Often the easiest thing to do is criticise a proposed solution without offering an alternative. I empathise with Finance Minister Tito Mboweni who has the tough task of crafting a budget in difficult economic times, and steer the mutiny-prone ship that is South Africa from stormy seas to more placid waters so that South Africans can ultimately enjoy the services they deserve from a government they elected democratically.

In his maiden budget speech 2019, Finance Minister Tito Mboweni says the public service wage bill is unsustainable. He may be correct. He says the solution is to reduce it by R27 billion over the next three years, and that this will be done by encouraging public servants to take early retirement. This may not be a prudent budgetary approach  as it will have inevitable negative consequences. He deftly left it to Minister of Public Service and Administration Ayanda Dlodlo to “outline the details of the early retirement framework”.

Armed with this rather blunt budgetary scalpel, Minister Dlodlo is certain not only to be on a collision path with public servants but is also likely to ratchet up the cost of her surgical intervention on the fiscus. And who do you think will take the fall for the fallout that may follow?

Public servants are members of defined benefit pension funds, like the Government Employees Pension Fund (GEPF). The benefits of each employee are defined based on the employee’s pensionable salary, pensionable service and actuarial factor. Each of these cannot be determined with any accuracy in advance as this is an exercise that involves complex actuarial principles.

For that reason, some observers may argue that the Finance Minister’s announcement that Minister Dlodlo will reduce the public service salary bill by R27 billion over the next three years is either a pipe dream or a deliberate election stroke. At least three sets of considerations would explain such scepticism.

Firstly, for the Minister to know in advance by exactly how much the public service wage bill will be reduced over the next three years, he must be in possession of facts on which the employment of general actuarial principles is based in order to arrive at a desired outcome. These include

  • how many employees will take early retirement over the next three years;
  • what the salary of each of these employees will be over that period;
  • whether each of these employees will receive salary increases over that period and how frequently;
  • how many of these employees will die before taking early retirement;
  • how many of the employees who take up early retirement will be married in the next three years;
  • how many of these employees will resign before taking up early retirement;
  • how many of them will be dismissed for misconduct over the next three years;
  • how many of them will be re-employed and therefore return on the government payroll.

These are facts that even the pension fund’s actuary cannot know in advance, and in respect of which he can only make an educated guess using the tools of his trade, actuarial principles. So, presumably untrained in the actuarial science discipline, how can the Minister possibly know all this? Did he seek advice from people skilled in this discipline? What did they tell him? Is he following their advice or did he ignore such advice?

Secondly, for the Minister’s project to be realised there will most probably have to be some government (employer) interference with the fund actuary’s work and that of the trustees who owe a fiduciary duty not to government but to the pension fund. The reason for this is that the means by which politicians want to achieve budget cuts in public service may not necessarily be aligned with what is in the best interest of the fund. That being so, government interference would not only be unlawful; it would also pose the risk of an unmitigated disaster for the financial soundness of the fund.

Thirdly, early retirement from a defined benefit fund is not cheap for the employer – in this case, the government – because there is an early retirement penalty to be paid. The amount of that penalty will depend on the salary of each employee taking early retirement at a given time, the number of employees taking early retirement, and the number of years that each affected employee has until reaching actual retirement age. Depending on these factors, that could be a huge bill. Ultimately, it is the taxpayer who pays that penalty.

So, until the Minister has shown the nation these numbers, it is difficult to understand how he could possibly have arrived at the R27 billion number as one by which the public service wage bill will be shaved over the next three years.

Now, let us consider each of the three grounds for scepticism on this announcement.

(1) On General Actuarial Principles, the Minister’s R27 Billion Project is Difficult to Grasp

Quite apart from the final salary that each employee who takes early retirement receives at the time, and the number of years recognised by the fund as pensionable years of service, an important factor that the fund actuary takes into account when calculating each retiring member’s actuarial interest (benefit) is what is called actuarial interest factors. This is an actuarial pursuit without which it is impossible to work out each employee’s early retirement benefit in defined benefit funds.

The computation of actuarial interest factors is a complex actuarial discipline which entails reliance on actuarial assumptions that fall into two broad categories, namely, demographic assumptions and financial assumptions.

