|
|||||
An Equitable Assessment of Rights and Wrongs by Dr Michael Nassim Appendix I: Extracts & Comments on the Discussion of the Ranson-Headdon Paper |
|||||
|
Appendix I:
Extracts & Comments on the Discussion of the Ranson-Headdon Paper Section
I: London, March 20th
1989 Roy
Ranson presented the paper in London8.
Though more could have been taken from the discussion, the following
suffices: Mr
J.H.R. Tonks among others defended the estate principle, and from him the
following extracts are pertinent: “Assuming that the fund continues to accept new business,
we could endeavour to reach the simplicity of the author’s situation by
considerably increasing the bonuses paid to the current generation of
policyholders. To my mind there are
two major objections to this course. First,
it is inequitable to pay the present generation considerably more than they have
earned. Secondly, if bonuses are
artificially increased in this way the fund will attract more new with-profits
business and so hasten the time when bonus distribution returns to normality.
At that time the new policyholders will become disenchanted with the
situation, because they have received less than their expectation.
Thus I believe that, in practice, the estate will continue in being for
the foreseeable future.” This is,
of course, a considerable understatement of what eventually materialised.
He concluded: “A fifth item needs to be added to the information to be
provided for a policyholder as set out in § 4.3.6.
This is the relative size of the estate of the office which he is
contemplating joining. That is not
to say that the policyholder would always place his business with the office
with the highest relative estate, and I accept that there are many policyholders
who would prefer an office with no estate, once they were able to understand
that such an office’s bonus philosophy is in line with the authors’, in that
it has paid full asset share in the past and proposes to do so on future claims.
Other policyholders may prefer to join an office which has a safety net
of an estate, even if they realise that this has been built up by paying less
than asset shares in the past, which could indicate a similar policy in future.
Whatever the outcome, I believe that they should have this
information, and that it is our duty to them to devise some means of doing
it.” Mr
C.S.S. Lyon reiterated the value and functions of the components of a
traditional estate, as did G.K. Aslet who followed him, e.g. “…It does not
seem to be difficult to explain that a mutual life office has an existence apart
from the interest of those who happen to form the current generation of
policyholders. This was surely the intention of the office’s founders.
To do otherwise seems to invite the opportunistic argument that the
office should be wound up immediately so that its riches can be distributed
among its members. Such
a course would not appeal to those who need insurance and not a windfall
profit.” –and again: “… The
authors also criticise the current emphasis placed on office strength, and I
realise that this is fashionable. I
do not want to imply that any single ratio can adequately describe the
characteristics of a life office, but its strength does have virtues beyond
that of preventing the DTI intervening in an office’s affairs, and
there are many policyholders who might wish that their advisers had paid more
attention to this point when recommending their policies.
A strong office has greater freedom to adopt the investment policy it
believes will be most profitable in the long run and thus most beneficial to its
policyholders. Such an office also
has more time to adjust to changed conditions before being forced to take
action.” Again with hindsight
this looks like an understatement. Mr H.W. Froggatt concisely and
efficiently summarised the issues surrounding an intrinsic mismatch of risk and
reward for the different categories of policyholder in relation to the
investment mix held. He began by saying that guarantees cost money, and one
could meet this by appropriate charges, and/or by reducing their impact by
holding a suitable proportion of fixed interest securities.
The latter is also particularly important if there is no reserve estate,
and so it is in the event doubly unfortunate to the extent that, in the scramble
for higher overall returns, it may later have been neglected (see R. & H.
section 3.2.7 & 3.2.8). H.W.
Froggatt continued: “...So, for
the authors’ office, it seems that the total guarantees are catered for by
adjusting the aggregate investment portfolio.
The office, as a whole, holds investments appropriate to its guaranteed
liabilities. It also exercises some
control by adjusting the reversionary bonus rates. In neither case is account taken of the very different levels
of guarantee being provided for policies of different original and outstanding
term; nor for policies of different classes.
