Category: Front Page

Do retirees hoard their superannuation?

Download the article as a PDF

Jim Bonham*

On 22 January 2021, the Australian Financial Review featured a front-page article by John Kehoe and Michael Roddan headed “‘Ever more’ super gets hoarded: Hume”. 1

In the same issue, Jane Hume (Minister for Superannuation, Financial Services, and the Digital Economy) provided an op-ed “Safety nets let frugal retirees spend savings without a super rise”. 2

On 23 January 2021, Kehoe followed up with an article entitled “Push for seniors to dig deep into super nest egg” in which he wrote:

“Superannuation Minister Jane Hume kicked off a national debate about retirement incomes this week …

“She said people needed to be more confident to spend – not hoard – retirement savings to improve living standards throughout their lives …

“The government’s retirement income review led by former Treasury official Mike Callaghan identified that many retirees died with most of their wealth intact and did not run down their super or tap equity in their home, so they might be saving too much”. 3

It is clearly an important national question.  Wealth includes the home and other assets as well as super, but because the regulatory, financial, market, liquidity, and social issues in relation to housing differ so much from those applying to super, this article focusses only on super.

Is it true that retirees hoard their super?  The answer is in three parts:

  1. Yes, in nominal terms, in some cases,
  2. No, in real terms (indexed to wages),
  3. No, when considered as an average across all retirees.

The Minister’s view, as presented in the op-ed2 and reported in the articles mentioned1,3, is rather different, but it is strongly supported by the Retirement Income Review – Final Report (20 November 2020), chaired by Michael Callaghan4 (“RIR Report”).

A couple of quotes give the flavour of the RIR Report’s attitude (page numbers refer to the pdf version4): 

page 23, “Most people die with the bulk of the wealth they had at retirement intact.”

page 56, “The evidence suggests that retirees tend to hold on to their assets … Alternatively they need not have saved as much …”

It seems the way is being paved towards downgrading the level of compulsion applying to super contributions for pre-retirees and tightening the requirements for withdrawal in retirement. Such changes may be damaging to retirees if they are not soundly based on facts and understanding.

The counter-arguments to the claim that super is being hoarded by retirees need to be fleshed out:

  1. Nominal hoarding

Superannuation kept in an allocated pension account, as is typical for retirees, is subject to minimum annual withdrawal limits.  Those rates have been halved for 2019-21 because of Covid-19, but normally they are: 4% below age 65, 5% for ages 65-74, 6% for 75-79, 7% for 80-84, 9% for 85-89, 11% for 90-94 and 14% over 94.

Provided that investment returns can keep up with the minimum withdrawal rates from an allocated pension it is possible, with care, to leave the capital (in nominal dollars) untouched and take only the investment returns as income.

However, this becomes increasingly difficult beyond age 80 as the minimum withdrawal rates increase well beyond 7%, or at much younger ages if investment returns are low.

Hoarding of nominal superannuation capital throughout retirement is therefore possible, but only for those who die early or invest well.

  • Real hoarding

Nominal dollars provide a poor base for comparison across long time periods.  Real values, indexed to wages, relate much better to community living standards.  If the super account maintains its nominal value for 15 years, it will have lost almost half its real value (assuming 4% p.a. long-term wages growth). 

Successfully hoarding real capital between 65 and 74 years of age would require nominal investment returns, net of fees, consistently above 9%.  This is possible during good times, but almost impossible in bad times, and it becomes far harder as the superannuant ages further.

There is a simple reason for that: the minimum drawdown rates are designed to prevent long-term hoarding, whilst enabling those who live a long life to continue to benefit from their savings.

  • Average hoarding

It is easy to trot out simple examples to show that capital can or cannot be preserved in various scenarios.  From a policy point of view, however, what matters is the true average behaviour of all retirees. 

In support of the claim that retirees do not consume their capital, the RIR Report4 cites a paper by Polidano et al 5, and re-plots Fig 2 of that paper as Fig 5A-12 on page 434 (pdf version).  That graph shows average superannuation account values at a point in time, as a function of age, thus neatly dodging the inflation issue. 

At first sight, that graph seems to confirm the hoarding thesis – although some drop-off in account value can be seen for ages in the late 70s.

On page 434 (pdf version), the RIR Report4 states: “Superannuation assets have tended to grow in retirement (Chart 5A-12), instead of declining as would be expected if assets were funding retirement”.

Polidano et al 5 similarly state that they find “little evidence that people, on average, run-down superannuation balances after reaching the preservation age (Figure 2).” 

Both comments support the hoarding hypothesis, but closer inspection reveals that the graph pertains only to the average of non-zero-balance accounts.  In other words, those accounts which have been totally withdrawn have been excluded.

This is an example of “survival bias”.  A similar situation can arise when back-testing share investment criteria against past data, if consideration is limited to companies that are still in business. Ignoring those that have failed can be an expensive mistake.

In the present case, the survival bias may or may not matter, depending on one’s purpose; but when the purpose is to establish that retirees hoard their super, it matters a great deal.

Fortunately, Table 1 of Polidano et al 5 provides valuable additional data: average account balances are listed there both for accounts with non-zero balances, and for all accounts – segregated further by gender.  That allows the survival bias effect to be both quantified and eliminated.  It is substantial: roughly 80% of males are shown as having exhausted their accounts by age 80.

To make the impact of the account survival bias easier to see, Fig A below plots the Alife data from Table 1 in Polidano et al 5, for all accounts and for non-zero-balance accounts. 

A discussion about how fast, if at all, people consume their superannuation in retirement must include all accounts to be meaningful.

As shown in the two solid all-accounts curves in Fig A – blue for males and orange for females – there is a strong and steady fall-off in the average all-accounts value throughout retirement, at least to the early 80s, by which time most or all of the average balance has gone.


The notion that retirees, averaged across the population, hoard their super is thus contradicted by the facts.

That is an important conclusion when considering superannuation policy.

[1]John Kehoe and Michael Roddan, “ ’Ever more’ super gets hoarded: Hume”, The Australian Financial Review, 22 January 2021, pages 1,2

[2] Jane Hume, “Safety nets let frugal retirees spend savings without a super rise”,  The Australian Financial Review, 22 January 2021, page 35

[3] John Kehoe, “Push for seniors to dig deep into super nest-egg”, The Australian Financial review, 23 January 2021, page 2


[5] Polidano, C. et al., 2020. The ATO Longitudinal Information Files (ALife): A New Resource for Retirement Policy Research, Working Paper 2/2020, Canberra: Tax and Transfer Policy Institute, available at .

A final version of that paper, in which some errors are corrected, has been published in The Australian Economic Review, September 2020, vol 53, No 3, pp 429-449.

* Jim Bonham PhD, BSc, Dip Corp Mgt, FRACI is a retired scientist and manager with a professional background which was initially in academic physical chemistry, and subsequently in applied research and development in the paper industry.  He has been running an SMSF since 2003 and has a keen interest in the retirement income system.

25 January 2021


On 28 January 2021 an abridged version of this article was published by SuperGuide (


On 31 January 2021 Senator Jane Hume, Minister for Superannuation was reported as saying, amongst other things, that  “[retirees are] passing away with most of their retirement savings intact” ; see “Hume urges retirees to use super capital, rather than just returns” by Emily Chantiri, Sunday Age,

The Tax Institute Submission | Proportional Indexation of the Personal Transfer Balance Cap

26 August 2020

Senator The Hon Jane Hume
The Assistant Minister for Superannuation, Financial Services and Financial Technology
The Treasury
Langton Cres
Parkes ACT 2600

CC: Robert Jeremenko, Retirement Income Policy Division, The Treasury

By email:

Download PDF

Dear Assistant Minister

Proportional Indexation of the Personal Transfer Balance Cap

The Tax Institute requests the Government to consider the reforms outlined below in relation to Division 294 of the Income Tax Assessment Act 1997 (ITAA 1997) – that is, the transfer balance cap (TBC) provisions.

Division 294 was included as a part of the 2016-17 Federal Budget Superannuation Reforms and introduced a cap on the amount of superannuation benefits an individual could transfer into pension (tax-free retirement) phase. The General TBC was initially set at $1.6 million1 upon commencement from 1 July 2017. The law includes the indexation, in increments of $100,0002, of the TBC in line with movements in the CPI.

Section 294-40 of the ITAA 1997 provides for individuals that have not reached their TBC to access proportional indexation of the TBC in accordance with their unused cap percentage.

The Tax Institute considers that this provision is overly complex and needs to be reformed. In the context of the TBC provisions, The Tax Institute submits that simpler rules would:

  • Assist members and their advisers to better understand and manage the TBC; and
  • Reduce the cost for both industry participants and the ATO in relation to administering these provisions.

The Institute submits that the removal of proportional indexation will achieve both of these objectives. Accordingly, the Institute’s submission is as follows:

  • Abolish proportional indexation of the Personal TBC and adopt standardised indexation in line with movements in the General TBC for individuals that have not maximised their Personal TBC prior to the indexation; or
  • If proportional indexation is retained:
    • it should be limited, or its application better targeted to individuals that are within close proximity of the General TBC; and
    • a permanent relief mechanism should be introduced to allow the Commissioner to disregard inadvertent and small breaches of the Personal TBC and not penalise individuals in cases where the excess transfer balance amount is removed within a specified period of time. There is a precedent that was included in the enabling legislation whereby TBC breaches of less than $100,000 were able to be rectified within 6 months (of the commencement of the legislation) and the breaches would not give rise to the application of notional earnings or an excess transfer balance tax liability3. We suggest adopting this transitional rule permanently (with some modification if required).

The requested reforms have been developed by The Tax Institute’s National Superannuation Technical Committee and are detailed in Annexure A for your consideration.

* * * * *

If you would like to discuss, please contact either me or Tax Counsel, Angie Ananda, on 02 8223 0050.

Yours faithfully,

Peter Godber




Part of the 2016-17 Federal Budget superannuation reforms introduced (from 1 July 2017) a lifetime cap on the amount of accumulated superannuation an individual could transfer into pension (tax-free retirement) phase. The General Transfer Balance Cap (TBC) was initially set at $1.6 million4.

Apart from some specific circumstances prescribed in the law, amounts in excess of the TBC must be withdrawn out of pension phase. The excess amounts could remain either in the accumulation phase with associated earnings taxed at 15 per cent or completely transferred out of the superannuation system. Notably, the General TBC would be indexed and allowed to grow in line with the CPI, in increments of $100,0005.

In addition, a Personal TBC was introduced within the enabling legislation6 which is generally equal to the General TBC applicable at the time when an individual makes their first transfer of capital to a retirement phase income stream.

The ATO administers individual compliance with the TBC via a Transfer Balance Account (TBA).

Proportional indexation

Section 294-40 of the ITAA 1997 provides for proportional indexation of the TBC to allow individuals to benefit from increases in the General TBC where they have not fully expired their entitlement to their Personal TBC.

The structure of proportional indexation, over time, results in individuals having a Personal TBC that differs from the General TBC relative to the time they commence a retirement phase income stream.

Complexity is further added as an individual can only obtain a proportion of each General TBC indexation increase (that is, $100,000), based on their Unused Cap Percentage (UCP).

The UCP applies only if the individual has not fully utilised their Personal TBC and is determined by identifying the individual’s highest TBA balance at the end of a day, at an earlier point in time, and comparing it to their Personal TBC on that day. The UPC is expressed as a percentage and applied against the indexation available for the General TBC. This process is required at each indexation point until such time as the individual utilises 100% of their Cap Space.

Once an individual has used their entire available Cap Space, their Personal TBC is not subject to further indexation, even if they later commute some or all of that pension to bring their TBA balance below their Personal TBC.