The broad principles in the calculation of actuarial interest factors can be simplified as follows:

  • Understanding the purpose and calculation of actuarial reserve values is critical in understanding the purpose and application of actuarial interest factors and why they change over time.
  • An actuarial valuation of a pension fund – done every two years in the case of the GEPF – is done in order to check (among other things) whether the fund has sufficient assets to pay out the benefits promised to members as they become due. If the GEPF has sufficient assets, it is considered to be financially sound. If not, it is financially unsound, and so additional contributions may be required from the employer or other remedial action (such as decrease in actuarial interest factors, less benefit payments to pensioners, more tax revenue collection) may have to be taken by the trustees of the GEPF.
  • The liability value that is determined at the valuation date (the member’s “actuarial reserve value” or “actuarial interest”) is only in respect of the portion of benefits that relates to service up to the valuation date. The portion of benefits in respect of service after the valuation date is funded by future member and employer contributions to the GEPF.
  • The GEPF is a defined benefit pension fund in which the benefit payable to a member on retirement is set out in the rules and calculated using formulae based on the member’s pensionable service and pensionable salary at the date of retirement. When each member will retire (e.g. at the normal retirement age or before normal retirement age) is not known before the event actually occurs, nor is the salary he or she will eventually be earning at retirement. The exact amount of money needed to meet the benefit that will be paid on retirement (or on death, disability or resignation) is not known in advance and the actuary estimates this amount as part of the actuarial valuation.
  • The member’s actuarial interest at the valuation date is therefore the actuary’s best estimate of how much is required in order to be able to pay out the benefits due to the member in terms of the rules of the GEPF whenever the member actually leaves or retires from employment at any time in the future. This is obviously impossible to predict for a specific member, but the actuary can estimate the average amount required for homogenous groups of members, e.g. by age, gender, etc.

In brief, the actuary estimates the value of the benefits that may be paid to the member at any point in the future and discounts this value to the current time, i.e. the assets required now, which together with the investment return that will be earned on these assets, will be sufficient to pay the benefits as they fall due.

As indicated earlier, the actuary makes a number of assumptions in these calculations which fall into two broad categories, namely, demographic assumptions and financial assumptions.

Demographic assumption is the basis upon which it is assumed that:

  • the member will leave the GEPF, e.g. the likelihood that the member will leave by retirement (or early retirement) or death or disability or resignation. This assumption is critical as it determines the form of benefit that is payable to the member. A resignation benefit, for example, is much less than a retirement benefit and so will cost the fund less;  
  • the date on which the member will leave the GEPF, e.g. the likelihood that the member will retire, die, resign or become disabled at each future age;
  • the member’s marital status at the date of exit, e.g. whether the member will be married and how old the member’s spouse will be. Marital status affects the value of the benefits payable on death in-service and on retirement (a pension continues to be paid to the spouse on the death of the pensioner).

Financial assumptions entail questions such as:

  • what annual salary increases and promotional salary increases the member will receive between now and his or her exit from the GEPF. This is important as it affects the value of the benefit that will be payable on exit.
  • what pension increases will be granted to pensioners in the future. This will affect the value of any pension paid to the member (and subsequently to his or her spouse if the person is married) if the member becomes a pensioner in the GEPF through either disability or normal or early retirement.
  • what the investment returns are that will be earned in future. The assets required at the valuation date to pay the future benefits will be higher or lower depending on the level of investment returns that will be earned on the assets of the GEPF.

The actuarial valuation incorporates all of these assumptions and calculates an actuarial reserve value for each member. The total of these values is then the GEPF’s liabilities and these are compared to the assets in the fund to determine if the GEPF has sufficient assets to pay benefits as they fall due.

The actuarial interest factors are tables of factors at each age (described as F(Z) and A(X) depending on whether the member is younger or older than age 55, respectively) which are applied to the member’s pensionable service and pensionable salary at the date of exit. The higher the interest factor, the higher the benefit. And these factors may increase or decrease at each fund valuation. That is not something the Minister would have known during his budget speech.

The demographic and financial assumptions used in calculating actuarial interest values, and therefore to determine the actuarial interest factors, are not arbitrary assumptions determined at the whim of the actuary.