In effect their policyholders are charged the same for guarantees
regardless of the level of guarantee which applies to their policies at the
time. This can leave room for
selection against the with-profits policyholder in some circumstances. Insurance operates by pooling
homogeneous risks. If risks are
significantly heterogeneous there comes a point at which it becomes worthwhile
to distinguish the different levels of risk and to underwrite and to charge
different rates of premium for them. In
the context of asset allocation for with-profits business, the relevant risk
element is the investment guarantee. So
it is appropriate to ask at what level of guarantee it is worthwhile to abandon
the simplicity of the authors and differentiate between policies with different
levels of guarantee?”
And in his conclusion he stated:
“The simplicity described in this paper has its price; and existing and
new policyholders- and their advisers, if they have any- ought, in the present
climate of disclosure, to be made aware of what these might be”. Section
II: Edinburgh, 19th
February 1990 The Ranson-Headdon paper had a
second outing the following year on the 19th February at the Faculty
of Actuaries in Edinburgh10. In
introducing it Christopher Headdon took the opportunity to explain the latest
developments, and to attempt to dissuade the meeting from discussing the estate
issue in so much depth as previously in London.
He began by saying that at the end of 1989 the market value of the Fund
assets stood at around £5,700 million and that after the transfer of
appreciation to revenue to support the bonus declaration the “investment
reserve” amounted to about £1000 million.
How much of the latter was accrued unconsolidated benefits and how much
was remnants of the previous estate was not explained.
Significantly, he went on to say that a further development to the bonus
system had been made: “Paragraph
3.2.18 of the paper foresees the development of the business so that total
policy proceeds, that is the sum of consolidated and unconsolidated benefits,
steadily accumulates from year to year in a way which would enable published
final bonus scales to be done away with. Such
an approach is consistent with our general philosophy in that we consider all
business of whatever term or duration in force to experience the same uniform
asset mix. At the time of writing, we had
seen this development of our bonus systems as a possibly long-term development. However, circumstances have changed, and we find it desirable
to make such a move sooner rather than later.
Accordingly, the new approach was introduced in respect of 1989.”
And a little later: “Our
recurrent single premium business is, effectively, equivalent to unitised
with-profits business. Since most
unitised with-profits business is fairly new, the terminal bonus element in
policy proceeds is not yet very significant, and most offices appear to be
taking a fairly straightforward approach to this part of total proceeds.
By
contrast, we have been selling such business since 1956, and have substantial
tranches of business which have been in force long enough to have attracted a
sizeable final bonus element. Arriving at an approach which deals with the unconsolidated
element of total benefits in an equitable manner, in the face of a very wide
degree of variability in the timing and amount of premium payments, is, we feel,
a problem which we have encountered in advance of most offices.
It would be interesting to learn if actuaries with offices now writing
unitised with-profits business have different ideas of how to cope with terminal
bonuses when the business becomes more mature.”
All too evidently the Equitable never solved this problem either, and
ended up with unreal bonuses that crossed over into genuine liabilities. Headdon
then went on to say that there was not one myth of estate, but two:
“ The first myth is that all offices have an estate, in the sense of a
body of assets which belongs to no-one……….. The second myth is that by
having an estate, an office’s policyholders are much better prepared than if
no estate exists. If one has an
estate, then it seems to us that one of two approaches can be taken.
The first is that the estate really is a cushion that is available if the
need arises. The second is that it
is money held in a kind of trust, to be maintained and passed on to future
generations. If
the first of those approaches is taken, then additional protection is available
on a one-off basis for one fortunate generation of policyholders.
However, once used, the estate ceases to
exist and the office moves to our position.
In the second case, the estate
cannot be utilised except for some temporary additional smoothing, since it
would otherwise not be preserved for the future.
In that case, are policyholders really better protected?
Indeed, should not product particulars state that the office’s policy
is to maintain an estate of x% of assets, and that a charge of y% on the
investment earnings otherwise available to policyholders will be made in order
to support that policy”. The
sophistry in this passage is so obvious that it needs no formal explanation, and
in the underlined portion Headdon has begged the question as to how the
Equitable’s own approach should have been put to its representatives and
policyholders. In the event, of
course, that explanation never materialised. In opening the discussion Mr
S.T. Meldrum said: “The paper is not a rigorous mathematical proof of a new
actuarial principle. It is a clear
and straightforward description of a bonus philosophy tinged with pragmatism.
