Responding to Callaghan: completing Australia’s retirement income system

Jan 6, 2021

The Retirement Income Review (Callaghan) Report concluded that the Australian retirement income system is effective, sound and its costs are broadly sustainable.

Credit – Unsplash

While the Review’s terms of reference were ‘to establish a fact base’ rather than make recommendations, the Report has a very clear overall message: that we should now focus on settling the pensions phase of the system, moving on from the focus up until now on the accumulation phase. That is a message many of us have been pressing for several years now.

Should the Government decide not to proceed with the legislated increase in the superannuation guarantee (or at least not the full increase to 12%), citing the Report’s evidence that optimal use of savings accumulated at the current 9.5% can deliver adequate retirement incomes for those on median incomes or below, there must be a quid pro quo. The Government should fix the outstanding problems in the pensions phase that are causing sub-optimal use of savings and leaving gaps for particular vulnerable groups. Unless these problems are resolved and retirees fully utlilise their savings, then as the Report finds, an accumulation rate of 9.5% will not deliver adequate retirement incomes.

Unfortunately, Callaghan is not very helpful about the solutions to these problems, notwithstanding its support for funds to be required to provide regular estimates of expected savings in terms of the income stream they can generate and for people to be guided to income streams that include reasonable protection against market and longevity risks and against thinking of the savings as a ‘nest egg’ or for passing to the next generation. The Report highlights the underlying problem of complexity and suggests the need for much greater cohesion particularly between superannuation and the age pension, but it offers no support for any of the practical solutions it canvasses. Strangely, it also questions the value of the measures available to directly increase support for those facing the greatest financial stress without offering alternatives.

Optimal Drawdowns

The Report’s ‘optimal’ use of superannuation is questionable. It is doubtful that funds could offer the drawdown arrangements it involves: these are highly complex as they are designed to reduce real superannuation incomes as people get older to offset exactly real increases in the age pension, with superannuation balances dropping to zero at age 92. The optimal use also assumes that 5% of the savings will be directed into a deferred lifetime annuity from age 92: but the market for life annuities is already struggling and the likelihood of funds being able to offer an annuity that does not start to be paid until 25 or more years into the future seems extremely unlikely. This ‘optimal’ arrangement would also still leave unused some of the accumulated savings whenever the person died before age 92.

Much more likely models for sensible use of superannuation, consistent with the objectives of the system, would include life annuities from retirement (requiring most of the savings) with the remainder available for spending in the first decade or so for holidays and other active retirement living consumption; a small amount might be retained as insurance against unforeseen events but, as Callaghan rightly says, there is significant public insurance for health and aged care. Such models would of course need to vary depending on the person’s accumulated savings and the extent to which they should sensibly rely on the age pension, including for managing longevity risk. But for most people, the simplicity and security involved in having some form of life annuity starting at retirement should be most attractive. And if the funds claim they cannot offer life annuities, the Government should step in and sell them.

The Report’s emphasis on avoiding any real increase in total retirement income (super plus age pension) while acknowledging that the pension itself should increase in line with community incomes, would also cause some anomalies. As people age, retirement incomes at lower income deciles would become increasingly compressed. The more likely models involving CPI-indexed life annuities would, setting aside the funds kept for early consumption, lead to a pattern of regular income growth somewhere between wage indexation (for those reliant solely on the pension) and CPI indexation for those not reliant on the pension at all. Such a gradation would avoid the anomaly the Report’s ‘optimal’ model would cause.

While following the 2014 Financial System Inquiry the Government has lent support to requiring funds to offer ‘Comprehensive Income Products for Retirement’ (CIPRs), it is now high time to advance that agenda and to clarify what the CIPRs should look like. It is disappointing, however, that the Callaghan Report fails to provide useful guidance in this regard, especially as the progress in introducing CIPRs has been very slow and without them the Report’s key conclusion does not stand up.

Further work is also still needed to confirm or otherwise that key conclusion that appropriate use of superannuation savings would lead to adequate retirement incomes for those on median income at the current 9.5% superannuation guarantee. Despite my previous advocacy for the SG to increase to 12%, my suspicion now is that Callaghan’s conclusion will not be far out. Of course, those on higher incomes need to supplement the SG with voluntary savings. The Report mentions the FitzGerald finding in 1993 that, in the absence of the age pension, people would need to save around 18% for an adequate retirement income, a rate not dissimilar to the 20% OECD average contribution rate (or the rates applying in most existing Commonwealth Government schemes (including employee contributions)).

Cohesion and the Age Pension Means Test

The Report has a whole chapter on ‘Cohesion’ which hints in particular at the need for a simpler age pension means test so people can plan ahead sensibly. But it then suggests that a merging of the separate income and assets tests, or the Henry Report suggestion of a single income test with deeming of income from assets, would be too hard and expensive. It also seems to dismiss widespread concerns about the current assets test taper, highlighting the cost of any relaxation and that, under its (questionable) optimal drawdown model, net retirement incomes would still increase with marginal increases in savings.