The demographic assumptions are determined from regular experience investigations of the GEPF, i.e. the profile of how members have retired, died, resigned or become disabled in the past and at what ages members have tended to leave are determined by examining the records of the GEPF. The longevity of pensioners, and their spouses if they are married, is also examined by investigating the mortality experience of the GEPF’s pensioners. This experience is used to construct “decrement tables” which are used in the next actuarial valuation.

The demographic assumptions derived from the experience investigations will change over time for various reasons, such as:

  • The profile of the GEPF’s membership changes over time, e.g. a greater proportion of female members.
  • Behavioural changes with regard to resignation and retirement, e.g. greater employment mobility could result in more resignations at younger ages or financial conditions could result in less members retiring before normal retirement age.
  • Mortality improvements, due to such issues as better healthcare and living conditions, result in less members dying during employment and pensioners living longer in retirement.
  • Promotional salary increases (above general annual salary increases) are treated like a demographic assumption as past increases can be derived from the GEPF’s member data and used as an assumption for future increases.

So, the demographic assumptions derived from each experience investigation, historically done every four years or so, after interrogation by the GEPF’s Benefits and Administration Committee and the trustees for reasonability, are normally used in the next actuarial valuation to determine actuarial reserve values. The changed assumptions can result in higher or lower actuarial interest values. Neither the Minister nor his advisors can possibly know this in advance.

The financial assumptions, namely, future general salary increases, pension increases and investment returns, are all linked to future expected inflation. For example, future general salary increases can be assumed to be a constant amount above inflation each year, pension increases a constant percentage of inflation each year, and investment returns reflect a “real return” above inflation each year. The salary increase and pension increase “gaps” to inflation can be regularly tested and reviewed based on the GEPF’s experience.

Future expected inflation and real investment returns can be derived from the financial markets at any point in time, e.g. expected inflation can be derived by examining the yields on each of appropriate conventional government bonds and inflation linked bonds.

These changes in the financial assumptions derived from the market and the experience of the GEPF can result in higher or lower actuarial interest values.

Given that financial markets are constantly moving, different financial assumptions could theoretically be derived every day. If an actuarial valuation were done for the GEPF every day this would result in different actuarial interest values (ignoring any change in membership, service and salary details) and therefore different actuarial interest factors. This would be theoretically appropriate as the member would then receive the actuarial interest value applicable on his or her date of exit.

For administrative and practical reasons, and because formal actuarial valuations are only undertaken every two years, revised actuarial interest factors are only derived as part of each formal actuarial valuation of the GEPF and used for a period of two years until the next actuarial interest factors are derived.

Revised actuarial interest factors are therefore not determined in order to provide higher or lower benefit values to exiting members. They are revised to ensure that the member receives his or her actuarial interest in the GEPF based on demographic and financial assumptions that are applicable at the relevant time. This actuarial interest is the present value of expected future benefit payments and changes whenever the demographic and financial assumptions change.

Payment of an amount other than the appropriate actuarial interest value could be considered unfair to the member (if the amount is lower) or unfair to the GEPF (if the amount is higher).

Put differently, the benefit paid to the member is always the same, namely, it is the member’s actuarial interest in the GEPF. It is the amount of the actuarial interest that changes over time as the actuarial interest factors change, based on the demographic and financial assumptions that are applicable at the relevant time.

So, actuarial interest factors can increase or decrease depending on the demographic and financial assumptions that the fund actuary has made during the bi-annual actuarial valuation. It is thus difficult to grasp how the Minister and his advisors could possibly know, three years in advance, what those actuarial interest factors will be in three years’ time when there is every possibility (if not probability) that these may be revised upwards or downwards within the next two years depending on demographic and financial assumptions used by the fund actuary based on the fund’s financial soundness at valuation date.

(2) Short of Government Interference with Trustees and Actuary, it is Difficult to See How the Minister’s R27 Billion Project Can be Achieved

As pointed out above, with reference to general actuarial principles, it is difficult to understand how the Minister’s project of reducing the public salary wage bill by a specific amount (R27 billion) over the next three years by encouraging public servants to take early retirement can be pre-determined.

But even assuming the Minister could achieve this Herculean task of shaving off R27 billion from the public service wage bill in the next three years, is early retirement of public servants a wise move? Those who understand how defined benefit funds, like the GEPF, work may say no. Here is why.