This is best illustrated with the following examples (adapted from the Explanatory Memorandum)7:

Example – simple

Danika first commenced an $800,000 retirement phase income stream on 18 November 2017. A Transfer Balance Account is created for her at this time. Her Personal TBC was $1.6m for 2017-18. As Danika had not made any other transfers to her retirement phase account, her highest TBA balance is $800,000. She has used 50% of her $1.6m Personal TBC.

Assuming the General TBC is indexed to $1.7m in 2020-21, Danika’s Personal TBC is increased proportionally to $1.65m. That is, Danika’s Personal TBC is only increased by the UCP of 50% of the $100,000 indexation increase to the General TBC. As such, Danika can now transfer a further $850,000 to commence a retirement phase income stream without breaching her Personal TBC. Notably, earnings on the original $800,000 were not taken into account in working out how much of the unused cap was available.

Example – complex

On 1 October 2017, Nina commenced a retirement phase income stream with a value of $1.2m. On 1 January 2018, Nina partially commuted $400,000 from this income stream to buy an investment property.

Nina’s TBA balance on 1 October 2017 was $1.2m and on 1 January 2018, it was $800,000 (after the $400,000 partial commutation).

In 2020-21, assume the General TBC is indexed to $1.7m. To work out the amount by which Nina’s Personal TBC is indexed, it is necessary to identify the day on which her TBA balance was at its highest. In this case, the highest balance was $1.2m on 1 October 2017. Nina’s Personal TBC on that date was $1.6m. Therefore, as she has used up 75% of her Personal TBC, Nina’s UCP on 1 October 2017 is 25%.

To work out how much her Personal TBC is to be indexed, Nina’s UCP is applied to the amount by which the General TBC has indexed (i.e. $100,000). Therefore, Nina’s Personal TBC in 2020-21 is $1.625m.

In 2022-23, assume the General TBC is indexed to $1.8m (i.e. by another $100,000). As Nina has not transferred any further amount into the retirement phase, her UCP remains at 25%. Her Personal TBC is now

$1.65m (i.e. 25% of the further indexation increase of $100,000). In this year, Nina decides to transfer the maximum amount she can into the retirement phase.

This will be her Personal TBC for the 2022-23 year ($1.65m) less her TBA balance of $800,000. This means Nina can transfer another $850,000 into the retirement phase without exceeding her Personal TBC.

Once Nina has used all of her available Cap Space, her Personal TBC will not be subject to further indexation. Notably, this is the case even if Nina later partially commutes some of her retirement phase income stream and makes her TBA balance fall below her Personal TBC.

Range of Potential Personal TBC

Members (A to D) Cap Space utilised in 2018 – 2021 (assuming indexation in 2021).

 Personal TBC (Start)YearCap Space AccessedTotal UsedUnused Cap %8Personal TBC (End)Year
C$1,600,0002018$ 700,000$ 700,00057%$1,657,0002021

Members (A to D) Cap Space utilised in 2022 – 2025 (assuming further indexation in 2025).

 Personal TBC (Start)YearCap Space AccessedTotal UsedUnused Cap %Personal TBC (End)Personal TBC Year
B$1,638,0002022$ 500,000$1,500,0009%$1,647,0002025
C$1,657,0002022$ 900,000$1,600,0004%$1,661,0002025

As illustrated above, the amount of proportional indexation largely depends on an individual’s UCP, each time there is indexation of the General TBC.

The result of this process is an individual’s Personal TBC is likely to be unique to themselves if they are in the category where they do not utilise their full TBC at the point of commencing a retirement phase income stream.

The burden is therefore likely to fall to individuals with smaller superannuation balances.

Current Issues and Challenges

Overly complex calculation and too widely targeted

The above examples demonstrate the complexity of the proportional indexation calculation which, over time, will require very effective record management systems to track and be effective in ensuring individuals are not subject to inadvertent breaches of their TBC.

According to the Explanatory Memorandum9 (EM) to the enabling legislation, the introduction of the TBC was to better target superannuation concessions thereby ensuring the system’s sustainability over time. The EM estimated the introduction of the TBC would affect less than one per cent of Australians with a superannuation interest. If this estimate is accurate, then arguably this particular calculation which is carried out on the full population (in retirement phase) represents a very inefficient and costly tax administration for what was meant to only have a one per cent application.

Design not hitting target

By its very nature, proportional indexation essentially rewards and benefits those who are able to defer the commencement of their retirement phase income stream to a later date. As a consequence, it may actually encourage those who are ready to retire (with amounts in excess of the General TBC) to defer and leave their superannuation balance in the accumulation phase until indexation does occur. Such behaviour would appear contrary to the intent and objective of proportional indexation.

Inconsistent treatment with other superannuation caps

The proportional indexation of the Personal TBC is unprecedented when compared with other indexation that takes place under the tax law. For instance, the Low Rate Cap Amount10 which represents a lifetime cap on the tax-free level of superannuation lump sums an individual can receive in their lifetime is indexed in line with AWOTE, in increments of $5,000. There is no proportional reduction based on how much of the previous Low Rate Cap Amount has remained unused.

Too many rates, caps and thresholds

It has taken industry participants and the ATO considerable time and effort to educate individuals on the concept of the TBA (including debits and credits) and the consequences of breaching their Personal TBC. At times, there has even been some confusion with the Total Superannuation Balance (TSB) which has also been pegged to the General TBC for the purposes of assessing eligibility to the following superannuation concessions:

  • Non-concessional contributions caps (NCC) and associated bring forward caps;
  • Government co-contributions;
  • Tax offset for spouse contributions; and
  • Segregated method to determine their earnings tax exemption (SMSFs and small APRA funds only).

We submit that there was a logical nexus and a familiarity developed with the $1.6 million between the General TBC and the TSB (at least for the above mentioned superannuation concessions). However, as proportional indexation comes into effect with an individual potentially having a Personal TBC not equalling the General TBC, The Tax Institute is concerned that the Personal TBC will further exacerbate what is becoming a very complex suite of rates, caps and thresholds operating within the superannuation system.

A lack of consistency has been applied historically across the various caps and eligibility criterion for superannuation concessions. Some caps are indexed based on either CPI or AWOTE whereas others are referenced to a specific date and/or data set.

The non-concessional bring forward non-concessional contributions cap (that allows some individuals to use 3 years of the non-concessional contribution cap in a single year), provides an example of the anomalies apparent in the reference points for access to superannuation concessions.

The bring-forward non-concessional contributions cap is assessed based on an individual’s TSB, with thresholds referencing both the General TBC and the General NCC.

Source: Paragraph 5.47 of the Treasury Laws Amendment (Fair and Sustainable Superannuation) Bill 2016 – Explanatory Memorandum.

Specifically, the references to $1.4 million, $1.5 million and $1.6 million reflect a calculation that takes a multiple of the NCC (originally set at $100,000) and subtracts it from the General TBC11.

Notably the table above gets very complicated and may produce anomalous eligibility thresholds given the indexation of the General TBC is based on CPI and the General NCC on AWOTE. This problem is further exacerbated if, in a given income year, the NCC is indexed but the General TBC is not (due to the higher $100,000 incremental hurdle).

Access difficulties to Personal TBC

Currently, financial planners and advisers have difficulties accessing clients’ TBA and TSB balances given the information is only accessible via the ATO and restricted to the individual and their tax agents.

Without this information, it is extremely difficult for advisers to provide appropriate financial advice which will not cause their clients to inadvertently breach their caps. Unfortunately, we foresee this problem only getting worse if and when proportional indexation comes into effect with an individual’s Personal TBC diverging from the General TBC12. This ultimately affects the quality of the advice provided and may lead to errors, giving rise to complaints and the need to remediate.

Reliance on ATO

Of particular concern is that the Personal TBC determination rests solely with the ATO, who in turn must rely on all Fund providers to report all relevant member’s transactions accurately and in a timely fashion. Any errors or omissions in the reporting may lead to subsequent re-reporting which may trigger a re-calculation of the Personal TBC. However, the individual and/or their financial adviser may have acted and relied upon the ATO’s calculation.

Notably in a recent AAT case13, a taxpayer submitted that they were led into error by misleading content on the ATO website when managing the excess over their Personal TBC. However, the Tribunal found in favour of the ATO and also confirmed that the Tribunal had no jurisdiction to address any application against the ATO even if its website could be said to be misleading This was the case, notwithstanding there was acknowledgement that the taxpayer had relied on the ATO information in good faith and made an honest mistake.

Furthermore, there appears to be no rights to object to any inadvertent breach of the individual’s Personal TBC in such a situation and is a real concern given the fact that the Commissioner does not explicitly have any discretionary powers to disregard an excess transfer balance (even if information obtained from the ATO contributed to the breach).

Culpability for errors or omissions unclear

As alluded to in the previous issue, it may also be unclear as to who is culpable for an inadvertent breach of the cap as a consequence of any error or omission in member reporting. With such a large superannuation system and the volume of transactions that occur, one can expect reporting errors

or omissions to arise from time to time that require re-reporting to the ATO. To the extent re-reporting causes a subsequent breach of the cap and/or denial of any related superannuation concessions, it is unclear who is culpable for that breach. This is particularly the case if the individual has multiple accounts with various Fund providers and there are errors or omissions experienced by more than one Fund provider.

Severe tax consequences for excess transfer balances not rectified

The breaching of the TBC may often lead to an individual disputing their excess transfer balance with the ATO and refusing to commute the excess. Individuals may give explicit instructions to their Fund not to process a commutation authority issued by the ATO. However, Trustees have no choice but to act and process a commutation authority that it receives14. Failure to do so within the prescribed 60 days will result in the income stream ceasing to be a retirement phase income stream15.

This in turn results in the income stream ceasing to qualify for the earnings tax exemption16 with the eligibility deemed to have been lost from the start of the income year in which the commutation authority was not complied with. Notably this tax outcome poses an administrative challenge particularly for large funds that operate retail Master Trusts and offer their members superannuation interests in unitised asset pools. Specifically, it becomes extremely difficult to even determine what portion of the income and gains from the unitised (pension) asset pools is attributable to that particular member. Furthermore, the Trustee would also need to assess the tax character of that income and gains, plus any applicable tax offsets in order to determine the effective earnings tax rate that ought to have been applied to the deemed assessable income.

More importantly, the Trustee’s compliance with a commutation authority may give rise to a complaint by the member for not complying with their explicit instruction or request not to process.

Outcomes and Conclusion

The current superannuation system is overly complex and costly to administer. There is merit in making it much simpler for all stakeholders.

The Tax Institute submits that simpler rules would assist in:

  • Members and their advisers better understanding and managing the TBC; and
  • Reducing the cost for both industry participants and the ATO to administer these rules.

The removal of proportional indexation specifically achieves both these objectives.

As with other existing superannuation caps, adopting the General TBC will ensure every individual has the same TBC regardless of when they entered the retirement phase. This will provide certainty to individuals and their advisers to better plan for their retirement, without having to rely on gaining access to ATO information.

The call for permanent relief for inadvertent and small breaches is aimed at red tape reduction and thereby increasing the overall efficiency of the superannuation system. At present, it can take up to 120 days before an excess transfer balance amount is withdrawn from retirement phase. Any such breaches tend to involve every party within the system and is therefore a very costly and time- consuming exercise. We submit this particular option strikes a good balance between maintaining the integrity and efficiency of the superannuation system. It should also obviate any breaches which may result from any misreporting and re-reporting of member transactions by Fund providers and reduce the number of disputes and complaints raised by the individual against the Trustee and/or the ATO.