But even with the high drawdowns the Report uses, the marginal increase in net retirement incomes from marginal additions in savings are very low as the graphs in the Report reveal. This suggests that, at the margin, the trade-off between incomes during working life (reduced by such extra savings) and incomes in retirement (very slightly increased at best) is extremely poor, and certainly not justified when the extra savings are compulsory. It is concern about just such trade-offs that the Report elsewhere emphasises when questioning the need to increase the SG.

The Report states that evidence that the current means test affects savings behaviour prior to retirement is weak. That is hardly surprising given how complex the means test arrangement now is. But if the system was ever to achieve the cohesion the Report advocates more generally, then people will have knowledge ahead of retirement of the marginal impacts from savings and, after retirement, of how best to integrate their superannuation and any age pension entitlement. The Report also notes the cost of lowering the taper and that those who would benefit would be people with more substantial savings. What is not explored is some combination of relaxing the taper and reducing the thresholds (reversing the 2017 changes which have caused such problems in the first place).

The Report shows how the separate income and assets tests (unique to Australia) cause inequities in the treatment of people with similar resources, and hints that a merged means test might assist. But the Report concludes that ‘it would be challenging to design a merged test that achieves the objective of the current dual means test but is less complex’. I am not at all convinced that this is so. The example chosen for Appendix 6B is particularly complex with suggested income equivalents of assets rising with age (the same pattern the Report criticises when currently used for minimum drawdown rules). David Knox and I have suggested a much simpler version (drawing on the experience of the test that existed in the 1960s and 1970s) with a standard deemed income factor, slightly above the price of an indexed life annuity at age 67, and with actual lifetime annuities treated as income only but in full. Thresholds for the savings included would be reduced, particularly for home-owners.

One of the impacts of such a merged means test would be an effective relaxing of the current assets test taper, as the deemed income would be subject to the existing 50% income test taper.

More work is needed on the design of a merged means test, but if we are to achieve cohesion and ensure people can properly and confidently plan how much to save and how to use their savings to achieve, with some age pension, an adequate and secure retirement income, this is an essential component. The challenge of also achieving cohesion with aged care arrangements – also referred to in the Report – might best be addressed later when the Royal Commission into Aged Care reports.

Consideration should also be given to complementary administrative reforms to limit the frequency of reporting about means and the frequency with which age pension entitlements are adjusted. A worrying fact mentioned in the Report is that, on average, age pension entitlements are currently varied four times a year!

Support for Vulnerable Groups

While concluding that the level of the pension is sufficient to protect people from poverty if they own their own homes, the Report highlights the extent of financial stress faced by retirees in private rental accommodation and those involuntarily retired before age pension age. But the Report does not lend support to the most obvious measures to reduce this stress.

It suggests that increasing rent assistance would have only limited impact on the different position of renters and home-owners and would cost a lot; it suggests – with no clarification – that ‘a broader approach to assisting renters in retirement appears necessary’.  Increasing rent assistance towards the level of subsidies available to those in public housing (as has been recommended as long ago as 1995 as well as in the Henry Report) would seem to deserve a high priority despite the Report’s scepticism. It should have happened in response to the 2010 Harmer Report on pension adequacy rather than the Rudd Government’s far more expensive across-the-board increase in the pension. There is also a case for addressing the supply of low cost housing but this should complement, not be in place of, increasing rent assistance.

For the early retirees on Jobseeker, the Report merely notes that ‘any change in the rate of the Jobseeker Payment must consider its broader implications to all age groups and many recipients may re-enter the workforce’. The case for increasing Jobseeker is already overwhelming, but the case for those over age 60 is particularly strong – almost 60% of those becoming unemployed at age 60 never achieve 12 months without social security assistance before reaching age pension age, the Report reveals.


The Report rightly draws attention to the inequities involved in treating assets held in the home differently to other assets. There is little doubt that, in principle, the value of the home should be taken into account in the means test (and also treated more equally with other savings in the tax system). But home ownership is rightly seen as one of the pillars of the retirement incomes system, providing both security and income-in-kind. With home ownership, the age pension ensures adequate protection from poverty.

The difficulties with including the home in the means test go well beyond the obvious political ones given over 100 years of exemption. Apart from identifying an appropriate threshold (or thresholds based on location), it would be essential to ensure easy access to an income stream funded by the home assets for those left with insufficient age pension support.

The current Pension Loans Scheme goes a little way towards helping people draw on their home assets to supplement their age pension, but there would be considerable advantage if many more home-owners had access to income streams financed from their home assets, particularly amongst the current cohorts of retirees who do not have substantial superannuation savings. This would enhance the retirement incomes system whether or not action was taken to include the home in the means test. It should not be difficult to design market-based schemes with appropriate consumer protection to guarantee occupancy and to limit the net reduction in the person’s home equity. If the market has difficulty offering such products, however, there is a strong case for the Government to extend the existing scheme, based on repayment from the eventual estate.