(3) Early Retirement may be a Costly Exercise Ultimately for the Taxpayer

The suggestion by the Finance Minister that the public service wage bill be reduced by R27 billion over the next three years by encouraging public servants to take early retirement may be considered by some to be irresponsible.

There are two provisions of the Public Service Act that regulate early retirement: s 16(2A) and s 16(6). These must be read together with the applicable provisions of the GEPF Law, specifically, s 17(4) and Rule 14.3.3 of the GEPF Rules. From a plain reading of these provisions, it becomes rather clear that the early retirement option may not be the most efficacious or financially prudent.

Section 16(2A) says

“(2A) (a)          Notwithstanding the provisions of subsections (1) and (2)(a), an officer, other than a member of the services or an educator or a member of the Agency or the Service, shall have the right to retire from the public service on the date on which he or she attains the age of 55 years, or on any date after that date.”

Section 16(6) of the Public Service Act says:

“(6) (a)             An executive authority may, at the request of an employee, allow him or her to retire from the public service before reaching the age of 60 years, notwithstanding the absence of any reason for dismissal in terms of section 17(2), if sufficient reason exists for the retirement.

(b)            If an employee is allowed to so retire, he or she shall, notwithstanding anything to the contrary contained in subsection (4), be deemed to have retired in terms of that subsection, and he or she shall be entitled to such pension as he or she would have been entitled to if he or she had retired from the public service in terms of that subsection.”

Section 16(4) says an employee who reaches age 60 may be retired from public service subject to the approval of the relevant executive authority. Excluded are members of the armed forces, teachers and state security personnel.

The material differences between s 16(2A), on the one hand, and s 16(6), on the other, are these: s 16(2A) sets early retirement age, as a right, at 55 (normal retirement age is 60) but does not entitle the employee to full retirement benefits as if s/he had reached age 60. Section 16(6), on the other hand, does not confer early retirement as a right. The employee requires the approval of the executive authority (usually the relevant Minister); “sufficient reason” is required for ministerial approval; an employee can take early retirement at any age before 60; and the employee is entitled to full retirement benefits as if s/he had attained the age of 60.

In practice it is s 16(6), for obvious reasons, that is susceptible to much abuse. A 30 year old engineer at Eskom could under this section conceivably take early retirement, receive the pension of a 60 year old, and be re-employed on contract for his scarce skills set to the same position from which he retired, earning the same salary while receiving a pension.

But it is the effect of s 17(4) of the GEPF Law, read together with Rule 14.3.3(b) of the GEPF Rules, that may raise some eyebrows and which render this early retirement option less efficacious from a prudent budgeting point of view.

Section 17(4) of the GEPF Law says:

“If any action taken by the employer or if any legislation adopted by Parliament places any additional financial obligation on the Fund, the employer or the Government or the employer and the Government, as the case may be shall pay to the Fund an amount which is required to meet such obligation.”

The “action” by the employer would be the approval by the executive authority of early retirement in terms of section 16(6)(a) of the Public Service Act. Once that “action” takes place, Rule 14.3.3(b) is triggered. That rule says where a public servant who has been in public service for at least 10 years reaches the age of 55, or where a teacher reaches the age of 50, that employee’s early retirement benefit will be reduced by one-third of one percent for each complete month between his actual date of retirement and the retirement date as prescribed by the rules. In other words, where an employee takes early retirement at 55, the employer is liable for one-third of one percent for every month between the date of that employee’s early exit and the month on which he turns 60.

That one-third of one percent is the early retirement penalty, or pensionable service “buy back” that is for the employer’s account, i.e government which in essence means the taxpayer. It is the additional financial obligation on the fund which could never have been foreseen by the fund actuary when determining that employee’s actuarial interest at fund valuation stage with reference to demographic and financial assumptions.

Depending on the seniority of the civil servants who are prevailed upon to take early retirement, the indispensable skill they possess, the number of months they still have to go until reaching actual retirement age, and their commensurately high final salaries, the one-third of one percent that government will have to pay in each case may run into millions of rand and the overall figure may not justify the R27 billion sought to be shaved off the salary bill – especially if many of these will in any event be rehired in the same positions (typically on fixed-term contracts) in order to keep the state machinery ticking over.

Is this a prudent budgetary solution to a public service wage bill? Perhaps not. But what alternative does the opposition have?