Whilst the proportional indexation methodology is designed to target higher member balances, its universal application is more likely to impact superannuation funds with lower balance members. The Tax Institute considers it to be undesirable to significantly increase the complexity inherent in a system that has limited application against the object of the measure. There is still time to remove this provision from Division 294 before the first indexation event occurs and we urge the Government to carefully consider this submission.


1  Subsections 294-35(3) of the ITAA 1997.

Subsection 960-285(7) of the ITAA 1997.

3 Section 294-30 of the IT(TP) Act 1997

4  Subsections 294-35(3) of the ITAA 1997.

Subsection 960-285(7) of the ITAA 1997.

6 Treasury Laws Amendment (Fair and Sustainable Superannuation) Act 2016 [No. 81 of 2016].

7 Treasury Laws Amendment (Fair and Sustainable Superannuation) Bill 2016 – Explanatory Memorandum.

8 Unused cap percentage effectively is rounded up to the nearest per cent as a result of subsection 294-40(2)(c) of the ITAA 1997.

9 Paragraph 3.373 of the Treasury Laws Amendment (Fair and Sustainable Superannuation) Bill 2016 – Explanatory Memorandum.

10 Section 307-345 of the ITAA 1997.

11 Subsections 292-85(2), (3), (4) and (5) of the ITAA 1997.

12 Indexation of the General TBC will first occur (in the following income year) if and when the December quarter CPI figure reaches 116.9.

13 Lacey and Commissioner of Taxation (Taxation) [2019] AATA 4246.

14 Section 136-80 in Schedule 1 to the TAA 1953.

15 Subsection 307-80(4) of the ITAA 1997.

16 Subdivision 295-F of the ITAA 1997.

Click here for The Tax Institute’s original submission published 28 Aug 2020 by THE TAX INSTITUTE


The early release of superannuation: the financial consequences

16 August 2020

Jim Bonham

Early release of superannuation

Limited early release of superannuation has been a part of the government’s support to people suffering financial stress caused by the COVID-19 pandemic.  This has been welcomed by some but strongly opposed by others because it is seen as a corruption of compulsory saving and disruptive to super funds. 

The financial implications

This paper takes a quantitative look at the financial implications of a $10k early release, for three hypothetical individuals Continue reading

Consequences of increasing the superannuation guarantee rate

12 March 2020

Jim Bonham

1       Introduction

Between now and 2025 the compulsory “superannuation guarantee” (SG) contribution to superannuation is legislated to increase in steps from 9.5% of gross income to 12%.

This move is being opposed by some people, particularly the Grattan Institute (see, amplified by op-eds in the press.

Unfortunately, formal “think tank” and academic reports tend to be inaccessible to the average reader. Calculations may be opaque; and journalists often manage to make the impending increase look quite complicated and confusing.

It does not have to be so.  This short note explores the immediate consequences of the legislated increase in the SG rate from 9.5% to 12% and introduces an alternative proposal to increase the SG rate to 10%, or even leave it unchanged, and drop the contribution tax entirely.

Click here to download PDF version

2       The issues

The table below lists what I perceive to be the main points of concern, and a brief comment on each.  This is provided for context, not as a detailed commentary on any specific position.

Perceived problem Comment
(a)Retirees already have enough money so there is no need to beef up super. Depending on investment returns, current SG contributions will only provide an initial retirement income of 14% to 25%, or so, of final employment income (depending on investment choices).
(b)Increasing the SG rate will depress incomes. The government has reportedly asked ANU to advise specifically on this issue.  Gross incomes will fall by 2.23% (given assumptions detailed in the text), but there is an alternative.
(c)Increasing retirees’ assets will disproportionately reduce their age pension entitlement This reflects a problem with the structure of the age pension, not super.  In any event it will take a decade or so to become significant, leaving plenty of time to fix the structural issue.
(d)The budget can’t afford the cost The cost to the government of the planned increases, per employee, is equivalent to about 0.5% of their gross income – the equivalent of a modest tax cut.

Each of these issues is discussed in more detail below.  The intention is not to provide detailed rebuttals of any specific point of view, but rather to add context to the upcoming increase, and to suggest an alternative approach.

3       How much does the SG provide?

It can be a daunting task to work out how much superannuation one will have in retirement, what its real value will be and how that might relate to one’s income needs.

Fortunately, help is available from on-line calculators such as the excellent one provided by ASIC   ( which also provides detailed actuarially determined estimates of long term investment returns, fees and earnings for several common investment styles, as well as estimates of inflation and wages growth (as reflected in rising living standards).

A common measure, used by the OECD for example, for assessing retirement funding systems is the replacement ratio, which is the initial income in retirement divided by the final employment income.  (Obviously, this only makes sense for someone who remains in steady employment up to retirement and doesn’t apply to those with a more fractured employment history). 

It is generally accepted that a replacement ratio of 70% represents good practice and, in the absence of better information, it seems to “feel” about right.

Fig 1 shows the replacement ratio expected just from current SG contributions and their compounded investment returns, assuming

  • SG rate is 9.5%, taxed at 15%
  • Wages growth is 3.2%
  • Length of employment is 45 years
  • On retirement, superannuation is converted to an allocated pension from which 5% per annum is drawn as income in the early years.
  • Complications such as contribution caps are ignored.

The simple conclusion from Fig 1 is that, however the superannuation account is invested, the SG contributions alone will not provide anything like a 70% replacement ratio. 

Most people will need to supplement their SG contributions substantially with further voluntary superannuation contributions, the age pension, or other investments outside superannuation, in order to live at a level anything like what they were used to.

There is thus a lot of scope to increase the SG contributions, which goes a long way toward refuting Issue 2(a). 

4       How will the SG rate increase affect pre-retirement incomes?

To keep things simple, we’ll exclude from consideration those who are on a very low income, those who are subject to Division 293 tax (incomes over $250,000) and those who are already at or near the concessional contribution cap. We’ll also assume that all income derives from employment.  Finally, in the interests of simplicity, we’ll assume that the increase takes place in one step rather than being staged over several years.

It is highly likely that employer bargaining power is such that increasing the SG contribution rate will not affect total income packages (i.e. gross income plus SG contributions).  The calculations below assume that this is so – it is a key assumption of this paper.

Note, however, that in real life salary negotiations are not necessarily cut and dried.  So, a push to restore a previous total package value might not be immediate but be buried in subsequent increments, or it might manifest as additional pressure in future negotiations.

To be able to work this through mathematically, however, we make the simplifying assumption that incomes will adjust immediately.

It’s also important to keep in mind that while a mathematical model produces precise, neat and tidy results, these are only as good as the initial assumptions – the real world is much messier.  The important function of an analysis such as this one is not so much to produce precise predictions, but rather to lay bare the way in which key variables (in this instance: income, income tax, SG rate and contribution tax) all interact.  Better understanding should lead to better decision making.

With these cautions in mind, let’s move on.  Some straightforward arithmetic, illustrated in Table 1, shows that the immediate consequences of increasing the SG rate will be as follows:

  • SG contributions will rise by 23.5%
  • Gross incomes will fall by 2.23%
  • Net SG contributions will rise by 23.5%, corresponding to 1.90% of initial gross income
  • Government income tax receipts will fall by an amount which depends on income.

Table 1 shows how the numbers work out for an initial gross income of $100,000:

Table 1

  • The top line reflects the assumption of no change in the total income package
  • so, there must be a drop in gross income (2nd line)
  • and therefore, the government’s income tax receipts will fall (3rd line)
  • as will the individual’s net income (4th line).
  • The SG contribution goes up by 23.5% (5th line).
  • The government claws back an extra 23.5% contributions tax (6th line)
  • leaving a net contribution which is also 23.5% higher than before (7th line).

In summary, the superannuation account of the individual currently earning $100,000 nets an extra $1,897 per year.  In the short term, this is a zero-sum game (the savings have to be paid for): $1,295 is provided by the individual (reduced net income offset by lower income tax) and $603 is provided by the government (reduced income tax receipts offset by higher contribution tax).

In other words, the individual saves more, and the government also contributes.

Although this is a zero-sum game in the short term, that is not the case in the long term.  Superannuation savings provide a massive investment resource for the nation, and a more financially secure retiree population will require less government support.  There is a large net benefit to the nation from supporting and incentivising long-term saving.

Although Table 1 is worked for $100,000 initial gross income, the same 2.23% fall in gross income and 23.5% increase in net SG contribution occurs for any other initial income. 

The boost to SG contributions then flows through to provide a valuable 23.5% increase in the value of SG contributions and their accumulated investment returns at any time through to retirement, and consequently the same percentage increase in both earnings and earnings tax.

A partial response to Issue 2(b), therefore, is: yes, the planned increase in SG rate will depress gross incomes by 2.23%. 

5       How is the cost shared between government and individual?

Before the increase, the net SG contribution is 8.075%, after allowing for the 15% contributions tax, so a 23.5% increase in that corresponds to 1.90% of the initial gross income.  That 1.90% must be paid for, and as we have seen the cost is shared between the individual and the government.

Fig 2 shows the split for a wide range of initial incomes, the structure in the graphs reflecting the complicated structure of income tax rates.

The cost to government averages about 0.5% of gross income (for incomes between $50,000 and $180,000) and that helps put Issue 2(d) in context: it is of similar magnitude to a modest income tax reduction.

The cost should not be onerous for the government and could be funded by cancelling or reducing less important programs, or by working with greater efficiency (meant literally, not as a euphemism for sacking people which only pushes costs back to individuals).

Incidentally, the cost to government is sometimes compared to the cost of fixing other significant problems, such as Newstart.  This is the wrong way to evaluate the priority of a project: it should be compared to the least important project, which can most easily be dropped, not to other important projects.

6       How quickly will the effects be felt?

The effects on net income and taxes discussed above will be immediate, but the impact on retirement income will take time to evolve – about a decade for effects to become noticeable and four decades for the complete benefit. 

Superannuation operates over the very long term, which means the sooner problems are fixed the better. The current financial climate does not justify delay.

Two issues which will eventually emerge but will be insignificant for the first couple of decades are:

  • Earnings taxes on superannuation investments will increase by 23.5%.
  • Age pension entitlements will decrease for people on low-to-moderate-incomes.

Both benefit the government.  However, the age pension needs significant modification to correct other fundamental problems:

In short, there is plenty of time and opportunity to make sure that Issue 2(c) will not become a problem.

7       An alternative proposal

The above calculations highlight something quite bizarre about concessional superannuation contributions: the superannuation guarantee compels people to save for their retirement, but the contributions tax immediately undermines that – now you see it, now you don’t!

The system would be much neater and easier to understand if the contributions tax were abolished.

That would also make voluntary concessional contributions (up to the cap) more attractive, thus encouraging more saving, but let’s look more closely at what it would mean for compulsory SG contributions.

As we have seen the upcoming increase in SG rate will increase compulsory net contributions to superannuation by 23.5%, given the assumption that gross incomes are unaffected, so let’s take that as an objective and see how it would be achieved without the contributions tax.

The answer is that the SG rate then only needs to be increased to 10%, rather than 12%.  Net contributions will increase by 23.3% which is almost identical to 23.5%, but the split in cost between the government and individual is changed significantly.

Table 2 shows the detailed figures for a gross income of $100,000:

Table 2

and Fig 3 shows the split in costs between government and individual for a range of gross incomes:

From the government’s point of view, this proposal is more expensive by about 1% of gross income than the increase currently legislated – it is still the equivalent of a modest tax cut across the board.  Staging this change over several years would further reduce the budgetary shock. 

