusion was a deliberate,
but inevitable result of his methodology. That much is common
cause. Consequently, this is not a case in which a material mistake
of fact arose within the meani ng of the rule explained in Pepcor
Retirement Fund and another v Fi nancial Services Board and
25
another 2003 (6) SA 38 (SCA), para 47. Respondents' criticism is
aimed at the methodology a dopted by De Wit. Shorn of
unnecessary elaboration, Lowther's objection amounts to no more
than that he would not have adopted the same methodology as De
Wit because, so he said, it led to a result which eventually turned
out to be demonstrably unfair and unreasonably prejudicial to the
respondents.
[36] Since De Wit's decision was taken before the advent of both
the final Constitution and the Prom otion of Administrative Justice
Act 3 of 2000 (PAJA), it was acce pted by counsel on both sides,
rightly in my view, that the legal substructure for respondents' case
is to be found in s 24(d) of the interim Constitution. This subsection
– which was re-enacted as a t ransitional measure pending the
promulgation of PAJA in item 23(2 )(b) of schedule 6 to the final
Constitution – vested in every per son the right to administrative
action 'which is justifiable in rel ation to the reasons given for it'.
Although the subsection expanded th e ambit of judicial review, it
did not abolish the well established distinction between review and
appeal (see eg Carephone (Pty) Ltd v Marcus NO and others 1999
(3) SA 304 (LAC) 315C). Nor did it introduce substantive fairness
as a criterion for judging the valid ity of administrative action. That
26
much is clear from the explanat ion of the analogous provisions of
item 23(b) of schedule 6 in the ma jority judgment by Chaskalson
CJ in Bel Porto School Governing Bo dy and others v Premier,
Western Cape and another 2002 (3) SA 265 (CC) 291F-292G.
These provisions, the Chief Justic e said, encapsulate and extend
the common law grounds of judici al review – legality, procedural
fairness and rationality – as they have been devel oped over the
years in England and South Af rica. For good reason, he said,
judicial review of administrative action has always distinguished
between procedural fairness and substantive fairness. The
substantive unfairness of a dec ision in itself, has never been a
ground for review. Thereafter, t he Chief Justice proceeded to
express himself as follows (at paras 88, 89 and 90):
'I do not consider that item 23(2 )(b) of Schedule 6 has changed this and
introduced substantive fairness into our law as a criterion for judging whether
administrative action is valid or not. The setting of such a standard would drag
Courts into matters which, according to the separation of powers, should be
dealt with at a political or administrative level and not at a judicial level. This is
of particular importance in cases such as the present, in which the issue
relates to difficult and complex policies …
I do not understand the Carephone case, or any of the cases that have
followed it, to hold otherwise. What they require for a decision to be justifiable,
27
is that it should be a ra tional decision taken lawfully and directed to a proper
purpose.
If that is the case, and if the decision is one wh ich a reasonable authority
could reach it would, in my view, meet the requirements of item 23(2)(b) ….'
(Cf also Minister of Environmental Affa irs & Tourism and others v
Bato Star Fishing (Pty) Ltd 2003 (6) SA 406 (SCA) paras 46-50
and Bato Star Fishing (Pty) Ltd v Minister of Environmental Affairs
and others 2004 (4) SA 490 (CC) 512H-513A.)
[37] In this matter, De Wit gave a full explanation as to why he
decided to adopt the methodology th at he did. Though he realised,
he said, that a determination of the funding percentage based on
audited figures at every transfer date would be more accurate and
therefore more satisfactory to everybody concerned, that approach
was, in his considered view, ruled out by considerations of delay
and expense. He therefore had to compromise. Both his reasons
for compromising and the compro mised methodology itself were
thereafter endorsed by the advis ory board appoint ed to oversee
the transfer process on behalf of the variou s interest groups. The
members of that board include d two professors of actuarial
science, Professors As her and Marx. Amongst the affidavits filed
on behalf of respondents, is one depo sed to by Prof Asher. In this
28
affidavit he, inter alia , made the following unequivocal and
unqualified statements:
'I confirm that, as stated by De Wit, both Prof Marx and I were at all material
times aware of, and satisfied with, the professional methodology applied by
De Wit to the determinations and described by him in his affidavit.
