The retirement phase of superannuation
Require superannuation trustees to pre-select a comprehensive income product for members’ retirement. The product would commence on the member’s instruction, or the member may choose to take their benefits in another way. Impediments to product development should be removed.
Government should require superannuation fund trustees to pre-select an option for members to receive their superannuation benefits in retirement. Details of the pre-selected option would be communicated to the member during their working life. At retirement, the member would either give their authority to commence the pre-selected option or elect to take their benefits in another way. This approach would simplify decisions at retirement and deliver better outcomes for retirees (Figure 8: Stylised example of decision making for superannuation benefits). No income stream would commence without the member’s instruction.
The pre-selected option should be a comprehensive income product for retirement (CIPR) that has minimum features determined by Government. These features should include a regular and stable income stream, longevity risk management and flexibility. CIPRs would be low-cost and include a ‘cooling off’ period. Their design could vary with the member’s known characteristics, such as the size of their superannuation benefits, and take account of the possibility of cognitive impairment at older ages.
A combination of underlying products would likely be required to provide these features; for example, an account-based pension paired with a pooled product that provides longevity risk protection. To offer these products, funds may need to partner with another provider, such as a life insurance company.
Regulatory impediments to developing retirement income products, which include tax policy settings, need to be removed. These changes should not discourage the use of CIPRs or other products that provide longevity risk protection.
Figure 8: Stylised example of decision making for superannuation benefits
- Better meet the needs of retirees, including those who are disengaged or less financially sophisticated, and provide a more seamless transition to the retirement phase of superannuation.
- Achieve the objectives of the superannuation system (discussed earlier in this chapter) by strengthening the focus on providing retirement incomes.
- Improve Australians’ standard of living during their working lives and retirement through greater risk pooling.
Problems the recommendation seeks to address
Managing multiple financial objectives and risks in retirement is complex. For example, retirees may seek to maximise income while trying to retain flexibility to meet unexpected expenses and manage longevity, investment and liquidity risks. Individuals have to manage these problems without the guidance that exists in the accumulation phase, where funds are required to offer simple, low-cost default accounts. MySuper is an accumulation product only, despite the Super System Review recommendation that MySuper products include a retirement income stream.54
Information from stakeholders suggests that many retirees find it challenging to navigate the transition to the retirement phase of superannuation. When DC members notify their superannuation fund of their retirement, many funds recommend they speak to an affiliated financial adviser. Research has demonstrated that the quality of this advice can vary significantly.55 Anecdotal evidence suggests that some advisers have limited knowledge of longevity risk and how it can be managed.56 Although the Inquiry makes recommendations to improve the quality of advice, it will take time for such improvements to occur. Chapter 4: Consumer outcomes explores this issue in further detail.
In any case, many people do not seek professional advice, and funds and advisers overwhelmingly recommend account-based pensions. Stakeholders advise that, for less financially literate individuals, the simplest option is to take the entire benefit as a lump sum because other options can be difficult to understand and may require completing complicated forms. A recent survey commissioned by AustralianSuper found that “… 85% of pre-retirees are not confident in having an informed conversation around retirement income”.57 Managing income and risks can be particularly difficult for people later in retirement if they suffer from cognitive impairment.
The complexity of retirement decisions is compounded by behavioural biases.58 Mandatory superannuation contributions have been used to overcome behavioural biases in saving behaviour, such as decision making that disproportionately focuses on the short term; however, these biases do not end at retirement. In part, behavioural biases explain the dominance of account-based pensions and lump sums.
Limited range of income stream products
Australians who wish to convert their superannuation assets into an income stream generally have the choice of an account-based pension or an annuity. A well-functioning market would be expected to provide a wider range of products that meet different needs and preferences. This would allow people to combine products to achieve their desired levels of income, risk management and flexibility. However, there are tax, regulatory and other impediments to developing innovative retirement income products.