From the individual’s point of view, the cost has reduced to about a quarter of a percent of gross income, which in normal times most people would not notice.

However, these are not normal times: the aftermath of this summer’s fires and the developing coronavirus scenario mean that many people are or will be under severe financial pressure.  The government is currently working to provide significant stimulus in response.  The government has also reportedly asked ANU to advise the government on whether the upcoming increase will affect incomes.

As shown above, they certainly will do so, and it is tempting to see this as a strong argument against making any increase at all.

However, it is easy in times of crisis to neglect long term issues, banking problems for future generations.

The government could find it attractive, therefore, to demonstrate a continued commitment to long term saving by dropping the contributions tax, while leaving the SG rate at 9.5% so there is no additional cost for individuals. 

If that approach is followed, the boost to net SG contributions will be 17.6% instead of 23.5% – a little less, but still a sizeable improvement for the long term. 

To see what that would mean, we return to Fig 1 and consider someone who chooses a “Balanced” investment option for their super.  Using ASIC’s figures, that would give a replacement ratio of 20% under the current rules for the assets derived from SG contributions. 

The initial retirement income would thus be 24.7% of final employment income under the current plan (23.5% improvement), or 23.5% of final employment income if the SG rate remains at 9.5% and the contributions tax is dropped (17.6% improvement).  Either way, it is a significant improvement, while still leaving a considerable gap to be filled by extra voluntary saving, or the age pension, depending on the retiree’s circumstances.

8       About the author

Jim Bonham (BSc (Sydney), PhD (Qld), Dip Corp Mgt, FRACI) is a retired scientist (physical chemistry). His career spanned 7 years as an academic followed by 25 years in the pulp and paper industry, where he managed scientific research and the development of new products and processes. He has been retired for 14 years and has run an SMSF for 17 years.  He will not be affected by any change to the superannuation guarantee.


Save Our Super submission: Consumer Advocacy Body for Superannuation

Click here for the PDF document

13 January 2020

The Manager
Retirement Income Policy Division
Langton Cres
Parkes ACT 2600

Save Our Super submission:  Consumer Advocacy Body for Superannuation

Dear Sir/Madam

Save Our Super has recently prepared an extensive Submission to the Retirement Income Review dealing in part with the many ‘consumer’ issues triggered by the structure of retirement income policy and the frequent and complex legislative change to that policy.  

That Submission was lodged with the Review’s Treasury Secretariat on 10 January 2020 and a copy is attached to the e-mail forwarding this letter. It serves as an example of the analysis of super and retirement policy and of the advocacy that superannuation fund members, both savers and retirees, can contribute.  Its four authors’ backgrounds show the wide range of experience that can be useful in the consumer advocacy role. 

We note both the policy and the advocacy consultations are running simultaneously, exemplifying the pressures Government legislative activity places on meaningful consumer input.  Consumer representation is necessarily more reliant on volunteer and part-time contributions than the work of industry and union lobbyists and the juggernaut of government legislative and administrative initiatives.

Given the breadth, complexity and fundamentally important nature of the issues raised for the Retirement Income Review by its Consultation Paper, we have prioritised our submission to that Review over the issues raised by the idea of a Consumer Advocacy Body.  This letter serves as a brief submission and as a ‘place holder’ for Save Our Super’s interest in the consumer advocacy issues.

The idea of a consumer advocacy body is worthwhile in trying to improve member information, engagement and voice in superannuation and in the formation of better, more stable and more trustworthy retirement income policy.  It should help government to understand the perspectives of superannuation members.

Save Our Super was formed from our frustration at the evolution of superannuation and broader retirement income policies.  We contributed as best we could to the rushed and heavily constrained Government consultations on, and Parliamentary Committee inquiries into, the complex retirement income policy changes that took effect in 2017. One example of our inputs is .

For the consultations on the Consumer Advocacy Body for Superannuation, we limit our comments here to point 1 on the Consultation’s brief web page, :

Functions and outcomes: What core functions and outcomes do you consider could be delivered by the advocacy body? What additional functions and outcomes could also be considered? What functions would the advocacy body provide that are not currently available?”

Key roles

  • Consult with superannuation fund members on their concerns, including issues of legal and regulatory complexity, frequent legislative change and legislative risk which has become destructive of trust in superannuation and its rule-making.
  • Commission or perform research arising from consultations and reporting of member concerns.
  • Tap perspectives of all superannuation users, whether young, mid-career, or near-retirement savers, as well as of part- or fully self-funded retirees.
  • Publish reporting of savers’ concerns to Government, at least twice-yearly and in advance of annual budget cycles.
  • Contribute an impact statement – as envisaged in the lapsed Superannuation (Objectives) Bill – of the effects of changes to any legislation (not just super legislation) on retirement income (interpreted broadly to include the assets, net income and general well-being of retirees, now and in the future).

The advocacy body should:

  • take a long-term view, and could be made the authority to administer, review or critique the essential modelling referred to in Save Our Super’s submission to the Retirement Income Review. Ideally, the  Consumer Advocacy Body should have the freedom to commission Treasury to conduct such modelling, and/or to use any other capable body.
  • give appropriate representation and support to SMSFs, and be prepared to advocate for them against the interests of large APRA-regulated funds when necessary.  
  • advocate specifically for the very large number of people with quite small superannuation accounts, when their interests are different from those of people with relatively large balances.

The biggest risk to the advocacy body in our view is that it would over time be hijacked by special interest groups, or hobbled by its terms of reference.  Careful thought in its establishment, key staffing choices and strong political support would be helpful to protect against these risks. 

Membership issues

  • Membership of the Body should be part-time, funded essentially per diem and with cost reimbursement only for participation in the information gathering and consumer advocacy processes.  A small part-time secretariat could be provided from resources in, say, PM&C or Treasury. 
  • Membership opportunities should be advertised.
  • Membership of the Body should be strictly limited to individuals or entities that exist purely to advocate for the interests of superannuation fund members. (This would include any cooperative representation of Self-Managed Superannuation Funds.) We would counsel against allowing membership to industry entities which might purport  to advocate on behalf of their superannuation fund members, but might also inject perspectives that favour their own commercial interests.
  • Membership should include individuals with membership in (on the one hand) commercial or industry super funds and (on the other hand) Self-Managed Superannuation Funds.  We see no need to ensure equal representation of commercial and industry funds, though we would be wary if representation was only of those in industry funds or only commercial funds.
  • We offer no view at this stage on whether the Superannuation Consumers Centre would be a useful anchor for a new role, but we would suggest avoiding duplication.

Functions not currently available

The consultation asks what functions the Consumer Advocacy Body for Superannuation could perform that are not presently being performed.  SOS’s submission to the Retirement Income Review and earlier submissions on the changes to retirement income policy that took effect in 2017 shows the range of superannuation members’ advocacy concerns that are not at present being met.  

Prior attempts to establish consultation arrangements for superannuation members appear to us to have focussed mostly on the disengagement and limited financial literacy of some superannuation fund members.  Correctives to those concerns have heretofore looked to financial literacy education and better access to higher quality financial advice. Clearly such measures have their place.

But in the view of Save Our Super, these problems arise in larger part from the complexity and rapid change of superannuation and Age Pension laws, and in the nature of the Superannuation Guarantee Charge. Nothing predicts disengagement by customers and underperformance and overcharging by suppliers more assuredly than government compulsion to consume a product that would not otherwise be bought because it is too complex to understand, too often changed and widely distrusted.

There needs to be more consumer policy advocacy aimed at getting the policies right, simple, clear and stable, as was attempted in the 2006 – 2007 Simplified Super reforms.

Other issues

In the time available, we offer no views on questions 2,3 and 4, which are more for government administrators.

Yours faithfully

Jack Hammond, QC

Founder, Save Our Super

Click here for the PDF document

Submission by Save Our Super in response to Retirement Income Review Consultation Paper – November 2019

Submission by Save Our Super in response to Retirement Income Review Consultation Paper – November 2019

by Terrence O’Brien, Jack Hammond, Jim Bonham and Sean Corbett

10 January 2020

Click here for the full PDF document


  1. The Review’s Terms of Reference seek a fact base on how the retirement income system is working.  This is a vital quest.  Such information, founded on publication of long-term modelling extending over the decades over which policy has its cumulative effect, has disappeared over the last decade.
  2. Not coincidentally, retirement income policy has suffered from recent failures to set clear objectives in a long-term framework of rising personal incomes, demographic ageing, lengthening life expectancy at retirement age, weak overall national saving, low household and company saving and a persistent tendency to government dissaving.
  3. A new statement of retirement income policy objectives should be:
    • to facilitate rising real retirement incomes for all;
    • to encourage higher savings in superannuation so progressively more of the age-qualified can self-fund retirement at higher living standards than provided by the Age Pension;
    • to thus reduce the proportion of the age-qualified receiving the Age Pension, improving its sustainability as a safety net and reducing its tax burden on the diminishing proportion of the population of working age; and
    • to contribute in net terms to raising national saving, as lifetime saving for self-funded retirement progressively displaces tax-funded recurrent expenditures on the Age Pension.
  1. With the actuarial value of the Age Pension to a homeowning couple now well over $1 million, self-funding a higher retirement living standard than the Age Pension will require large saving balances at retirement.  It is unclear that political parties accept this.  It seems to Save Our Super that politicians champion the objective of more self-funded retirees and fewer dependent on the Age Pension but seem dubious about allowing the means to that objective.
  1. Save Our Super highlights fragmentary evidence from the private sector suggesting retirement income policies to 2017 were generating a surprisingly strong growth in self-funded retirement, reducing spending on the Age Pension as a share of GDP, and (prima facie) raising living standards in retirement (Table 1). (Anyone who becomes a self-funded retiree can be assumed to be better off than if they had rearranged their affairs to receive the Age Pension.)  Sustainability of the retirement system for both retirees and working age taxpayers funding the Age Pension seemed to be strengthening. These apparent trends are little known, have not been officially explained, and deserve the Review’s close attention in establishing a fact base.
  1. Retirement policy should be evaluated in a social cost-benefit framework, in which the benefits include any contraction over time in the proportion of the age-eligible receiving the Age Pension, any corresponding rise in the proportion enjoying a higher self-funded retirement living standard of their choice, and any rise in net national savings; while
    the costs include a realistic estimate of any superannuation ‘tax expenditures’ (this often used term is placed in quotes because it is generally misleading – see subsequent discussion) that reduce the direct expenditures on the Age Pension. Such a framework was developed and applied in the 1990s but has since fallen into disuse.
  1. Policy changes that took effect in 2017 have suffered from a lack of enumeration of the long-term net economic and
    fiscal impacts on retirement income trends. They also damaged confidence in the retirement rules, and the rules for changing those rules. Extraordinarily, many people trying to manage their retirement have found legislative risk in recent years to be a greater problem than investment risk. Save Our Super believes the Government should re-commit to the grandfathering practices of the preceding quarter century to rebuild the confidence essential for long-term saving under
    the restrictions of the superannuation system.
  1. Views on whether retirement policy is fair and sustainable differ widely, in large part because the only official analysis that has been sustained is so-called ‘tax expenditure’ estimates using a subjective hypothetical ‘comprehensive income tax’ benchmark that has never had democratic support.
  1. This prevailing ‘tax expenditure’ measure is unfit for purpose. It is conceptually indefensible; it produces wildly unrealistic
    estimates of hypothetical revenue forgone from superannuation (now said to be $37 billion for 2018-19 and rising); and it presents an imaginary gross cost outside the sensible cost-benefit framework used in the past.  It also presents (including, regrettably, in the Review’s Consultation Paper) an imaginary one-off effect as though it could be a
    recurrent flow similar to the actual recurrent expenditures on the Age Pension.
  1. An alternative Treasury superannuation ‘tax expenditure’ estimate, more defensible because it has the desirable characteristic of not discriminating against saving or supressing work effort, is based on an expenditure tax benchmark. It estimates annual revenue forgone of $7 billion, steady over time, not $37 billion rising strongly. 
  1. Additional to the four evaluative criteria proposed in the Consultation Paper, Save Our Super recommends a fifth: personal choice and accountability. Over the 70-year horizon of individuals’ commitments to retirement saving, personal circumstances differ widely.  As saving rates rise, encouraging substantial individual choice of saving profiles to achieve preferred retirement living standards is desirable.
  1. We also restate a core proposition perhaps unusual to the modern ear: personal saving is good. The consumption that is forgone in order to save is not just money; it is real resources that are made available to others with higher immediate demands for consumption or investment. Saving and the investment it finances are the foundation for rising living standards. Those concerned at the possibility of inequality arising from more saving should address the issue directly by presenting arguments for more redistribution, not by hobbling saving.
  1. While retirement income ‘adequacy’ is a sensible criterion for considering the Age Pension, ‘adequacy’ makes no sense as a policy guide to either compulsory or voluntary superannuation contributions towards self-funded retirement. Adequacy of self-funded retirement income is properly a matter for individuals’ preferences and saving choices.
  1. The task for superannuation policy in the broader retirement income structure is not to achieve some centrally-approved
    ‘adequate’ self-funded retirement income, however prescribed. It is to roughly offset the government’s systemic disincentives to saving from welfare spending and income taxing. Once government has struck a reasonable, stable and sustainable tax structure from that perspective, citizens should be entitled to save what they like, at any stage of life.
  1. The Super Guarantee Charge’s optimum future level is a matter for practical marginal analysis rather than ideology. Would raising it by a percentage point add more to benefits (higher savings balances at retirement for self-funded retirees) than to costs (e.g. reduced incomes over a working lifetime, more burden on young workers, or on poor workers who may not save enough to retire on more than the Age Pension)?
  1. The coherence of the Age Pension and superannuation arrangements is less than ideal. Very high effective marginal tax rates on saving arise from the increased Age Pension assets test taper rate, with the result that many retirees are trapped in a retirement strategy built on a substantial part Age Pension.  Save Our Super also identifies six problem areas where inconsistent indexation practices of superannuation and Age Pension parameters compound through time to reduce super savings and retirement benefits relative to average earnings. These problems reduce confidence in the stability of the system and should be fixed.
  1. Our analysis points to policy choices that would give more Australians ‘skin in the game’ of patient saving and long term investing for a well performing Australian economy.  Those policies would yield rising living standards for all, both those of working age and retirees.  Such policies would give more personal choice over the lifetime profile of saving and retirement living standards; fewer cases where compulsory savings violate individual needs, and more engaged personal oversight of a more competitive and efficient superannuation industry.