I confirm also, that at the time, there appeared to be no possibility of De Wit
being allowed to defer making his determinations until after all the data for the
September 1994 triennial evaluati on had been obtained and audited. It
appeared that considerable covert pressure was being placed on the
Department of Finance to ensure that the transfer process was completed as
quickly as possible. We were instru cted that the transfer process had to
proceed as quickly as possible …
I have read both Lowther's criticisms of De Wit and De Wit's response to
Lowther's affidavit. I remain convinc ed that within the constraints of the
difficult process that was imposed upon hi m, De Wit made his transfer value
determinations in a manner that meets the professional standards expected of
a reasonable actuary.'
[38] Respondents also filed an affi davit by Mr Peter Milburn-Pyle,
an independent actuary of so me 40 years experience who
specialises in the valuation of pension funds and who was at one
stage the chief actuary of the Financial Services Board. He
confirmed that after consideri ng De Wit's methodology and
Lowther's criticism thereof, he came to the conclusion that De Wit
29
performed his task in a professional manner and that he followed a
reasonable methodology in doing so.
[39] Particularly in the light of the training, skills, experience and
intricacies involved in the applicati on of actuarial science, I believe
that this is a matter where judi cial deference is appropriate. Of
course, I do not mean judicial timidity, but judicial deference as
explained in Bato Star Fishing (Pty) Ltd v Minister of
Environmental Affairs supra para 47. In these circumstances it is
almost self-evident, I think, th at the respondents have failed to
make out a case that De Wit' s methodology was not one which an
actuary could reasonably have adopt ed, ie that De Wit had failed
to act rationally in the execution of his brief. For these reasons the
appeal must, in my view, succeed.
UNREASONABLE DELAY
[40] The remaining issues stem from the challenge of the
application on grounds of un reasonable delay. Given my
conclusion on the merits, these is sues can make no difference to
the outcome of the appeal. Neverthe less I consider it necessary to
decide them lest it be thought that the court a quo's view on when
the defence of unreason able delay will be uph eld, is endorsed by
this court. To begin with, I revert to the facts.
30
[41] In their founding papers , the respondents gave no reasons
why their application for the review of De Wit's determination of the
funding percentage was launched nearly four years after the
implementation of that decision in August 1995. In the answering
papers the appellants pertinent ly raised the defence of
unreasonable delay. In support of this defence they pointed out
that, apart from the fact that nearly four years had lapsed since the
decision had been implemented, t he respondents' application was,
on their own version, triggered by De Wit's valuations of the AIPF
as at 30 September 1994 and 31 March 1995. According to the
appellants, both these valuations were freely available at the
Department of Finance in July 1996. There was therefore no
reason why the application coul d not have been launched shortly
thereafter. Moreover, appellants pointed out, Prof Marx, who held
a Chair at the University of Pretoria, was a member of the advisory
board that oversaw the transfer process to its completion, while the
objections raised by the respondent s arose from decisions taken
by De Wit with the full knowl edge and approval of the advisory
board. All this, the applicants st ated, was apparent from the
minutes of the meetings of the advisory board that were circulated
to all the associated instituti ons, including the University of
Pretoria. Furthermore, the appellants contended, the Unisa case
31
was launched in October 1998 a nd the respondent s must have
become aware of those proc eedings shortly thereafter.
Nonetheless, it took them anot her eight months to launch the
present proceedings.
[42] As to the prejudice caus ed by this alleged unreasonable
delay, the appellants pointed out that about 80 per cent of the
members of the AIPF withdrew from the fund over the transfer
period. Assuming that the pres ent application were to be
successful and all the transferring members were to become
entitled to a further payment representing the difference between a
funding level of 60 per cent and, say, 66 per cent, the total of the
capital amount involved would be approximately R500m. To this
should be added the interest at bank rate from date of transfer to
date of payment prescri bed by regulation 2( 4)(b) of the transfer
regulations. The effect of all this would be, depending on when the
matter is finally resolved, that the AIPF may be required to pay out
amounts of between one billion and one and a half billion rand, or
between approximately 15 per cent to 22 per cent of the total
assets of the fund.