We must thank the authors etc”. He
also declared a natural sympathy with the approach taken in the paper, because
he had qualified at the Equitable. His
position was therefore more informed than most, and he stated it thus:
“The authors in paragraph 1.1.1 find themselves frustrated at industry
obfuscation in the form of resistance to un-bundling of expenses and an emphasis
on “strength”. They find their
escape in borrowing some of the concepts of unit-linked insurance and applying
them to the with-profits contract and in particular to one form of that which
makes up the bulk of their office’s business, a form of recurring single
premium with-profits pension accumulation policy.
Each premium paid secures a
slice of the fund calculated by accumulating the premium less expenses at a
guaranteed accumulation rate of 3.5% (i.e. the Guaranteed Interest Rate or GIR)
to the chosen pension date. Effectively
this is the “sum assured” of the policy.
Reversionary and terminal bonuses are declared on this sum assured with
the intention of returning full value to the policyholder at his pension date,
but with a “smoothed” investment return.
Since
1987 results have been shown to policyholders additionally in present value form
and for new contracts such as personal pensions this is the only form.
The contract is then effectively a unitised with-profit.
The authors in paragraph 3.2.6 define the policyholders key concern in
bonus declaration as the total proceeds with the question as to how much can be
declared and how much to emerge as terminal bonus as of secondary importance.
The increasing guarantee given by declared reversionary bonuses is
however central to with-profits business. In paragraph 3.2.15 the authors
describe how the final bonus lifts the declared bonus to an appropriate asset
share subject to averaging and smoothing. The
smoothing occurs two ways: firstly
over durations at that point in time, and secondly over time.
The essence of smoothing is
insurance of the investment risk. It
is a system which has inherent appeal but I have some practical difficulty with
its application. My first problem relates to the
smoothing over time of an investment risk whose first move is downwards.
It does not appear possible to do this within the ordinary understanding
of prudence unless there is some other source of capital available.
This appears to have been the situation described in Appendix A for the
triennium ending 31 December 1976. The
situation was then met by releasing unnecessary margins in the valuation basis.
It is in this sense that I describe the approach in this paper as one of
pragmatism rather than one of rigorously proven theory.
It is in the area of
policyholder’s self insurance of investment risk that I believe a lot more
could be done to make the techniques more scientific.
A life company can self-insure mortality risks with the experience, good
or bad, reflected in the returns to with-profit policyholders.
But to do so it must have some measure of the expected risk to be charged
initially to each policyholder and on the basis of which the subsequent
experience is then spread. This is
the principle of equitable assurances. I
cannot see a corresponding principle of equitable investments described here.
How sure are the authors that
the averaging over contracts fully reflects the risks of those contracts?
How sure are the authors in averaging over time of where the market then
stands? Hindsight is a great help,
but without it the consensus of all the players in the market is that the next
move is as likely to be down as up. The
market reflects all that is known. I
would like to hear more on how the authors improve on this.
The process of an increasing
level of guarantees under each policy as bonuses are declared while the totality
of policies still participates in the investments underlying the whole fund is
one whose pricing basis lies in option pricing theory…Although the practical
difficulties may be large I recommend that some thought be given to this because
it is in this area that I am least comfortable” Here S.T. Meldrum seems to be
making much the same points as H.W. Froggatt had previously, although starting
from the analysis of risk rather than from guarantee.
The underlined portion is now important in terms of what should have been
disclosed in view of how matters were actually represented, as well as to whom. Also relevant to issues of
disclosure was his conclusion: “The
authors have done us a great service in exposing these issues and there is a lot
of material here which those present can use to compare with the philosophy of
their own offices. I would like to
think that more offices would be encouraged to expose their philosophy in this
manner. I look forward at least
to reading a mini version in the many company booklets which are shortly to be
produced”. Mr P. Kilgour provided a
comprehensive and well-balanced overview of the situation.