About the authors

Terrence O’Brien is an honours graduate in economics from the University of Queensland, and has a master of economics from the Australian National University. He worked from the early 1970s in many areas of the Treasury, including taxation
policy, fiscal policy and international economic issues. His senior positions have also included several years in the Office of National Assessments, as resident economic representative of Australia at the Organisation for Economic Cooperation and Development, as Alternate Executive Director on the Boards of the World Bank Group, and as First Assistant Commissioner at the Productivity Commission.

Jack Hammond LLB (Hons), QC is Save Our Super’s founder. He was a Victorian barrister for more than three decades.
He is now retired from the Victorian Bar. Prior to becoming a barrister, he was an Adviser to Prime Minister Malcolm Fraser, and an Associate to Justice Brennan, then of the Federal Court of Australia. Before that he served as a Councillor on the Malvern City Council (now Stonnington City Council) in Melbourne.

Jim Bonham (BSc (Sydney), PhD (Qld), Dip Corp Mgt, FRACI) is a retired scientist (physical chemistry).  His career spanned 7 years as an academic followed by 25 years in the pulp and paper industry, where he managed scientific research and the development of new products and processes.  He has been retired for 14 years has run an SMSF for 17 years.

Sean Corbett has over 25 years’ experience in the superannuation industry, with a particular specialisation in retirement income products. He has been employed as overall product manager at Connelly Temple (the second provider of allocated pensions in Australia) as well as product manager for annuities at both Colonial Life and Challenger Life. He has a commerce degree from the University of Queensland and an honours degree and a master’s degree in economics from Cambridge University.


Click here for the full PDF document

Retiree time-bombs

By Jim Bonham and Sean Corbett

1 Abstract

The complexity of superannuation and the age pension conceals at least 6 time-bombs – slowly evolving automatic changes to the detriment of retirees – caused by inconsistent indexation: Division 293 tax, currently only for high income earners, will become mainstream.Shrinking of the transfer balance cap relative to the average wage (which is a measure of community living standards) will reduce the relative value of allocated pensions.Shrinking of the transfer balance cap, relative to wages, will increase taxation on superannuation in retirement. Prohibiting non-concessional contributions when the total superannuation balance exceeds the transfer balance cap will constrict superannuation balances more over time.The age pension will become less accessible, as the upper asset threshold shrinks relative to wages.Part age pensions, for a given value of assets relative to wages, will reduce.

 For PDF version click here

2 Introduction

The Terms of Reference of the Review of the Retirement Income System require it to establish a fact base of the current retirement income system that will improve understanding of its operations and outcomes.

 Important goals are to achieve adequate retirement incomes, fiscal sustainability and appropriate incentive for self-provision. The Retirement Income System Review will identify:

  • “how the retirement income system supports Australians in retirement;”
  • “the role of each pillar [the means-tested age pension, compulsory superannuation and voluntary savings, including home ownership] in supporting Australians through retirement;”
  • “distributional impacts across the population over time; and”
  • “the impact of current policy settings on public finances.”

For the detailed Terms of Reference follow the link at

Terrence O’Brien and Jack Hammond have made a number of suggestions for the Review to consider (  In particular, at the close of their paper they wrote:

9. Let the modelling speak

“Only long-term modelling can show which measures are likely to have the best payoffs in greatest retirement income improvements at least budget cost. Choice of which measures to develop further are matters for judgement, balancing the possible downside that extensive policy change outside a superannuation charter may only further damage trust in retirement income policy setting and in Government credibility. “

This paper expands on that point, by presenting the results of straightforward but informative modelling which shows how the age pension, superannuation and voluntary savings (including home ownership) operate and interact – particularly over extended time periods.

A disturbing problem, because it is not obvious, arises from the inappropriate, or no, indexation of various parameters within both the superannuation system and the age pension system. This was briefly mentioned by Sean Corbett (, but has otherwise gained little or no attention.

Over the medium to long term, this inappropriate indexation will result in higher taxation, reduced age pension and superannuation for many people and a generally worse retirement outcome.  If this is the deliberate intent of the government, it should be declared.  Otherwise the problem should be corrected.

These indexation issues are the retiree time-bombs.  The Review must come to grips with them, whether or not they are intentional, and they are the focus of this paper.

3 The importance of the average wage

The impact of superannuation on an individual typically extends from the first job, through retirement to death; or perhaps even further until the death of a partner or another dependent.  For many, the age pension provides critical income through all or part of their retirement.

Accordingly, formal analysis of retirement funding often must extend over many decades, making suitable indexation an extremely important matter.

The full age pension is indexed to the Consumer Price Index (CPI) or the Pensioner and Beneficiary Living Cost Index, or to Male Total Average Weekly Earnings (MTAWE) if that is higher, which it usually is.  In May 2019 MTAWE was $1,475.60 per week or approximately $76,730 per annum.

This method of indexing is intended to allow retirees receiving a full age pension to maintain their living standard relative to that of the general community – MTAWE being taken as an indicator of that standard – and it is critical to the design of the full age pension.

For long term modelling ASIC’s Money Smart superannuation calculator suggests a combination of 2% for CPI inflation, plus 1.2% for rising living standards, giving an inflation index of 3.2% (  In this paper, that figure of 3.2% is assumed to represent future wages growth. 

This is a modelling exercise, not a prediction, so the precise value of wages growth assumed in this paper is not particularly important – use of a somewhat different figure would only change details, not the big picture – but the distinction between CPI growth and wages growth is important, the latter usually being higher.

4 Division 293 tax

Division 293 tax (see—information-for-individuals/) is a good place to begin this analysis because the problem is straightforward.

Division 293 tax has the effect of increasing the superannuation contributions tax for high income earners.  The details don’t matter here, as we are only concerned with the threshold: $300,000 when the tax was introduced in 2012, subsequently reduced to $250,000 in 2017. 

During this period the average wage has been rising steadily, which means the reach of this tax is extending further and further down the income distribution.  Eventually, if the threshold is not increased, it must reach the average wage, by which time Division 293 will have become a mainstream tax.

Fig 1 shows the actual and projected values (in nominal dollars) of the tax threshold and of the average wage.  The two are projected to be roughly equal within about 30 years.

If Labor had been successful in the 2019 election, the Division 293 threshold would have been further reduced to $200,000 so that, barring further changes, it would have been equal to the average wage in about 25 years.

Fig 2 shows the same data replotted by dividing the Division 293 threshold for each year by the average wage for that year and expressing the result as a percentage.  Expressing data relative to the average wage in each year, as in Fig 2, is often an informative way to show long term trends.

If the Division 293 threshold remains unchanged in nominal dollars, it will continue to decrease relative to the average wage, and so a greater percentage of taxpayers will be caught each year.  They will find their after-tax superannuation contributions, and hence their balance at retirement, will decrease as will their subsequent retirement income from superannuation. 

This will reduce the effectiveness of superannuation as a long-term savings mechanism.  People’s confidence in superannuation will be eroded, as it thus becomes less effective.

Failure to index the Division 293 threshold to wages growth (which would confine its effect to the same small percentage of high wage earners in future years) is taxation by stealth and makes it the 1st time-bomb.

The very simple solution is for the government to commit to indexing the Division 293 threshold in line with wages growth. 

5 The transfer balance cap

5.1       The value of allocated pensions

The transfer balance cap, currently $1.6 million, is the maximum amount with which one can start a tax-free allocated pension in retirement.  Any additional superannuation money must either be withdrawn or left in an accumulation account where taxable income is taxed at 15%.

Unlike the Division 293 threshold the transfer balance cap is indexed, but it is indexed to CPI inflation (assumed to be 2% in this paper) rather than to wages (3.2%), and adjustments are only made in $100,000 increments.

Indexing the transfer balance cap to CPI means that over time the starting value of an allocated pension account will become lower relative to the average wage, because the latter rises faster.

This is illustrated in Fig 3, where projected values of MTAWE are shown in orange; and the minimum (5%) allocated pension drawn by someone starting retirement at age 65-74, with an allocated pension account balance equal to the transfer balance cap, is shown in blue.

Despite the fact that the minimum allocated pension payment amount increases every couple of years, the average wage increases faster.

The following table shows some of the data behind Fig 3, for someone age 65 to 74 retiring in 2019, compared to retirement in 2050:

Year Transfer Balance Cap Allocated Pension (AP) MTAWE Ratio AP/MTAWE
2019 $1.6 m $80,000 $76,730 104%
2050 $2.9 m $145,000 $203,724 71%

This is a big drop in relative living standard for superannuants: 104% of MTAWE in 2019 down to 71% in 2050.

This degradation, relative to community living standards, of the allocated pension provided by the transfer balance cap is the 2nd time-bomb for the transfer balance cap.