Despite the heterogeneous nature of retirees, at least 94 per cent of pension assets are in account-based pensions, which provide flexibility but lack risk management features and may not deliver high levels of income from a given accumulated balance.59,60 The lack of a significant market for products with longevity risk protection sets Australia apart from most other developed economies.61 Evidence suggests that the major worry among retirees and pre-retirees is exhausting their assets in retirement.62 An individual with an account-based pension can reduce the risk of outliving their wealth by living more frugally in retirement and drawing down benefits at the minimum allowable rates.63 This is what the majority of retirees with account-based pensions do, which reduces their standard of living.64,65 The difficulty in managing this risk is also exacerbated by the uncertainty as to how long a retiree will live.66
The potential gains to members, the economy and taxpayers from a more efficient retirement phase are significant and warrant intervention. Higher income in retirement and a wider range of retirement income products would better meet the varied needs of retirees. The economy will benefit if the growing proportion of people in retirement can sustain their level of consumption.
Combinations of products enable retirees to balance the three desired features of retirement income products: high income, risk management features and flexibility (Figure 9: Desired features of retirement income products). Pooling of longevity risk would give retirees greater confidence to consume. Alternatively, by improving the superannuation system’s efficiency in providing retirement income, people may be able to save less during their working lives to reach a given level of retirement income.
Figure 9: Desired features of retirement income products
Private provision of longevity risk protection could benefit taxpayers and the broader economy. It would shift some of the longevity risk borne by Government, as the provider of the Age Pension, to the private sector. Assets underlying products with longevity risk protection could be invested with a longer time horizon, helping to fund long-term investments and develop the corporate bond market in Australia as funds seek more investments that provide a steady flow of income.
The Interim Report broadly identified four options that would enable the retirement phase to better achieve the objectives of the superannuation system and position Australia to manage the challenges of an ageing population:
- Recommended: Require superannuation trustees to pre-select a CIPR for members.
- Recommended: Remove impediments to retirement income product development.
- Provide policy incentives that encourage retirees to purchase retirement income products that help manage longevity and other risks.
- Mandate specific retirement income products (in full or in part, or for later stages of retirement).
Option costs and benefits
Require superannuation trustees to pre-select a CIPR for members
Under this option, superannuation fund trustees would be required to pre-select CIPRs for members to receive their superannuation benefits. Members would retain the freedom to take their benefits in another way if they desired. Appropriately designed pre-selected options would offer an effective, low-cost means of improving retirement income and risk management for superannuation fund members without limiting personal freedom and choice. This option would also help to develop markets for a wider range of income products with enhanced risk-management characteristics.
Greater use of products that pool longevity risk could significantly increase retirement incomes. For many retirees, incomes from CIPRs could be 15–30 per cent higher than those from the current typical strategy of drawing the minimum amount from an account-based pension.67 The Inquiry notes that one of the primary reasons why incomes are significantly higher in products that pool longevity risk is that they reduce bequests from superannuation. Although the system should accommodate bequests, it should not do so to the detriment of retirement incomes.
Pre-selected options would simplify the process of using an accumulated balance to generate an income stream. The member guides that exist for the accumulation phase currently cease at retirement. Pre-selected options would bring the policy philosophy in the retirement phase closer to that of the accumulation phase.
People can benefit from pre-selected CIPRs in different ways. CIPRs can improve retirement incomes and risk management outcomes, especially for the disengaged and those who are less financially sophisticated, by providing a comprehensive option that balances a number of objectives and risks. The design of CIPRs can also guide more engaged members by providing a framework for decision making. Johnson and Goldstein (2013) suggest such an approach would be effective for three main reasons:68
- Effort — real or psychological costs of moving from the pre-selected option.
- Implied endorsement — products are perceived to be recommended.
- Loss aversion and reference dependence — decisions are affected by the starting point.
Pre-selecting options can have a significant influence on decision making. This was shown in an Australian experiment that involved individuals allocating savings between account-based pensions and annuities. The study found that the distribution of allocations was strongly clustered around the pre-selected allocations (Chart 4).