[43] The capacity of the AIPF to c arry this loss, so the appellants
contended, had been substantially affected by the changing
32
composition of the membership of the fund. The withdrawal of
members of the AIPF had left the fund with a substantially reduced
proportion of contributing me mbers. Whereas the number of
contributing members had dropped by 91,3 per cent, the total
number of pensioners had dropped by only 13 per cent. As time
passed since 1995, they said, th e membership of the fund had
aged further, the numb er of contributing members had shrunk
even further and the number of pens ioners had increased. If the
present application had been brought without delay, the AIPF
would have been able to ameliora te the position in a number of
ways by, for example, negotiating an increase of contributions from
contributing members and/or thei r employers or by adopting a
different asset management strategy with a view to maximise short
term returns on investment. Apar t from all this, the appellants
averred, they were severely ham pered in the prepa ration of their
defence so long after the event.
[44] The first respondent was the only applicant who deposed to
a replying affidavit. The other resp ondents did not respond to the
appellants' allegations of unreas onable delay at all. The first
respondent's answer to these alle gations was in essence that,
although he had known since about August 1995 that De Wit had
33
determined the funding percentage of the AIPF at 60 per cent, he
had had no reason to question the accuracy of this determination
until he became aware of the Unisa case which was launched in
October 1998. He conceded that the present proceedings were
triggered by De Wit's subsequent valuations of the AIPF that were
both released in 1996. He denied, however, that these valuations
were freely available at the time . In any event, first respondent
contended, he had no reason to s eek access to these valuations
or to the minutes of the advisory board nor, for that matter, to seek
the advice of Prof Marx until he realised that there was something
amiss with De Wit's determination. That only happened, he said,
when he heard of the Unisa case. Consequently, the time period
which preceded the Unisa case should not be counted for the
purposes of deciding whether or not there had been an
unreasonable delay. With reference to the period that had elapsed
between the Unisa case and the launch of the present
proceedings, first respondent's co ntention was that in all the
circumstances eight months could not be regarded as an
unreasonable delay.
34
[45] The court a quo found that the defence based on the lapse of
time could not be sustained. Its reasons for this finding were
formulated as follows:
'There is no reason to doubt the word of first applicant that he was never
aware of the possibility of review until he got wind of t he Unisa application.
Even if he had sight [of the valuation r eports pertaining to 30 September 1994
and 31 March 1995] there would not have been reason for him to obtain
elucidation which, in this case, would have entailed obtaining actuarial advice.
… The argument that there was a duty on the applicants after the
determination had been communicated to them, to go behind the decision and
to establish whether it could be taken on review, must be rejected.
That means that I am of the view that the period preceding the Unisa
application cannot be counted for the purposes of deciding whether there had
been an unreasonable delay. What is left, is a matter of eight months. If one
looks at the application, its technical complexity and the number of persons
involved, I simply cannot find that a period of eight months was unreasonable
to bring the application to fruition. …'
[46] Since PAJA only came into operation on 30 November 2000
the limitation of 180 days in s 7(1) does not apply to these
proceedings. The validity of th e defence of unreasonable delay
must therefore be considered with reference to common law
principles. It is a longstanding rule that courts have the power, as
part of their inherent jurisdiction to regulate their own proceedings,
35
to refuse a review application if the aggrieved party had been guilty
of unreasonable delay in initiating the proceedings. The effect is
that, in a sense, delay would 'va lidate' the invalid administrative
action (see eg Oudekraal Estates (Pty) Lt d v City of Cape Town
and others [2004] 3 All SA 1 (S CA) 10b-d, para 27). The raison
d'etre of the rule is said to be twofold. First, the failure to bring a
review within a reasonable time may cause prejudice to the
respondent. Second, there is a publ ic interest element in the
finality of administrative de cisions and the exercise of
administrative functions (see eg Wolgroeiers Afslaers (Edms) Bpk
v Munisipaliteit van Kaapstad 1978 (1) SA 13 (A) 41).
[47] The scope and content of the rule has been the subject of
investigation in two decisions of this court. They are the
Wolgroeiers case and Setsokosane Busdiens (Edms) Bpk v
Voorsitter, Nasionale Ve rvoerkommissie en 'n Ander 1986 (2) SA
57 (A). As appears from t hese two cases and the numerous
decisions in which they have b een followed, application of the rule
requires consideration of two questions:
(a) Was there an unreasonable delay?