5.2 Taxing super in retirement

Because the transfer balance cap will grow more slowly than wages, one expects that in the future an increasing proportion of a retiree’s superannuation will be held in taxable accumulation accounts rather than in tax-free allocated pension accounts.

This is another instance of taxation by stealth.  It is the 3rd time-bomb.

5.3 Non-concessional contributions

Non-concessional contributions (from post-income-tax money) are currently limited to $100,000 per annum.  That limit is indexed to wages growth, like the $25,000 limit on concessional (pre-tax) contributions.

However, non-concessional contributions are only permitted when the total superannuation balance is less than the transfer balance cap which, as we have seen, is indexed to CPI.  The effect of this indexation mismatch is that the ability to make non-concessional contributions will automatically shrink, relative to wages and living standards, in the future.  Those who rely on substantial non-concessional contributions late in their working life will be particularly affected.

This is the 4th time-bomb.

6  The age pension

6.1 Basic structure

The age pension is indisputably complicated, and a brief review may be helpful.

Detailed descriptions can be found at

A somewhat simplified description is as follows:

The full pension for a member of a couple is lower than for a single person. 

The full pension is reduced by the application of tests on assets and incomes – whichever test gives the greater reduction is the one that applies. 

Assets reduce the age pension by $78 per annum per thousand dollars’ worth of assets (often expressed as $3 per fortnight per $1,000) above a threshold.

Despite frequent claims to the contrary, including on the Human Services website quoted above, the retiree’s home is assessed by the asset test.

The assessment is achieved by applying an asset test threshold which is lower for homeowners than for renters (currently by $210,500 for singles), but the effect is exactly the same as using the renter threshold and treating the home as a non-financial asset worth $210,500 – indexed to CPI.

Regardless of how the process is defined mathematically, claiming that the home is not assessed is simply false.

Income reduces the age pension by 50 cents per dollar of income earned above a threshold, except that

  • The first $7,800 per annum of employment income is not counted
  • Income from financial assets (bank accounts, shares, superannuation accounts etc) is deemed to be 1%, for asset value below a threshold, and 3% above that; then the deemed income is used in the income test.

The full age pension, for singles or members of a couple, is usually indexed to MTAWE as discussed previously.  Every other relevant figure (the income test deeming threshold, the asset test threshold, the assumed value of the home – or, equivalently, the homeowner’s asset threshold) is indexed to CPI.

Apart from the home, the most significant asset which pensioners own is likely to be their financial assets, so a convenient way to describe the age pension graphically is to plot the value of the pension against the value of financial assets – as shown in Fig 4 for a single renter and for a homeowner, with no significant assets or income other than the financial assets.

Additional non-financial assets would shift the asset-test-controlled region further to the left.  Additional income lowers the income-test-controlled part of the curve.  For example, in Fig 5, each person is assumed to earn $20,000 per year.

Graphs such as Fig 4 allow visualisation of the complex behaviour of the age pension, which can otherwise be very confusing.  For example, a quick glance at Fig 4 shows that for the case considered, owning a home will reduce the age pension by nearly $20,000 per year for a single age pensioner with around $600,000 worth of financial assets but the effect is far less with $400,000 worth of assets.

The curves also show the steep asset-test-controlled region, where the age pension decreases at a rate of 7.8% ($78 per annum per $1,000 of assets), which tends to overwhelm the increase in actual earnings from the assets.  Although it is a major problem in itself it will not be discussed in detail here, in the interest of brevity.

6.2 The age pension time bombs

Figs 4 and 5 give snapshots at a particular point in time.  To examine the behaviour of the age pension over extended periods, however, it is best to relate asset values and income to the average wage.

This is done in Fig 6 which shows the projected curves over the next 4 decades for a single renter, with no other assets, and in Fig 7 for a member of a homeowner couple.  Note that each curve is calculated using the projected average wage for that year.

Except for part of the full age pension area of the curves, where they all overlap because the age pension is indexed to wages growth, the curves show a steady trend towards lower age pension for a given financial asset base.

Because only the full pension is indexed to wages growth and other parameters are indexed to CPI, it is mathematically inevitable that the structure of the system will automatically change slowly over time.  Relative to living standards, as indicated by the average wage:

  • Part age pensions will become harder to get, as the upper asset threshold shrinks.
  • Part age pensions for a given value of financial assets will reduce.

These are the 5th and 6th time-bombs.

What we are seeing here is universal: if different components of the system are indexed in different ways, the structure of the system will automatically change over time – even if the age pension is radically restructured.

6.3 A case study: effects of indexation on the individual retiree.

It is obvious in Figs 6 and 7 that someone beginning retirement in future years will receive less age pension for a given value of assets, when both are expressed relative to the average wage. i.e. to community living standards.

Fig 7 shows how the 5th and 6th time-bombs  work for a specific case: a single homeowner who retires at age 67 with 8 times MTAWE ($614,000 in 2019) in an allocated pension account, from which only the age-based minimum is withdrawn, and who has no other income or assets.

Four cases are shown, for retirement in 2019, 2030, 2040 or 2050.

The allocated pension is assumed to be invested in a “balanced” fund returning 4.8% nominal, less 0.5% investment fees (these values are taken from the ASIC Money Smart superannuation calculator, neglecting the small administration fee).

Because it is assumed that the retiree in each case starts with 8 times MTAWE in the allocated pension account, both the account balance and the minimum withdrawal, as a percentage of MTAWE only depend on age.  Thus, there is only one line for allocated pension income in Fig 8.  It falls fairly steadily as capital is depleted in the account.

As the asset value falls, the retiree eventually becomes entitled to a part age pension.  Starting in 2019, this retiree would begin getting a part age pension within a couple of years.  The retiree starting in 2050 would have to wait a few years longer.

Thereafter, the later retiree always gets a lower age pension, relative to MTAWE.  The full age pension – which each of these retirees approaches in their early 90s – is, however, constant as a percentage of MTAWE as already discussed.

The cause for the different treatment of full and part age pensioners is that the full pension is indexed to wages growth, while most of the factors controlling the part age pension are indexed to CPI.  If Fig 8 is reworked for a different set of assumptions (for initial balance, investment returns and withdrawal rate), the detailed shapes of the curve will alter, but the general conclusions will be unaffected.

The obvious solution is to index all parameters to wages growth, and then all the curves for part age pensions in Fig 8 will be the same and the system will be stable over time.

As it stands, the part age pension is designed to slowly become harder to get, and less generous (for a given asset value relative to living standards) – another demonstration of the 5th and 6th time-bombs.

7 Conclusion

This paper, based on relatively straightforward spreadsheet modelling, has exposed a number of time-bombs in the structure of superannuation and age pension.  These time-bombs are not particularly obvious, but they have the effect of surreptitiously increasing taxation, decreasing superannuation pensions and making the age pension harder to get and less generous.  This will reduce the prosperity of retirees and hence of the country as a whole.

The time-bombs all have their origin in failing to index all parameters in the same way.  The natural choice for an index is wages growth, because that is directly related to living standards, but if some other index is used it is still important that it be used consistently throughout the system to maintain stability.

Even if the superannuation and age pension schemes are radically altered, it remains important to index all relevant parameters in the same way.  Otherwise these time-bombs will be inevitable.

8 Disarming the time-bombs

The Review of the Retirement Income System panel is scheduled to produce a consultation paper in November 2019.

The 6 time-bombs discussed in this paper merit consideration because they touch directly on so many of the issues flagged in the Terms of Reference: “adequate retirement incomes”, “appropriate incentives for self-provision”, “improve understanding”, “outcomes”, “the role of each pillar”, “distributional impact … over time” and, most importantly, “establish a fact base”.

Identification of these time bombs is a contribution to the establishment of a fact base on retirement incomes.  Fortunately, the time bombs can be disarmed by regularising indexation throughout the superannuation and age pension systems.

30 October 2019


For PDF version click here

Save Our Super suggestions for Review of Retirement Income System

BY TERRENCE O’BRIEN AND JACK HAMMOND on behalf of themselves and Save Our Super

28 June 2019

When more individuals save for self-funded retirement above Age Pension levels, their savings contribute funds and real resources for reallocation through the financial sector to fund investments. Such an economy will be more dynamic and efficient than one which relies more on incentive-deadening taxes for redistribution through the Age Pension.

For PDF version click here

Table of contents

Save Our Super suggestions for Review of Retirement Income System

Save Our Super offers the following preliminary ideas for the Review of the Retirement Income System. We regard such a review as highly desirable and potentially path-breaking if well directed, but fraught with dangers for the Government and threatened with futility if not properly handled.

1. Embed a clear statement of Government retirement income objectives in the Review’s Terms of Reference

The Superannuation (Objective ) Bill 2016 attempted to legislate an objective for superannuation, as if that would somehow guide future governments’ detailed regulatory and tax decisions for superannuation. It did not identify the objective for the Age Pension, nor explain how the two elements ought to interact in the overall retirement income system.

Both Save Our Super and the Institute of Public Affairs criticised that attempt, and suggested improvements.

The effort to define objectives is much better set in the broader retirement income framework now envisaged.

Saving is a contested issue of philosophical vision.

To most Australians, saving is the process by which those prepared to delay gratification and consumption make real resources available to those with an immediate need for them. Savings are not just a pile of money that Scrooges sit over and count.  From the dawn of history, when families saved some of autumn’s grain to provide seed for next spring’s planting, saving in every culture has been the means by which living standards have grown and the next generation has been given more opportunities than their parents.

Saving funds investment. In the modern economy, it provides both the finance and, indirectly, the real resources that are allocated through capital markets to the businesses or loans that produce the biggest increase in the community’s living standards. If Australian investment cannot be financed by Australian saving (either by households, companies or governments running budget surpluses), it has to be financed by borrowing from foreigners or accepting direct foreign investment in Australian projects.

Viewed against that backdrop, household saving is good. Raising household savings, just like eliminating government budget deficits (ie stopping government dis-saving), reduces Australian reliance on selling off assets to foreigners or contracting foreign borrowing. People should be able to save as much as they wish, ideally in a stable government spending, taxation and regulatory structure that does not penalise savings or distort choice among forms of saving.  In such an ideal framework, they should be allowed to save for retirement, for gifts, for endowments, for bequests or for any purpose for which they choose to forgo consumption.

Specific tax treatment of long term saving (such as capital gains tax, and tax treatment of the home and superannuation) is necessary to reduce the discouragement to saving from government social expenditures and from levying income tax at rising marginal rates on the nominal returns on saving. But critics regard such specific treatment as ‘concessions’ to be reduced or eliminated. They want to limit what saving can occur, and tax what does occur. Critics think of private saving not as the foundation of investment, but as the wellspring of privilege and intergenerational inequity.

We suggest the Terms of Reference for the Review should in its preamble set the Government’s practical and philosophical aspirations for household saving and the retirement income system. We suggest the preamble to the Terms of Reference should highlight:

  • The retirement income system as a whole should aim to ensure that as the community gets richer, retirees should through their own saving efforts over a working lifetime, both contribute to and share in those rising community living standards.
  • The Age Pension should be focussed as a safety net for those unable to provide for themselves in retirement because of inadequate periods in the workforce or otherwise limited earnings and saving opportunities.
  • As the population ages, superannuation saving for retirement is likely to be a growing part of the national savings effort. Buoyant growth in superannuation finances investment and lending, and helps support rising living standards. (Conversely, a rising dependence on the Age Pension would spell only a higher tax burden).
  • The design of the retirement income system must be:
    • stable;
    • set on the basis of published, contestable modelling; and
    • evaluated for the long term, namely, the 40 or so years over which lifetime savings build, and the 30 or so years in which retirees can aspire to enjoy whatever living standards they have saved for.
  • The Age Pension and superannuation systems and the stock of retirement savings should be protected as far as practicable by grandfathering assurances against capricious adverse changes in future policy. Such changes create uncertainty and destroy trust in saving and self-provision for retirement.