Chart 4: Proportion of assets taken as an annuity with different pre-selected allocations
Source: Bateman et al.69 This study referred to pre-selected options as defaults. The allocation to annuities was pre-selected for participants who were able to change the allocation as they desired.
Although designing CIPRs would be more complex than designing accumulation accounts, trustees are well placed to select appropriate CIPRs for their members, within a defined framework. ASFA argues “… trustees should exercise their fiduciary duty to consider longevity, market risk and inflation risk in designing post-retirement arrangements in much the same way they need to consider investment risk and insurance needs throughout the accumulation stage”.70 Funds would incur the costs of selecting, establishing and maintaining CIPRs, but these costs would be justified by the benefits to members.
However, CIPRs would be less beneficial to individuals with very high or very low superannuation balances. Those with small balances are likely to continue to take their benefit as a lump sum and rely primarily on income from the Age Pension. Individuals with very high balances may be able to generate satisfactory retirement income from an account-based pension, drawn down at minimum rates.
A change in the way superannuation benefits are taken could affect Age Pension costs, as discussed in the Implementation considerations section later in this recommendation.
Remove product development impediments
As discussed in the Interim Report, regulatory and other policy impediments are hampering the development of retirement income products such as deferred lifetime annuities (DLAs) and group self-annuitisation (GSA) schemes.71 These impediments include rigid standards to qualify for earnings tax exemptions. The Inquiry supports the review of retirement income stream regulation being undertaken by Government, which is examining ways to reduce or remove barriers to developing a market for DLAs.
However, increasing the range of products alone will not be sufficient to improve outcomes for retirees significantly. Behavioural biases and other system incentives will continue to impede the widespread use of pooled longevity risk products, despite evidence that many individuals would be better off.72
Provide policy incentives to encourage the use of retirement income products that manage longevity and other risks
Several submissions support using tax and Age Pension incentives to encourage take-up of income products with longevity risk protection. The Australian Institute of Superannuation Trustees (AIST) recommends “… policy incentives that encourage retirees to purchase retirement income products that help them deliver their optimal retirement experience”.73 In the past, favourable treatment of products with better risk management features by the Age Pension means test has been shown to increase their use.74
Although policy incentives would be likely to increase the use of superannuation to generate retirement incomes, incentives generally increase the cost of the retirement income system to taxpayers. However, it is important that tax and Age Pension settings do not discourage people from using CIPRs.
Mandate specific retirement income products
Mandatory use of certain retirement income products would achieve effective risk pooling and ensure superannuation is used to provide income throughout retirement. It would also overcome issues with adverse selection in products with longevity risk protection.
Mandatory use of products that pool longevity risk could disadvantage groups with lower life expectancies.75 This could be mitigated by pricing products to reflect the characteristics of members, including their life expectancy, although this would be complex and add costs.
People tend to have diverse needs in retirement, and no given product or combination of products will be appropriate for everyone. Many submissions caution that compulsory income streams could result in poor outcomes for some individuals and stifle innovation. Although this option could help achieve the objectives of the superannuation system, it would remove individuals’ flexibility to tailor their retirement plans to suit their needs and is not consistent with the Inquiry’s philosophy.
Pre-selected CIPRs and greater use of longevity risk pooling at retirement could significantly improve the superannuation system’s efficiency in providing retirement incomes and better meet the needs of retirees.
In making this recommendation, the Inquiry sought to balance the desire to increase system efficiency in providing retirement incomes with a degree of individual freedom and choice. The Inquiry favours an approach that preserves freedom and choice. However, if introducing pre-selected CIPRs does not achieve the intended objectives of this recommendation, Government could consider forms of defaults that commence automatically on retirement. Tax and Age Pension incentives could also be used to better achieve the objectives of the superannuation system.
