Issue: March/May 2008


Crunched by numbers

Basic demographic projections suggest that the costs of funding National Treasury’s proposed reforms cannot possibly be justified by the retirement provision from which a relatively small percentage of the population will benefit. HEIN H HERBST argues that the ‘other policy objectives’, related to social security, will either hit taxpayers hard or be at the expense of other government services..

A distinguishing feature of National Treasury’s second discussion paper on social security and retirement reform is the glaring absence of any attempt to analyse future demographic trends of the South African population.Such an analysis, based on trends clearly discernable during past years, should be an absolute prerequisite. This is especially so if it is accepted that the policy proposals on retirement provision (should they be implemented) will only mature fully after a contributory period of 30 to 35 years.

A presentation in South Africa by Jorge Piņera, the father of retirement fund reform in Chile, made it abundantly clear that the full impact and benefits of retirement fund reform only become evident two to three decades from inception. While it may be difficult to predict and quantify the demographic composition of any population 20 to 30 years from now, some empiric research is surely preferable to shooting in the dark.

Data from the 2007 mid-year population estimates, issued by Statistics South Africa in Statistical Release P0302, indicates:


Group African % Coloured % Indian/Asian % White %
Over age 60 2 482 300 6,52 314 400 7,4 126 900 10,81 857 800 19,71
Over age 65 1 617 400 4,25 201 900 4,75 80 300 6,85 575 300 13,22
  • The total South Africa population was 47,9 million;
  • For 2007, life expectancy at birth was 48,4 years for males and 51,6 years for females, with (average) life expectancy for both estimated at 50 years;
  • The estimated adult HIV-prevalence rate was equal to 11,1 percent of the population, with woman in the age group 15 to 49 having the highest HIV-prevalence rate of 20,4 percent. The implications of the latter figure on future fertility rates, as well as the life expectancy of this age group, are evident;
  • The total South African population over the age of 60 years was 3,78 million, equal to 7,9 percent of the estimated mid-year population;
  • The total South African population over 65 years of age was 2,47 million, equal to only 5,17 percent of the estimated mid-year population;
  • Different demographic trends between the different racial groups are evident from the mid-year population estimates (see table).
  • Only the Indian/Asian population group shows any semblance of a normal age distribution curve based on five-year interval age cohorts, while the three remaining population groups all indicate abnormal distribution curves.

Possible future trends

During the past decade, life expectancy at birth has been steadily declining. Latest estimate of Stats SA is 50 years for both sexes. During the next decade, based on past trends, this figure is likely to decline further. Only extensive demographic modelling could produce reliable predictions within acceptable parameters of accuracy.

The percentages of the over-60 (or over-65) age groups in three of the four main population groups will undergo further reduction. Contributing factors are:

  • The white population in the age cohorts 20-24 to 40-44 indicates a large “missing” population. Only in the age cohorts 45-49 does the population distribution revert to “normal”. This “missing” population in the respective age cohorts can be ascribed mainly to emigration. The effect of this “missing” population will dramatically reduce the white population of retirement age (60 or 65) within two decades from now.
  • The already “abnormal” normal distribution curve of the black and coloured populations will, because of higher mortality rates in the fertile age cohorts, result in a decrease of these populations reaching “normal retirement age” (whether defined as 60 or 65 years of age);
  • Net effect of these anticipated trends will be that the percentage of the population reaching retirement age (60 years or 65 years) is likely to halve by 2025, reducing the percentage of population of retirement age to about 1,8 million if retirement age is determined at 60 years, or 1,25 million if retirement age is determined at 65.

As a percentage of population at mid-2007, this will equate to an estimated four percent of population if retirement age is assumed to be 60 years, or 2,5 percent if it is 65 years.

Already provided

It’s generally acknowledged that, since the promulgation of the Pension Funds Act in 1956, South Africa has developed a private retirement system admired even by “developed” countries. The occupational and retirement annuity schemes have been nurtured and encouraged by two “pillars of inducement”, viz a lack of tax-funded social security systems (except for the modest old-age pension introduced in 1928) as well as tax incentives to provide for own-retirement capital over the economically active life of an employee.

Savings endeavours over 50 years have brought South Africa to the stage where, according to the latest Registrar of Pension Funds report, existing funds have some nine million contributing members and an asset base of about R900 billion (doubtless now much higher because of subsequent share-market appreciation).

On the nine million members, some double counting occurs where the same person is a member of two or more pension funds. It is more difficult to quantify correctly the number of pensioners in private funds.

In a feasibility study of retirement reform released by the Department of Social Development last September, the total age-eligible population is put at slightly over three million for mid-2005. Of these, it’s stated that 2,1 million actually receive the social old-age pension (soap). And of these, 165 000 are ineligible to receive the pension (for reasons not given). A total of 645 000 people are age-eligible, but not eligible if the means test is applied.

So it’s fair to conclude that there are about 800 000 people receiving non-state pensions. The Social Development report refers to this group as “income-ineligible older people”. It must be conceded that this figure is not necessarily correct as an indication of the number of people of retirement age who receive a pension, as the income source disqualifying them from the soap might be other investment income, or application of the means test might also have disqualified them. In the absence of a more reliable figure, assume the number of pensioners in pension funds (including retirement annuities) is about 800 000.

Based on data in the Social Development report, about two million people are eligible by age and income for the soap. This equals 4,2 percent of the total population, assuming the Stats South Africa mid-2007 population estimates are correct.

Costs versus benefits

National Treasury’s paper estimates the total cost to the fiscus of the “wage subsidy proposal” to fund the “social security retirement arrangement” at R20 billion to R30 billion a year. Based on the lower figure, the cost to the fiscus (taxpayer) over a 20-year period will be R400 billion (at 2007 rand values). If the projected demographic trends materialise, how is it justified that such a magnitude of tax resource can be allocated for the estimated 2,5 percent to four percent of total population reaching retirement age? This question must be asked, and answered, in the context of competing demands for funds and services from central government.

At the soap’s level of R870 per pension per month, the cost to the taxpayer in the 2007-08 fiscal year was roughly R22 billion. National Treasury’s paper is silent on funding arrangements for the existing group of soap beneficiaries after 2010 when its proposed dispensation is supposed to take effect. The question arises whether the funding arrangement as it now exists will continue, or whether the liability will be transferred to the proposed “National Social Security Fund” supported by a “social security tax” piggy-backed onto income tax.

Hard conclusions

If the primary rationale of National Treasury’s proposals is retirement provision at age 60 or 65, the cost to the fiscus is unjustified. The amount of money required to provide for a small percentage of the population, who’ll reach retirement age in 20 years’ time, is heavily disproportionate against competing demands on government for basic services.

But if other policy objectives set out in the National Treasury proposals (unemployment cover, death and disability benefits, and “an inducement for low wages to rise to a targeted minimum wage”) are the primary rationale, they are made under the guise but not with the intent of retirement reform.

Instead, they’d represent a massive tax-funded scheme to provide for unemployment and employee benefits that are linked to an increased minimum wage. This has little to do with retirement provision at retirement age.

Hein Herbst is chairman of the R7,5 billion Cape Joint Retirement Fund whose membership base comprises 23 000 municipal workers.