2. Avoid policy prescription of savings targets or permissible retirement income standards

Some commentators have proposed the idea of a ‘soft ceiling’ on levels of retirement income saving acceptable to policy. That approach derives a level of acceptable retirement income by working backwards from the observed historical pattern of retirees’ spending, which declines with age, especially after age 80. According to those views, the fact that some retirees continue to save even after retirement is regarded as a sign of policy failure and excessively generous tax treatment (rather than of recently rising asset values). Saving is treated, in effect, as allowable to those of working age, but to be discouraged beyond a certain point, and prevented for retirees. According to this analysis, we already have “more than enough” money in retirement.

The Grattan Institute suggests a savings target such that all but the top 20 per cent of workers in the earnings distribution achieve a retirement income of 70 per cent of their pre-retirement income over the last five years of their working lives. For those in the top 10 per cent of the earnings distribution, a replacement rate of 50-60 per cent of pre-retirement earnings is “deemed appropriate”. (Approved retirement income for the second decile is not specified.)

Even after discouraging saving in this way, Grattan also recommends the introduction of inheritance taxes.

The Terms of Reference should make it clear that the Government does not support such ideas. It should emphasise it regards saving for retirement as beneficial to the community, and does not wish to limit it by arbitrary targets.

3. Prevent another ‘Mediscare’

Possible changes to the Age Pension, its means tests, compulsory superannuation contributions, superannuation taxation or regulation will be inevitably contested.

There is now zero public trust in the stability and predictability of retirement income policy. That results from the reversal, in 2017, of Age Pension and superannuation policies which, after extensive research and consultation, were introduced as recently as 2007. In addition, public trust has been eroded by the 2019 Labor election platform to make wide-ranging increases in taxes on long-term savings (that is, the Capital Gains Tax, franking credit and negative gearing proposals).

No other area of policy has more complex interactions and regulations from policy changes than the retirement income field. Complexity is such that legions of financial planners specialise in advice on the interaction of income tax, superannuation, the Age Pension and aged care arrangements.

No other area of policy takes longer lead times (40-plus years) to produce the full effect of policy change, and has the capacity to impose irrecoverable losses in living standards on vulnerable people that they can do nothing to manage or avoid. People, late in their working career or those already retired, are rightly extremely cautious about policy-induced reductions in retirement living standards they have long saved towards. It is easy for political opportunists to exploit that caution.

No review of policy will get to first base if it can be misrepresented by political opponents as creating uncertainty, destroying lifetime saving plans or retirement living standards.  Given recent experience, many voters are understandably receptive to a fear campaign of misrepresentation, including those forced to make compulsory savings throughout their working life; those dependent on the full or part Age Pension; wholly or partially self-funded retirees; and indeed all those presently retired, close to retirement, or those who have responded lawfully to legislated incentives to save as previous governments intended. If aroused to uncertainty, these groups can destroy a government.

Many of the changes that would usefully be addressed by a review of retirement income policy are potentially political third rail issues if poorly handled.  To take just one example, consider how the family home is treated under the Age Pension asset test and in the structure of Age Pension payments. Think tanks of the left and right alike have recommended that treatment be changed to take more account of wealth in the family home. Without insurance against ‘Mediscare’-type attacks, a potentially important avenue of reform would instantly be used as a scare. Government would have to either instantly rule out any change (compromising the review) or watch the reform exercise die while still suffering the political fallout as ‘the party that wants to tax your home’.

So even sensible proposed changes in policy would be discounted as untrustworthy, disruptive and unlikely to endure without careful protections. No assurance by any political party that “it is not intending to make any change” will be believed for a minute. However, these problems are avoidable with the good management sketched below.

4. Disarm scaremongering by an absolute, up-front guarantee of grandfathering

The simple, tried and proven way to disarm the ‘Mediscare’ tactic and ensure an open, constructive and intelligent Review is to make an upfront, unconditional guarantee: the Review of retirement income will be instructed to avoid any recommendations which would significantly adversely affect anybody who has made lawful savings for retirement, and who is presently retired, or too close to retirement to make offsetting changes to their life savings plans.

That grandfathering guarantee should be absolute and unconditional, referring to the use of similar practices in Australia’s history of superannuation changes from the Asprey report in 1975 through to 2010. The force of any guarantee would be increased if the Government now grandfathered some or all of the 2017 measures initially introduced without grandfathering, in the manner discussed below.

Such unconditional grandfathering would not destroy the retirement income, economic or fiscal benefits of undertaking reform. The very reason that retirement income policy changes take a long time to have their full effect is a good reason for starting policy change early, grandfathering those who made their retirement income savings under earlier rules to ensure implementation of the reforms, and letting the benefits of reform build slowly over time.

5. Propose means to rebuild and preserve confidence and trust in future consideration of retirement income policy changes

Recent policy design efforts have tried to encourage new superannuation products, such as those to address longevity risk. But such effort, necessarily focussed on the distant future of individuals’ retirements, is futile if no one trusts superannuation and Age Pension rule-making any more. If savers cannot trust the Government from 2007 to 2017, or even from February 2016 to May 2016, why should they trust the taxation or regulation of products affecting their retirement living standards 40 years in the future?

To restore the credibility of any changes emerging from the Review, the terms of reference should encourage renewed examination of ideas such as the superannuation charter recommended by the Jeremy Cooper Charter Group, or possible constitutional protection of long term savings and key parameters of the retirement income system.

6. Rebuild credible public, contestable, long-term modelling of the effects of change on retirement incomes

Retirement incomes are the result of complex and slowly developing interrelationships between demographic change, growing community incomes, rising savings, government budget developments, and Age Pension and superannuation policies. Formal, published, long-term modelling of these relationships is an essential tool to understand the impact of demographic change and of policy settings. Formal modelling facilitates public understanding and meaningful consultation, and helps build support for future retirement income reform.  Such public modelling was integral to the development of the Simplified Superannuation package in 2006, implemented from 1 July 2007, but was lacking from the 2017 reversal of those reforms.

In about 2012, the Commonwealth Treasury stopped publishing long-term modelling in this field with the last of its published forecasts using the RIMGROUP cohort model. The then-projected impacts on the Age Pension through to 2049 from the Super Guarantee measures of 1992 and the Simplified Superannuation reforms of 2007 were for a large decline in uptake of the full Age Pension accelerating from about 2010, but an increase in the uptake of part Age Pensions. There was projected to be only a small rise in the proportion of those age-eligible for the Age Pension who were fully self-funded retirees (Chart One).

The reason that the projected growth in self-funded retirement was slow is instructive: people who would, on pre-2007 policies, have been eligible for a full Age Pension could only slowly build their superannuation savings in response to the 2007 changes. Some of the first cohorts reaching retirement age would have sufficiently larger superannuation savings to be ineligible for the full Age Pension, but would still be eligible for a part Age Pension. Moreover, as they aged and exhausted modest superannuation savings, they would become eligible for a full Age Pension in later life.

Chart One: Treasury’s 2012 projected changes in pension-assisted and self-financed retirement, 2007-2049  

Source: Rothman. G. P., Modelling the Sustainability of Australia’s Retirement Income System, July 2012.

The 2017 reversals of the 2007 reforms have never been properly justified. There was no published modelling to suggest costs to the budget were higher than projected, or transition to higher self-funded retirement incomes was slower than projected. As Save Our Super warned at the time, the 1 January 2017 increased taper on the Age Pension asset test created a ‘death zone’ for retirement savings between about $400,000 and $1,050,000 for a couple who owned their home. For every extra dollar saved in that range, an effective marginal tax rate of up to 150 per cent sent the couple backward. (Similar death zones arise for other household types and single persons.)

Superannuation balances at retirement for males of $400,000 or more are common, so the practical burden of the 1 January 2017 perverse de facto tax increase could only be mitigated if a saver could quickly traverse the death zone through utilising high concessional and non-concessional contributions to accelerate late-career super savings.  But then the 1 July 2017 reductions in superannuation contribution limits scotched that hope, and compounded the damage of the 1 January 2017 change.

The longer those perverse 2017 incentives are left to operate, the stronger the incentives to build a retirement strategy around limiting superannuation savings and maximising access to a (substantial) part Age Pension. That will negate the objective of the Howard/Costello reforms to defeat adverse demographic budgetary impacts by encouraging rising self-funded retirement, growth in retirement living standards and reduced use of the Age Pension.

Outsiders will probably never know how the policy advice to make the 2017 policy reversal arose, but we speculate that the failure to publish long-term, contestable modelling since 2012 contributed to policies perversely destructive of retirement savings and encouraging tactical exploitation of access to the part Age Pension.

7. Highlight accelerated success in retirement income policy

As a result of the policies that applied up to 2017, we were witnessing a remarkable evolution of Australian retirement income outcomes that is passing unnoticed, because it is poorly explained and reported, and its end-point is still decades in the future. The combined effects of the 1992 Superannuation Guarantee process and 2007’s Simplified Superannuation are beginning to strongly reduce expenditures on the Age Pension much faster than was earlier projected.

The most recent projections of retirement developments, though only for 20 years out to 2038, were published in 2018 by Michael Rice for Rice Warner actuaries: The Age Pension in the 21st Century. (Treasury had a team leader on secondment to Rice Warner’s team of actuaries for the exercise.) The current trends are remarkable in themselves, but more remarkable in contrast to previous projections of how growing superannuation savings were changing the take-up of the Age Pension only slowly.

The proportion of those age-eligible for the Age Pension who draw a part Age Pension is still rising. (That growth comes from those previously eligible for a full age pension but now partly self-financing their retirement. So there is a net saving to the budget from this trend).  But the rise in the take up of the part Age Pension is not as much as earlier projected (Table 1, Panel 5).

What was originally projected to be only a very slight decline in the proportion of the age-eligible receiving any Age Pension (from 81 per cent in 2018 to 80 per cent by 2038), now looks likely to be a very large decline, to about 57 per cent (Table One, Panel 3, and Chart Two).

Table One: Rapid decline in Age Pension uptake projected to 2038

Notes: Intergenerational Report projections quoted are for years closest to 2018 and 2038. 2018 numbers were projections where earlier reports are cited, but are estimates of current data where a 2018 source is cited.
Sources: Intergenerational Reports for 2002, 2007 and 2015; Rothman. G. P., Modelling the Sustainability of Australia’s Retirement Income System, July 2012 (published again in the Cooper Report, A Super Charter: fewer changes, better outcomes, 2013); Rice, M., The Age Pension in the 21st Century, Rice Warner, 2018; Roddan, M., Pension bill falling as super grows, Treasury’s MARIA modelling shows, The Australian, 24 March 2019.

Put the other way around, the proportion of those age-eligible for the Age Pension who are instead totally self-funded retirees will have risen from some 31 per cent in 2018 to about 43 percent in 2038. This is a 12 percentage point rise in those totally self-funding, instead of the earlier projected 1 percentage point rise.

Reflecting this continuing gradual maturation of the system as it stood up to the 2017 policy reversals, spending on the Age Pension has already commenced declining as a share of GDP, instead of rising significantly as had been projected in early Intergenerational Reports. By 2038, spending on the Age Pension will be almost 2 percentage points of GDP lower than originally projected in the first Intergenerational Report in 2002.