(b) If so, should the delay in all the circumstances be condoned?
(See Wolgroeiers 39C-D.)
36
[48] The reasonableness or unr easonableness of a delay is
entirely dependent on the fact s and circumstances of any
particular case (see eg Setsokosana 86G). The investigation into
the reasonableness of the delay has nothing to do with the court's
discretion. It is an investigation in to the facts of the matter in order
to determine whether, in all the ci rcumstances of that case, the
delay was reasonable. Though th is question does imply a value
judgment it is not to be equated with the judicial discretion involved
in the next question, if it arises, namely, whether a delay which has
been found to be unreasonabl e, should be condoned (See
Setsokosane 86E-F).
[49] The finding of the court a quo was that the applicants' delay
was not unreasonable. The court did not decide that there had
been an unreasonable delay which it elected to condone. The
appeal against the court's finding un der consideration is therefore
an ordinary appeal against a findi ng of fact and law. It does not
involve an appeal agains t the exercise of a judicial discretion by a
court of first instance.
[50] Pivotal to the approac h adopted by the court a quo is the
proposition that, since there was no duty on the respondents to go
behind De Wit's decision and to establish whether it could be taken
37
on review, any delay which preceded their knowledge of
reviewability should be dis regarded. I cannot agree with this
proposition. It will inevitably lead to the result that however supine
and unreasonable the applicants might have been in their failure to
investigate the validity of an admin istrative decision affecting their
rights and however long it migh t have taken before they were
independently alerted to some flaw in the decision, the delay
caused by their ignorance s hould be disregarded. Acceptance of
the proposition will undermine the two considerations underlying
the recognition of undue delay as a substantive defence. There will
be no finality in administrative decis ions and those affected by the
review will have to suffer what ever prejudice comes their way
through the applicant's supine attitude.
[51] In my view there is indeed a duty on applicants not to take an
indifferent attitude but rather to ta ke all reasonable steps available
to them to investigate the revi ewability of administrative decisions
adversely affecting them as so on as they are aware of the
decision. These considerations are , in my view, also reflected in
both s 7(1) of PAJA and in the prov isions of s 12(3) of the
Prescription Act 68 of 1969. Whet her the applicants in a particular
case have taken all reasonable steps available to them in
38
compliance with this duty, will depend on the facts and
circumstances of each case. (Cf Drennan Maud & Partners v
Pennington Town Board 1998 (3) SA 200 (SCA).)
[52] Accordingly, it was legally in sufficient for the first respondent
simply to allege that he was unaware of any pot ential irregularities
in De Wit's determination until the Unisa application was launched
several years later. The ques tion remains: did he take all
reasonable steps available to inve stigate the reviewability of that
decision after he became aware that it had been taken? In my view
he did not. A reasonable person in hi s position would at least have
made enquiries, either from Prof Ma rx or from De Wit, as to how
the determination was made. This trail of enquiry would have led
him to the documents already avail able at his own university and
eventually to the facts that formed the basis of his review
application. Since the other respondents gave no explanation
whatsoever for their delay, which was prima facie unreasonable,
they cannot be in a better position than the first respondent.
[53] The finding that respondent s' delay was unreasonable, leads
to the next enquiry, that is, whether that delay should be
condoned. This question, I be lieve, should also be answered
against the respondents. Of prima ry concern is the severe
39
prejudice that the delay in t he setting aside of De Wit's
determination would have caused, not only to the appellants, but to
a large number of people and institutions who might have arranged
their affairs on the basis of it s presumed validity. Included among
these would obviously be the rem aining members of the AIPF and
those who have in the meantime taken retirement. The nature of
their prejudice appears from the appe llants' factual allegations that
I have referred to and which, though contradicted in some respects
by respondents, must be acce pted for purposes of motion
proceedings.
[54] For these reasons the appeal is upheld with costs, including
the costs of two counsel and the fo llowing order is substituted for
that of the court a quo:
'The application is dismissed with costs, including the costs
of two counsel.'
…..…………..
F D J BRAND
JUDGE OF APPEAL
Concur:
HARMS JA
MTHIYANE JA
CLOETE JA
COMRIE AJA