We all know about the low savings rate in South Africa and how few people in South Africa are able to retire comfortably.
The questions are ‘Why?’ and ‘What can we do about it?’, and there will always be ‘Whose fault is it?’.
First, a bit of background. Let’s start by looking at the original objective. A pension fund is designed to provide meaningful payments to members in their retirement: such funds are set up as a vehicle for saving for their post-working years.
To encourage this, most governments provide tax incentives and subsequent limitations to prevent retirees from taking a lump sum and either squandering it or making poor investments themselves so that it is soon all gone. It is now a highly regulated industry.
In any discussion of retirement funds, it is important to separate two types of fund, largely because of who bears the investment risk.
Defined benefit versus defined contribution funds
A defined benefit (DB) fund has a ‘sponsor’, usually the employer. The pensions is paid by way of a promise from the fund, often based on final salary and years of service, independent of the fortunes of the performance of the fund. The funds could suffer actuarial surpluses or deficits. While morally obliged to pass on any surplus to pensioners through increases in their pension payments, the sponsor bears the risk of the deficit and is responsible for covering it over a reasonable period. As a consequence, most have given way to or been converted to defined contribution (DC) funds.
Under the DC system, the assets of the contributing members equal the liabilities.
There is a further distinction within DC funds: those that allow payment of their own pensions and those that don’t. The only way a DC fund can get surpluses or deficits is by actually paying the annuitants. If a deficit does arise from this, clearly it would not be fair to ask the contributing members to reduce their share of the fund to cover this or to reduce the pensions that are being paid. However, pension increases could be halted. Or the employer could be asked to make a special payment.
To counter this problem (of mixing investments for the pensioners with those of the contributing members in a single fund) the rules of most retirement funds – and a condition in SA – allow, or require, any new retiree to take their share of the fund and purchase an annuity from an approved third party, usually a life office licensed to do so.
This means there is a constant outflow from ‘the fund’ as each retiree takes their share. It is also true that most of the assets of any retirement fund belong to those nearing retirement age. If funds pay their own pensions (rare), those assets also tend to be a large percentage of the total.
‘Fund performance’ myth
Note that ‘fund performance’ in the case of a DC fund is a myth! Yes, considered as an entity, the trustees need to see that the assets and new contributions are well managed but each member gets their own performance, depending on how large or small current contributions are relative to accrued assets.
Actuaries, whose specialist knowledge of mortality and compound interest, provide estimates of the value of the liabilities based on a set of assumptions. These are often for long periods, especially for DB funds, with pensions often for 40 years or more. But as it turns out, some assumptions have not been met in practice: mortality has been much lower than expected and investment returns have not met expectations either. They may do in the ‘long term’ but this offers little solace for those most impacted by low returns and lower-than-inflation increases in pensions.
For DC funds that do not pay their own pensions, there is almost no need for any ‘estimates’, as the investment returns are credited to each member monthly (and could be negative).
Another myth in DC funds, which benefits those doing calculations using assumptions related to age-related mortality and investment returns, is the need to provide the so-called ‘net replacement ratio’ (NRR). This is based on each member’s current assets and assumptions about expected future salary increases, to which expected investment returns are applied to get the expected fund value per member at the ‘expected retirement date’. Making further assumptions about the pension that could be expected on that fund value, the ratio is that monthly amount relative to the expected monthly salary at retirement date!
Note that the expected salary progression used in such calculations ought to be very different for different classes of contributing members. However, a general formula applied to all is usually used.
The reason I say that it is a myth is because many members do not stay with the same fund for their entire working life. Many executives will have come in much later than the new junior employees and most often will have put their previous employment retirement savings into a preservation fund – of which the current fund/trustees have no knowledge.
Expectations today and tomorrow
Worse, consideration of the output of an NRR analysis is often used to increase the contribution rate if there are many whose NRR is too low – indicating an inadequate pension – and an injunction that it is the responsibility of the trustees to ensure adequate pensions. As a general rule, ‘adequate’ is defined as two thirds of final salary. But the NRR calculations for the young usually show that they can be expected to retire comfortably – mainly as a result of an assumption of a reasonably high ‘real return’ on their future contributions for a long time.
The truth is that many will leave and grab their share of the fund (with taxation implications) and are thus most unlikely to retire comfortably, let alone well. This is likely to be stopped by legislation, as such behaviour at young ages has consequences for the state later.
Much is made of the costs associated with the investments: a much larger cost is that of administration which, by definition, is more complex when having to keep additional records on a monthly basis and reconciling funds received from the employer or contributor, especially if that extends to ‘member choice’.
A lot of this fails to answer the ‘What are we trying to achieve? question.
Policy statement ‘confidence’
Very few pensioners who retired more than 20 years ago are still able to get by on what were probably initially reasonable pensions.
An investment policy statement setting appropriate investment mandates – usually based on long-term assumptions of real returns from different asset classes (in the form percentages of the funds to each asset class, thereby creating a ‘benchmark’ to measure the performance of the managers as well as peer-group comparisons) – creates an often-spurious feeling of confidence about future outcomes.
In many cases, the chosen investment managers are left to decide whether to overweight or underweight any asset class (relative to the benchmark).
Does this all seem to be a far cry from concerns as to whether the pensioners are living well?
As in many situations, we have to look at the outcomes of all these efforts (which cannot be underestimated in terms of the expertise and time of professionals as custodians of Other People’s Money).
As mentioned earlier, the sad part of all of this is that there are very few pensioners who retired more than 20 years ago (a growing percentage due to longevity, itself a function of better lifestyles, modern medicine and lack of ‘hard times’ such as famine) who are still capable of getting by on what were, probably, initially reasonable pensions!
In Part 2 I discuss why it pays to be aware of the effects that inflation and interest will have on your pension pot.
* Liston Meintjes is a private wealth portfolio manager at NVest Securities.