Projections will doubtless evolve further. But the remarkable trends noted here are already surprising those working with current expenditure data.  The December 2018-19 Mid-Year Economic and Fiscal Outlook noted spending on the Age Pension had been overestimated by $900 million for reasons yet to be fully understood. The shortfall seems likely to involve the trends noted here, among other factors.

Chart Two: 2018 Projected proportions of the eligible population receiving the Age Pension, by rate of Age Pension

Source: Michael Rice, The Age Pension in the 21st Century, Rice Warner, p 31.

To most, the evidence of rising living standards in retirement, more self-funding through lifetime savings, less reliance on the Age Pension, a falling share of Age Pension spending in GDP and the disarming of the demographic and fiscal time bombs identified in earlier Intergenerational Reports would look like a policy triumph.

Further to this private sector modelling, in December 2018, an FOI request led to the first fragmentary public evidence of the initial uses of a new Treasury microsimulation model, MARIA , a “Model of Australian Retirement Incomes and Assets”. The model uses advances in data and computing power since Treasury’s 1990s RIMGROUP model was built to move from cohort modelling of age and income groups to microsimulation modelling of the population. This first report indicated Age Pension dependency falling markedly. Subsequent reporting of FOI information in March 2019 adds to public information that spending on the Age Pension is now falling towards 2.5% of GDP by 2038, a remarkable 1.6 per cent of GDP lower than was projected in the Intergenerational Report of 2007, the year the Costello Simplified Super reforms were enacted.

To give a sense of scale, 1.6 per cent of 2018 GDP is about $29 billion dollars. Even if GDP grew by 1% a year to 2038 (which would be a lamentable shrinkage in per capita GDP), spending on the Age Pension would be by then about $36 billion a year lower than previously projected, apparently wholly as a result of more people saving more in superannuation than was projected in the Intergenerational Report of 2007.

On 24 June 2019, more evidence of superannuation policy success became available. Analysis by Challenger, Super is delivering for those about to retire, noted that the average newly retired Australian is not accessing the Age Pension at all. Only 45% of 66-year-olds were accessing the Age Pension at December 2018 and only 25% of them were drawing a full Age Pension.  Of course, many of those might fall back on the Age Pension in later life when they exhaust their superannuation capital. Thus, if retirees are to remain wholly self-funded during their whole retirement, superannuation balances at retirement will need to keep rising. Other things being equal, 2017’s imposition of a $1.6 million cap and tax at 15% on amounts above that balance reduce the time that retirees can remain independent of the Age Pension.

Every additional person who wholly self-funds their retirement is, prima facie, achieving a better retirement living standard than they could have enjoyed by arranging their affairs to access only the Age Pension and to send the bill to working age taxpayers. This is not merely a budget success. An economy in which individuals save for retirement, contributing funds and real resources for reallocation through the financial sector to fund investments will be much more dynamic and efficient than one more dependent on the Age Pension, in which people pay incentive-deadening taxes for redistribution through the welfare system.

It is bewildering to us that the accelerated success of superannuation policy, which has helped people save for their desired retirement standard of living, is not being trumpeted from the rooftops.  Instead, the facts are dribbling out without explanation and context from FOI applications. Those facts are lost against the backdrop of incessant criticism from some commentators of More than enough saving and excessive revenue forgone from the tax treatment of superannuation.

It is vital for protecting the Government from a repeat of the backward steps on retirement income policy in 2017, for restoring the legacy of the Howard-Costello reforms and for timely identification of sustainable future reforms, to re-establish regular published, contestable and peer reviewed modelling of how retirement income policy is working. 

8. Commission initial modelling of three scenarios

We suggest Treasury should use MARIA to model three scenarios over a long-term time frame such as 2000 to 2060 to clarify the starting point for the Review of Retirement Income Policy.

Rather than study retirement income policy solely as a Commonwealth budget issue of the revenue hypothetically forgone in tax incentives for superannuation and the expenditure on the Age Pension, the modelling needs to be set in the fuller context of the Howard Government’s 2006-2007 analysis of Simplified Superannuation. Its output ought to include impacts on retirement incomes, the ‘RI’ in MARIA, not just on the budget. 

As noted above, the overarching objective of policy ought be to enable higher lifetime saving and rising living standards in retirement for those in a position to save for self-funded retirement, while preserving the Age Pension as a safety net for those unable to save for a better retirement living standard. Modelling should project implications for average self-financed and Age Pension retirement incomes under each scenario, as well as for government revenues and expenditures.

a) A baseline scenario

We suggest the first scenario for long-term modelling should be the projected effects by 2060 of the continuation of Age Pension and superannuation policies as at end 2016. That would be comparable against the earlier 2012 Treasury modelling, and would show the impact of another 6 to 7 years’ data on the maturation of the Super Guarantee (including scheduled future increases) and the Simplified Superannuation reforms of 2007.

b) A current policy scenario

We suggest a second useful scenario would be to model the current policies. When the current policy scenario is compared to the baseline scenario, that would give an indication of the effect of the change in the taper rate of the Age Pension asset test, the imposition in the retirement phase of a 15% tax on earnings on superannuation balances above $1.6 million, and tighter restrictions on concessional and non-concessional contributions. Effects on individual retirement incomes, as well as comparisons of effectson government revenue and expenditure over time, should be made between the two models.

It might be objected that the behavioural responses to the 2017 changes are too recent to have shown up in data and thus too difficult to model. But to assert that we can have no estimate of the likely effect of those policy changes on retirement incomes would be in effect to concede that they should never have been proposed or implemented.

c) Future policy change scenarios

A third useful scenario could involve empirically testing policy changes the Government wanted to explore, including grandfathering the changes introduced in 2017 and summarised in Table Two.

Any mix of measures selected should cohere around the Government’s retirement income strategy, to allow those who can save to raise their retirement living standards, to protect the efficacy and affordability of the Age Pension as a safety net for those who cannot, and to ensure as many as possible are in the first group.  The selection of measures must rest on the evidence of public, contestable, long-term modelling of outcomes on both retirement living standards and the government budget.

Because of these strategic and empirical imperatives, Save Our Super advocates that the grandfathering of all the measures of Table Two be enumerated and modelled, as well as the 2017 change to the means testing of the Age Pension and the impact of Superannuation Guarantee changes.

Table Two: Options for grandfathering 2017 superannuation changes

We illustrate one possible, strategically coherent path forward but without any implied prioritisation. Some potentially useful measures might:

  • Reduce the burden of the Superannuation Guarantee on the youngest (who have longest to fund their own preferred retirement living standards and face the highest competing demands on their early-career budgets) and the poorest (who will in any event accumulate insufficient savings over their working lifetimes to become ineligible for the Age Pension). This could involve either raising the cut-in point for the Superannuation Guarantee, halting its programmed rate increases, or both.
    • Against the merits of the Superannuation Guarantee must be set the cost that it forces a constant rate of saving for employees by their employers over the employees’ working lifetimes. In any event (but especially if the Government raises the Superannuation Guarantee rate), this is a particular burden on the young, those in tertiary study, those seeking to buy their first home, those establishing a family and those with low or punctuated career earnings.
    • One curious and little noted feature of the Superannuation Guarantee is that (broadly speaking) it applies to any employee over 18 who earns $450 gross or more a month. This extraordinarily low threshold has not been altered since the Super Guarantee was introduced at 3 per cent in 1992 – over a quarter of a century ago. At that time, the monthly $450 trigger corresponded to the then annual tax-free threshold in the income tax of $5,400. With the Superannuation Guarantee now at 9.5 per cent and scheduled to increase to 12 per cent, it is now a significant impost that falls as forgone wages on young and/or poor workers, when they have priorities of education, housing and family expenses much more pressing than commencing saving for retirement more than 40 years in the future. If the Superannuation Guarantee cut in at the monthly gross earnings equivalent to the current tax-free threshold, the trigger would now be $1517 a month, not $450 a month.
  • Remove the discouragement of saving from effective marginal tax rates of over 100%, encouraging saving by those who can save to escape reliance on the full Age Pension. This would require reversing the increased taper on the Age Pension asset test imposed on 1 January 2017 and reinstating the Costello reform of 2007.
  • Allow those who are able to save for their desired retirement standard of living, in the latter parts of their career, access to higher concessional and non-concessional superannuation contributionlimits, as shown in Table Two.
  • Acknowledge that self-funding a retirement standard of living which is higher than the Age Pension requires a large capital sum at retirement – the Age Pension for a married couple is estimated to have an actuarial value of over $1 million, with the costs rising in an environment of near-zero interest rates. At present, all political parties say they want an end (increased self-funding of retirement), but seem to attack the means to that end: a large capital sum accumulated at the end of the saver’s working career. With the continuing drift of interest rates towards zero, whatever unexplained calculations arrived in 2016 at the $1.6 million cap on superannuation should be re-examined, with a view to grandfathering the cap as shown in Table Two, raising it or abolishing it. The interest earnings from a $1.6 million sum is now almost 40% lower than it was in early 2016. Abolishing  the $1.6 million cap would re-capture many of the simplification benefits of the 2007 Simplified Super reforms, which were destroyed by the 2017 changes.

d) Strategic direction of future policy change scenario

These four classes of change have clear strategic directions: they are pro-choice, pro-personal responsibility and support rising living standards in retirement. They reduce emphasis on forced savings at a high and constant rate over the whole of working life from the earliest age and the lowest of incomes. They increase emphasis on saving at the rate chosen by individuals over their working careers in the light of their circumstances. The shift would likely result in faster late career savings after educational, family formation and mortgage commitments have been met. The shift would be pro-equity, in that it would avoid reducing the living standards of the youngest and poorest in the workforce, without ever helping many of them achieve retirement income living standards above the Age Pension. And it would reduce the constituency of voters denied growth in their own disposable incomes and supportive instead of increased government transfers to them for their early-career expenditures (such as childcare and other family benefits).

e) Budget effects of future policy change scenario

Of these four classes of change, the Superannuation Guarantee changes would raise significant revenue for the government budget, since higher incomes paid as wages would be taxed under normal income tax provisions, rather than at the lower rate for superannuation contributions. The other three measures would have a gross cost to budget revenue relative to the current measures, but would continue and likely accelerate the recent and faster-than-projected exit of retirees from dependence on the full Age Pension. That will save some future budget outlays, and it is unclear until public, contestable long-term modelling is published what the net effect on the budget would be, and its time frame.

Recall, however, that the origins of this debate were the demographic time-bomb facing Australia, and the intrinsically long-term challenge of building life-time savings for long-lived retirement. A measure that ‘breaks into the black’ even decades hence might be counted a success.

f) Retirement income effects of future policy change scenario

Whatever the net budget effects and their timing, it is clear a package such as sketched in scenario 3 will raise Australian’s retirement incomes and protect the sustainability of the Age Pension

9. Let the modelling speak

Only long-term modelling can show which measures are likely to have the best pay-offs in greatest retirement income improvements at least budget cost. Choice of which measures to develop further are matters for judgement, balancing the possible downside that extensive policy change outside a superannuation charter may only further damage trust in retirement income policy setting and in Government credibility.


For PDF version click here

7.30 Report interviews Save Our Super on Kelly O’Dwyer – 24 April 2017

Click here to view the ABC’s 7.30 Report interview of Save Our Super’s Jack Hammond QC and John McMurrick on Kelly O’Dwyer.

Click here for the transcript of the interview.

Grandfathering provisions

Save Our Super believes that major changes to the existing rules of the Australian superannuation system should not be made unless, at the same time, appropriate grandfathering provisions are included in the legislation.

“Grandfathering provisions” are qualifying clauses within legislation which exempt those people already involved in the activity with which the legislation deals.

Load more