The following extract from his commentary is relevant to the vexed
question of guaranteed versus unconsolidated benefits, and what the
corresponding duties of information might be:
“ The authors suggest that with-profits policyholders expect to receive
policy proceeds on maturity broadly equal to those that would have been achieved
under a unit-linked contract invested in a balanced managed fund.
I wonder whether all policyholders would agree on the investment mix
in a balanced managed fund. A lack
of clarity about the target at which they think their with-profits contract is
aiming is likely to result in confusion. The balanced managed fund
against which they should be comparing is one which is appropriate given the
guarantees in the contract, and any attempt to describe an appropriate portfolio
of investments for a with-profits contract must include a comment on the
relevance of the guarantees. The significance of these
guarantees is substantially reduced if declared bonus rates are at a level
financed by only a fraction of the full investment return being achieved.
The authors do not say so but I wonder if they are suggesting that in
future, policyholders can expect that, whether there is inflation or not,
declared bonuses will be at a level financed by a fraction of the investment
return achieved- thereby creating a continuance of the current picture for
maturities in that a substantial element of the maturity payout will be in the
form of a terminal bonus. This
seems in concert with their claim that with-profits business is a source of
capital- in fact I do not see how it is unless such a bonus policy is pursued. The forthcoming statements on bonus philosophy will probably
contain a relatively detailed explanation of a company’s current stance but
will not tie the company to a continuation of that stance.
Over the duration of their contracts, policyholders will be able to gain
access to up to date statements on bonus philosophy and will be free to make
their views known to the appropriate parties should any changes not meet with
their approval.”
To this one can only say: “Amen,
and if only!” Mr Kilgour also
supported the role and value of a “dowry” or estate in the usual ways, and
explained his own preferences for adopting an appropriate investment mix based
on duration elapsed and term to run, and which also reflected the level of
guarantees. Mr W.B. McBride spoke as one
whose office had strong parallels with the authors’ as a non-commission paying
mutual, the major part of whose business was also single premium with-profits,
which had been growing rapidly in recent years.
However he continued: “Our trading experience,
however, as is public knowledge, has been rather different. Again until recently, my office
would probably have subscribed, although more subconsciously than the authors,
to their view of the mythical nature of the rationale of the estate, even of its
existence. We would not hold that view
today. Instead, we recognise the estate as a precious attribute of
the office, inherited from the past, yes, but to be husbanded, and handed on to
future generations of with-profits policyholders. We would recognise the estate as the difference between the
value of the assets and of the published liabilities, plus the present value, at
the required rate of return, of the future stream of profits from the business
(other than profit attributable to policyholders). This measurement should reconcile with the difference between
the value of the assets and the total of the asset shares of current
policyholders. We would not consider the fact
that various methods of making these measurements exist, so that the size of the
estate at any moment is not an absolute and precisely measurable quantum, as any
barrier to accepting its reality. As
to its rationale, our experiences have reinforced the view, held by many
actuaries, that the estate is of benefit to the current generation of
policyholders in a number of ways, notably the power it conveys to the office to
smooth out the effect of fluctuations in the equity markets. Given that shares now change
hands every hour of the 24 somewhere in the world, and that there is
instantaneous reaction to news good or bad, smoothing power has never been more
important. Where there is no estate, and
the investment reserve represents the unconsolidated earnings attributable to
current policyholders, one would expect terminal bonus rates to be rather more
volatile than I believe has been evident in the declarations by the office of
the authors. I immediately
dismiss as unworthy the suspicion that perhaps they do not know their own
office’s strength ……. …..We do not claim that our
asset share calculations represent precise reality- they could not hope to in
practice and ought not to in theory, or the process reduces to unit linking- but
we believe that their relativity to one another is valid.
The process implies, of course, a significant degree of smoothing which
has recently been made practicable again for us by virtue of becoming backed as
a sub-fund of an international mutual, by a powerful estate.”