A step towards more equitable treatment of renters vis-à-vis home-owners would be to increase the gap between the assets test threshold for renters above that for home-owners.

Taxation of Superannuation

The Report highlights its assessment of the cost of tax concessions for superannuation and that they are concentrated on those with higher incomes. But this analysis is premised on how tax concessions are defined and the benchmark against which they are compared. The Annex at pages 407-409 briefly canvasses the issues involved but the approach taken remains questionable.

The Report argues that the concessions should be benchmarked against ‘what is’ rather than ‘what should be’, where ‘what is’ refers to the existing ‘comprehensive income’ tax treatment of other savings i.e. taxing both contributions and earnings at the individual’s marginal rate of tax but exempting any tax on final spending (or TTE). Some account is taken of the likelihood of changed savings behaviour under such a tax regime, but this is assumed to be small as the superannuation savings are mostly compulsory.

For good reason, however, superannuation could never be taxed this way. The Henry Report, like previous studies, argued that the TTE approach taxes savings excessively discouraging savings. Just because super is largely compulsory is surely no justification for over-taxing it. The degree to which TTE is excessive increases with the length of time the savings are held, as a recent report by TTPI demonstrates, because the tax on earnings is effectively a wedge that compounds. The research mentioned in the Report clearly favours TEE or EET regimes (particularly for longer-term savings) or regimes that also have a modest tax on earnings (TtE or EtT).

Yet the Report’s figures highlight the ‘concession’ on how earnings are taxed in particular, revealing that those on the highest incomes and who are the oldest (and hence have held their superannuation savings the longest) are gaining the most ‘concessions’. In terms of any reasonable counterfactual, that is misleading.

Moreover, it plays into the hands of those interest groups who believe there is a magic pudding of tax revenues available for redirection from the wealthy to their particular priorities.

The case for an EET regime for superannuation is actually very strong. Pension schemes overseas generally take this approach as we still do in Australia in the case of the old public sector benefits promise schemes. This is not just because most overseas schemes have no identifiable contributions (or earnings) that could be taxed but also because they are generally compulsory and cannot be accessed before retirement: they do not provide any taxable capacity when they are accumulating. Moreover, the savings are very long-term, generally around 40 years.

Research a few years ago commissioned by the Committee for Sustainable Retirement Incomes suggested that the post-Turnbull reformed ttE tax arrangements were broadly equivalent to an EET regime and hence about right (the CSRI accepted that shifting directly to the preferred EET regime is too hard now). The progressive tax on contributions flows through to the pensions phase, and hence the conclusion about being broadly equivalent to EET applies at all income levels.

Perhaps, as Callaghan seems to be pushing, the tax on earnings could be extended to superannuation held in the pensions phase without upsetting the CSRI research conclusion or causing excessive tax overall, but the Report’s view that 15% represents a ‘concession’ (and really should be increased) is inappropriate. If anything, I suspect the tax on earnings should be lower – perhaps a revenue neutral move to around 10% should be phased in for both the accumulation and pension phases.  (A recent report by the Tax and Transfer Policy Institute suggests that, if all forms of savings were from after-tax income, a flat tax of under 10% on the earnings would be revenue-neutral, progressive and efficiency-enhancing.)

An Agenda to Complete the Retirement Income System

Most likely, the Callaghan Report is correct that the SG does not need to increase to 12%. Further work might confirm this (and whether 9.5% is sufficient or moving to 10% is still advisable).

But the Report also demonstrates the need now to settle the pensions phase of our retirement income system. Foregoing the full increase in the SG to 12% (and the costs to revenue involved) requires a quid pro quo that does precisely this if we are to ensure an adequate retirement income for most retirees. This quid pro quo should include:

  • Funds to report regularly to members on the income streams in retirement their superannuation savings are likely to lead to;
  • Accelerating the move to require funds to offer CIPRs that guide people to optimal use of superannuation which, with any age pension entitlement, delivers adequate and secure retirement incomes (if necessary, involving the Government selling indexed annuities);
  • A simplified age pension means test that ensures higher savings do in fact lead to improved net retirement incomes while still concentrating assistance on those most in need;
  • A sizeable increase in rental assistance, towards that available to public housing tenants, and a sizeable increase in Jobseeker, particularly for the over 60s;
  • The introduction of a broader home equity release program with repayment from estates, if necessary managed by the Government.

Some of these measures need further development but the work required is not particularly complex nor are the financial and political hurdles involved excessive. Indeed, the final outcome would confirm the quality and effectiveness of the emerging Australian retirement income system and its deservedly high international standing.

This article was originally published by the ANU Crawford School of Public Policy.

Share and Enjoy !

Subscribe to John Menadue's Newsletter
Subscribe to John Menadue's Newsletter


Thank you for subscribing!