High-quality advice may be useful to some individuals to help them manage their financial affairs in retirement. Chapter 4: Consumer outcomes contains recommendations to improve the quality of financial advice.
Developing CIPRs will take considerable time. This recommendation should be implemented with sufficient lead time to allow superannuation funds to design products or form partnerships with other providers.
As pre-selected CIPRs are expected to influence retirees’ decisions, it is important they provide good outcomes for large numbers of members. This recommendation involves trustees potentially designing, advising on and delivering CIPRs. For these reasons, Government should establish a mechanism to ensure each CIPR provides the required features, which should be specified in regulation. Ongoing regulatory oversight will also be required. Meeting regulatory requirements should provide trustees with some protection against breaching their fiduciary obligations.
CIPRs should be offered as a pre-selected option to all members of APRA-regulated funds — not only to MySuper members. Decisions about retirement are more complex than most decisions made in the accumulation phase. Individuals who leave the default system are still likely to benefit from CIPRs. Self-managed superannuation fund trustees should not be required to design or offer CIPRs because the trustees are the fund’s members.
Government would need to consider how the Age Pension means test applies to new income stream products. In principle, the means test should not discourage products that manage longevity risk, should aim to provide neutral treatment of products with longevity risk protection, and should not make it difficult for individuals to smooth their income and consumption over retirement. Without some amendments to the Age Pension means test, some CIPRs could increase the cost of the Age Pension to taxpayers.76
Design of CIPRs
People have different needs in retirement and will value the three desired attributes of retirement products (income, risk management and flexibility) differently. CIPRs should deliver a balance of these attributes. As no single product has all these features, a CIPR is likely to be a combination of products. A working group convened for the Inquiry by the Actuaries Institute recommends “… a portfolio approach is likely to be more suitable than a single default product. A sensible default might include an account-based product and another product with longevity risk protection”.77 Superannuation funds may work with life insurance companies, other funds or other entities to provide CIPRs.
Many people will live for several decades after retirement. CIPRs should therefore provide exposure to growth assets to increase retirement income. Rice Warner advises that “… any investment period of 20 or more years requires a significant proportion of growth assets.”78 Using CIPRs will allow superannuation funds to take a longer-term investment perspective and reduce the need for retirees to worry about sequencing risk.
Although CIPRs may include a combination of products, members should still be able to transition smoothly from the accumulation phase to the retirement phase.79 Cooling off periods coupled with the provision of a (diminishing) return of capital in the event of early death may be appropriate for some pooled products. These products could be purchased using either a one-off payment or a series of premiums.
CIPRs could vary with known characteristics of the member, including the size of their superannuation benefits. A trustee could decide to recommend lump sum benefits to members with balances below a certain (low) threshold.
Examples of CIPRs that would provide the required features are described in Table 5.
|Longevity product (a)||Allocation to longevity product||Draw-down of account-based pension||Allocation to account-based pension|
|CIPR 1||DLA||23%||Exhaust balance at age 85||77%|
|CIPR 2||Deferred GSA||17%||Exhaust balance at age 85||83%|
|CIPR 3||GSA||75%||Minimum rates||25%|
(a) Deferred products commence payments at age 85.
The annual income expected to be generated by each of the above CIPRs for a male retiring at 65 years of age with a superannuation balance of $400,000 (a typical balance in a mature superannuation system) is shown in Chart 5: Expected annual income from example CIPRs. The chart shows the amount of income he can expect to receive if he is still alive at each age. The income from an account-based pension drawn down at minimum rates — the most common strategy used by retirees at present — is included for comparison.81
The income streams represented in Chart 5 are only illustrative. They highlight the benefits of pooling and the ability to draw down an account-based pension faster without the retiree running the risk of outliving their wealth. The expected income from products is sensitive to assumptions regarding investment returns, draw-down rates and mortality.