The omitted intervening section echoed previous observations on risk,
liability, asset matching and bonus distribution philosophy. Alas, the
underlined statement proved all too worthy in the event; the only argument now
being over how much was hubristic ignorance and how much was concealment. Mr C.E. Barton agreed with the
authors over the estate issue, but may have been mistaken in observing: “I
very much like the author’s preference for the term “non-consolidated
assets” rather than the somewhat mysterious term “the estate”.
When Redington coined the word “estate” in 1952 he was much more
concerned with solvency than with the equitable management of a participating
fund. Twenty nine years later,
against the background of “The Flock and the Sheep”, Redington’s concern
was with “non-consolidated assets” rather than what has been called an
“inviolate estate””. Did he
realise that this included the liability estate, or that the “asset estate”
might come to exist only in notional form as unconsolidated and unsupported
bonus statements? Mr M.D. Ross was under no such
illusions. He said:
“There is very much in the paper with which I agree and which reflects
my own views on bonus policy. However,
there are some points brought out in the paper with which I disagree; I think
they are based on what I would term as too serene a view of what future
conditions might bring… …Recognition must be given to
the likelihood, at some time or another, that the underlying pro rata asset
shares will fall below guaranteed payouts and some charge must be retained for
this- logically it would vary with policy term. While reference is made to the
pooling of investment mix I do not see this point emphasised in the paper.
I see it as important, very important unless the non-guaranteed element
can be set very high- perhaps at levels currently applying for long-term
policies. However, as I have said
on other occasions, it is rather difficult to see such high targets reflected in
current reversionary bonus declarations/unitised with-profits price increases. This leads me on to the section
entitled “the myth of the estate.” Having
done stochastic modelling it is not difficult to see the need for some free
assets, certainly with the current valuation regulations.
Of course the authors refer to the unconsolidated element of
policyholders’ asset shares as being available to meet finance strains and
guarantees. However again the
average duration of an office’s business is important and an office’s
ability to keep the unconsolidated element alone high enough at all times to
guarantee survival in a wide range of financial conditions has to be challenged.
Free assets are likely to be required over and above the
unconsolidated bonus element….. …In paragraph 3.3.5 reference
is made to the nature of the guarantees and its relevance in the context of
product design. Examples are given
of full value guarantees at a range of retirement ages, e.g. 50 to 75.
Interpretation of Regulation 62(2) in its strictest sense is likely to
prove very severe in terms of reserving requirements over time, again stochastic
modelling is likely to indicate that an office could not demonstrate statutory
solvency over time without very considerable free assets- yet the authors eschew
the need for this. Something must
be wrong. It may be that it is
envisaged that the terminal bonus element will always be maintained at a very
high level, but of course in these circumstances the granting of a full value
guarantee is not particularly meaningful. I
raise this, not particularly to take issue with the authors, but more because we
have found in our own modelling that a strict interpretation of Regulation 62(2)
can prove very severe for pension contracts.
I would be interested to know whether any other Appointed Actuaries would
assume all policyholders within the guarantee period would retire immediately or
if, what I would call, a more realistic view of this can be taken in
practice.” Mr Ross was right;
something was indeed wrong. More could be quoted from these
discussions, for example the contributions from A.D. Shedden, W.A.B. Scott, and
A. Eastwood. Scott, who gave
qualified support to the authors’ concept of the myth of the estate provided
that adequate financial strength was otherwise maintained, also entered the
following caveat: “I think we can agree that it
is this modified version of the managed fund concept which has to be
communicated to policyholders if the mystery of with-profits business is to be
dispelled. On this front the authors may be ahead of most of us,
assisted no doubt by the simple nature of the majority of their contracts and by
the absence of intermediaries in their dealings with policyholders.
Clearly there must be a risk of raising unreasonable expectations if
communication of this nature is not carried out with due care and attention but
on balance I am sure that this is the road down which we must go” In view of subsequent events,
this opinion suitably concludes the Appendix.
|
|
||||
