Coronavirus (COVID-19) updates from the Australian Government

Is there an aggregate superannuation savings gap


The projections in this paper indicate that replacement rates will be lifted by the Better Super reforms and continue to rise strongly reaching levels of 70% or more in the medium term, levels that are usually judged adequate by most commentators. The replacement rates here are higher than projected in the author’s 1999 paper, despite longevity increasing faster than previously expected. Based on this analysis, once the SG arrangements are fairly mature, prospects are quite bright for future cohorts of retirees and there is no savings shortfall or saving gap.

The Investment and Financial Services Association (IFSA, 2003) have argued that there is a saving gap based on work by the actuarial firm of Rice Walker (R-W). A Superannuation Savings Gap is a measure of the current shortfall in national savings between two amounts:

  • the national saving required to ensure “adequacy” in retirement; and
  • the amount saved in the superannuation system, and estimated to be saved up to retirement, by (most of) the current workforce.

IFSA have chosen to represent any shortfall as a lump sum amount, though it could be expressed as an additional amount that needs to be saved on a per annum basis over the future working lifetime of the current workforce. IFSA initially estimated this gap at $600 billion but have noted that their estimate has fallen when revisited in 2006 (IFSA, 2006). The revised estimate is about $450 billion.

Basically the R-W modelling appears to be a cohort group model quite like RIMGROUP, but somewhat less comprehensive. For example, the analysis performed by Rice Walker does not make any allowance for people with incomes below 0.5 x AWOTE or above 2 x AWOTE, nor does it directly model the value of the age pension. The author notes that his understanding of the R-W model is based entirely on (limited) published information.

Nonetheless the current analysis challenges their results. To resolve this difference the author has attempted to use the R-W framework to assess the gap (if there is one) and to try and understand the factors that might lead to a different conclusion being drawn. There are numerous possible factors involved, including.:

  • R-W uses a high gross measure of adequacy that exceeds 75% net for many income levels
  • R-W tends to treat the age pension as a bonus rather than as an integral part of retirement income.
  • They exclude savings outside super; and
  • Use a relatively low earning rate, around 1 percentage point lower than used in the current analysis (which is itself at the conservative end of historical 30 to 40 year averages).

While the above factors tend to overstate the gap, R-W use an assumption of 15% total contribution rate (including 4% voluntary) which is significantly higher than the average contribution rates used in RIMGROUP.

While some of these factors are matters for judgement and precision is difficult because of data limitations, the critical problem that the author has with the R-W analysis lies in R-W’s view of the role of the government age pension. and their unacceptably low estimate of the aggregate value of these pensions.

There seems very little basis for R-W or IFSA’s contention that the government age pension should be regarded as peripheral to consideration of retirement incomes rather than an integral and important part of those incomes. In 1998 the Government legislated to link pension rates to Male Total Average Weekly Earnings, guaranteeing that the single full-rate pension would be at least 25 per cent of MTAWE. Both the First and Second Intergenerational Reports ((Intergenerational Report, 2007)) show how critical future age pension costs are to future fiscal balances. The Second Intergenerational Report included projections of the proportions of the age pension population that are projected (using RIMGROUP) to receive a full, part or no government age pension. (Chart 5 )

Chart 5: Coverage Projections for the Age Pension, All Deciles

The impact of the projected higher wealth and income of retirees over time is shown in the clear pattern of decline of full rate pensioners and in the projected rise in part-rate pensioners and non-pensioners in the Chart. Notwithstanding the higher wealth and income of retirees over time, the age pension remains an integral part of most retirement incomes and is here to stay

Turning to consideration of the R-W estimate of the aggregate value of the age pension, we note that the R-W report estimates the value of the age pension during retirement at $100-200 billion (in 2003 dollars), a value which the author believes to be grossly understated.

Suppose that the estimate of the age pension value was at the upper bound ($200 billion), in comparison to R-W’s projected superannuation asset base for the cohorts in question of about $1,200 billion. The implication of this is that the report assumes that for every $1 in age pension paid to this cohort, the cohort will provide over $6 in private expenditure. Given that contribution rates appear to be too high, using R-W’s framework we can re-estimate private saving in the R-W model at about $900 billion. And so their ratio becomes 1 to 4.5. But published Treasury analysis of hypothetical cases (eg to the 2002 Senate Inquiry) and the data underlying Table 1, shows that for a typical worker earning around AWOTE, retirement income is made up about equally from private sources and from the age pension ie about one to one not one to four to six. This suggests the pension contribution to the R-W aggregate should be about $700 billion higher, entirely eliminating the claimed gap. The pension component will be even higher following the halving of the asset test withdrawal as part of Better Super. As a check RIMGROUP has been used to produce a separate estimate of the net present value of the age pension for approximately the group covered by the R-W report; a similar figure around $850 billion is calculated as the aggregate value of the age pension. In the RIMGROUP case the future pension payouts for the given income ranges have been brought back to a net present value using a 6% discount rate.

Using the structure of the R-W projections, accepting their sensitivity analyses as valid, but adjusting for what appears to be a major flaw in estimating the value of the government age pension, the author’s estimate of the savings gap is around zero, with the exclusion of savings outside superannuation and other factors overstating the gap broadly balancing the apparent overstatement of contribution rates. Dependent on assumptions made, after the Better Super changes, there may be a small savings surplus.