Chart 5: Expected annual income from example CIPRs82
For a 65-year-old male with a $400,000 accumulated balance (excludes Age Pension)
Source: Australian Government Actuary modelling.83
Retirees using CIPRs would obtain significantly higher and smoother private retirement incomes while reducing the risk of outliving their savings (Chart 5 and Table 6). This is achieved through the loss of some flexibility and smaller bequests for some. As a portion of each CIPR is invested in an account-based pension, individuals retain some flexibility.
|Expected income throughout retirement (NPV)84||Increase over account-based pension85||Increase over account-based pension (%)|
|Account-based pension drawn down at minimum rates||$275,000||_||_|
Source: Australian Government Actuary modelling.87
54 Commonwealth of Australia 2010, Super System Review Final Report, Part Two: Recommendation Packages, Chapter 7: Retirement, Recommendation 7.1, Canberra, page 207.
55 Shadow shopping research conducted by the Australian Securities and Investments Commission (ASIC) found that only 3 per cent of financial advice about retirement was “good quality”. ASIC 2012, Report 279: Shadow shopping study of retirement advice, ASIC, Sydney, page 8. This study was conducted before the Future of Financial Advice reforms.
56 Longevity risk is not covered in the Australian Securities and Investments Commission’s compulsory RG 146 qualifications for superannuation advisers. This could be improved by raising competency standards, as recommended in Chapter 4: Consumer outcomes.
57 AustralianSuper 2014, Data provided to Financial System Inquiry, 10 September 2014.
58 See Benartzi, S 2010, Behavioral Finance and the Post-Retirement Crisis, Allianz of America, USA, for a discussion of these biases.
59 Plan for Life 2014, Data provided to Financial System Inquiry, 23 June 2014.
60 A measure of income from a given accumulated balance, ‘income efficiency’ is the expected present value of income in retirement as a percentage of a product’s purchase price. The income efficiency of a 65-year-old male’s account-based pension, drawn down at minimum rates, is around 70 per cent. Australian Government Actuary, Data provided to Financial System Inquiry, 11 June 2014.
61 The size of Australia’s annuity market is only around 0.3 per cent of gross domestic product, compared with 28.8 per cent in Japan, 15.4 per cent in the United States and more than 40 per cent in some European countries. Organisation for Economic Co-operation and Development 2013, ‘Survey of annuity products and their guarantees’, paper presented at the Insurance and Private Pensions Committee meeting, Paris, 5–6 December.
62 More than half of the respondents to a survey were either worried or extremely worried about outliving their savings. When asked to identify the single most important feature in a retirement income product, twice as many members identified “income that lasts a lifetime” as the second most popular response. Investment Trends 2013, December 2013 Retirement Income Report, Investment Trends, Sydney. Note: Based on a survey of 5,730 Australians aged 40 and older. Results from another survey suggest that more than 90 per cent of Australians over the age of 50 believe that “money that lasts my lifetime” is somewhat important or very important. National Seniors Australia and Challenger 2013, Retirees’ needs and their (in)tolerance for risk, National Seniors Australia, Brisbane, page 10.
63 The regulatory prescribed minimum rates range from 4 per cent for people aged 55 to 64, to 14 per cent for those over the age of 95.
64 Most retirees draw down their account-based pensions at the minimum allowable rates. Rothman, G and Wang, H 2013, ‘Retirement income decisions: take up and use of Australian lump sums and income streams’, paper presented at the 21st Colloquium of Superannuation Researchers, Sydney, 9–10 July, page 19.
65 Research suggests that these below-optimal rates result in lower welfare for individuals. Bateman, H and Thorp, S 2008, ‘Choices and Constraints over Retirement Income Streams: Comparing Rules and Regulations’, Economic Record, vol 84, pages S17–S31.
66 Although the life expectancy of a 65-year-old female today is about 89 years, 10 per cent of 65-year-old females will die before they reach 77 years and 10 per cent will live past 100 years. Even if individuals knew their life expectancy (which is generally not the case), the probability of a 65-year-old dying at a particular age is no greater than about 5 per cent. Commonwealth of Australia 2009, Australian Life Tables 2005–07, Canberra, using 25-year mortality improvement factors.
67 Australian Government Actuary modelling prepared for the Financial System Inquiry shows that CIPR examples 1, 2 and 3 described in the Implementation considerations section of this recommendation increase expected income in retirement by 14 per cent, 30 per cent and 31 per cent respectively, excluding income from the Age Pension. Australian Government Actuary, Data provided to Financial System Inquiry, 10 October 2014.
68 Johnson, E and Goldstein, D 2013, ‘Decisions by default’, in Shafir, E (ed.), The Behavioral Foundations of Public Policy, Princeton University Press, Princeton, pages 417–427.
69 Bateman, H, Eckert, C, Iskhakov, F, Louviere, J, Satchell, S and Thorp, S 2013, Default and 1/N Heuristics in Annuity Choice, School of Risk and Actuarial Studies Working Paper 2014/1.
70 Association of Superannuation Funds of Australia 2014, Second round submission to the Financial System Inquiry, page 101.
71 Deferred lifetime annuities are a form of lifetime annuity where the commencement of income payments is delayed for a set amount of time after purchase. In a GSA scheme, participants contribute funds to a pool that is invested in financial assets. Regular payments from the pool are made to surviving members. GSAs allow pool members to manage idiosyncratic longevity risk but do not completely eliminate the risks associated with increases in life expectancies across Australia. GSAs differ from a lifetime annuity by not providing a guaranteed income stream. Instead, the adjustments to payments over time are subject to investment performance, mortality assumptions and experience.
72 Benartzi, S, Previtero, A and Thaler, R 2011, ‘Annuitization puzzles’, Journal of Economic Perspectives, vol 25, no. 4, pages 143–164.
73 Australian Institute of Superannuation Trustees 2014, Second round submission to the Financial System Inquiry, page 43.
74 Sales of annuities fell significantly after both the reduction in the asset test exemption in 2004 and its abolition in 2007. Plan for Life 2014, Data provided to Financial System Inquiry, 28 March 2014.
75 For a discussion of these issues, see Commonwealth of Australia 2010, Australia’s future tax system: Report to the Treasurer, Part Two, Detailed analysis, vol 1 of 2, Canberra, page 122.
76 Under the principles of the current means test, products with longevity risk pooling tend to increase Age Pension costs in the early years of retirement (due to faster depletion of assets when the assets test is binding) and reduce costs in later years (because of higher income when the income test is binding).
77 Working group convened for the Financial System Inquiry by the Actuaries Institute 2014, Retirement Income: options for managing Australia’s longevity risk, Sydney, page 1.
78 Rice Warner 2014, Second round submission to the Financial System Inquiry, page 14.
79 For example, an income product provided by a life insurance company could be paired with an account-based pension in the same way accumulation accounts include life insurance.
80 These allocations were designed to smooth retirement income from the CIPR (excluding the Age Pension). In practice, retirees would want to smooth total income, including the Age Pension. This would alter the proportion invested in deferred products. The current Age Pension means test makes it difficult to smooth total income.
81 The full set of assumptions underlying these results and a sensitivity analysis are available in Australian Government Actuary 2014, Towards more efficient retirement income products: Paper prepared for the Financial System Inquiry, November 2014.
82 Produced using a stochastic model. The aim of achieving a relatively smooth income stream is affected by market and mortality variations.
83 Australian Government Actuary, Data provided to Financial System Inquiry, 10 October 2014.
84 Net present value, rounded to the nearest $1,000. Includes retirement income only (not bequests).
85 Rounded to the nearest $1,000.
86 A similar increase in retirement income could be achieved with a combination of two thirds of assets in an account based pension drawn down at minimum rates and one third of assets used to purchase a GSA.
87 Australian Government Actuary, Data provided to Financial System Inquiry, 10 October 2014.