Australia's Future Tax System

Final Report: Detailed Analysis

Chapter A: Personal taxation

A2. Retirement incomes

A2–3 Responding to increasing life expectancies

As people live longer, they will require more options to manage their assets over a longer period. The retirement income system will need to become more flexible so that it can provide these options.

Principles

Policies requiring a person to invest their superannuation in a particular product, or restricting access to lump sums, should only be adopted where there is strong evidence that people are unable to make decisions that are in their best interests.

The retirement income system should be flexible to allow for the development of products that allow people to manage better their retirement income.

The public sector's primary role should be to support the private sector in the development of these products. The public sector should only enter this market where the private sector is unable to meet the needs of the community and should do so only on an actuarially fair pricing of risk.

Responding to higher life expectancies

Australians have one of the longest life expectancies in the world. Advances in health technologies have resulted in a marked increase in life expectancies since the 1970s (see Chart A2–11). This trend is likely to continue.

Chart A2–11: Average increases in life expectancy(a)

Chart A2–11: Average increases in life expectancy(a)

  1. Chart shows increase in the life expectancy of males. Females have had a similar increase in life expectancy but live longer than males.

Source: ABS (2008).

On current trends, men aged 60 years in 2047 are projected to live an average of 5.1 years longer than those aged 60 years in 2007 and women an average of 4.7 years longer (Australian Government 2007a). The probability that at least one person in a married couple both aged 60 will be alive by age 80 or age 90, ignoring future mortality improvements, is 89 per cent and 47 per cent respectively (Rawlinson & Cater 2008). The retirement income system will therefore need to provide a more diverse range of products to allow people to manage a longer retirement better.

The Age Pension currently provides a basic income for people who have limited means. However, many people prefer to have the security of knowing they will always have an income above the Age Pension.

The most popular income streams in the Australian market are allocated pensions and annuities, which account for over 85 per cent of the total purchased income stream market (Investment and Financial Services Association 2007). Allocated pensions are account-based with the value of the account dependent on how much a person takes as income and investment returns. The length of time a person can draw an income from an allocated pension therefore depends on these two factors. For this reason, an allocated pension cannot ensure security of income on its own.

Chart A2–12 shows the example of a person who invests $150,000 in an allocated pension at age 67. Their preference is to take an income of $15,000 (indexed) a year from their allocated pension. Based on the age of their parents, they consider they may live to at least age 90. If they take an income of $15,000 a year they will exhaust the assets in their allocated pension by age 80. If they want their allocated pension to pay an income to age 90 they will have to reduce the income they take to $9,500 a year. This is an ineffective way to deal with longevity risk as below average investment returns may result in a person not achieving their goal even after reducing their income. It can also reduce a person's standard of living in retirement and result in them bequeathing more assets than they wish.

Chart A2–12: Account balance of an allocated pension over time

Chart A2–12: Account balance of an allocated pension over time

Note: This assumes an initial account balance of $150,000 and an average return of 6 per cent per annum. The initial income is indexed to CPI at a rate of 2.5 per cent per annum.

Source: Treasury estimates.

Products are not available in the market to cover the broad range of preferences of retirees in achieving security of income. This is a structural weakness in the Australian retirement income system.

The main product on the market that does achieve this security of income is a guaranteed income for life. However, this product is unpopular among retirees, with only 374 sold in 2007 (Plan for Life Research 2007). This is due to a number of factors, including that these products are not seen as good value. Currently, their price is high due to information asymmetries between purchasers and product providers, as well as a lack of tools available to providers to manage the risk associated with longevity insurance. For example, long-dated bonds do not exist that would assist in the management of the investment risk associated with these products.

Given the diverse preferences of retirees, a single product is unlikely to satisfy all people who wish to manage their longevity risk. This suggests a need for product innovation within the Australian market.

Submissions state that providers have been reluctant to develop new products for the Australian market due to the prescriptive rules that set out what an income stream is. These rules were designed to ensure that the earnings tax exemption on superannuation pension assets supports only products that deliver a genuine income stream.

One argument is that the rule requiring a minimum payment to be made from a pension every year does not cater for deferred annuities. The potential for changes to these rules can increase the uncertainty faced by providers. Submissions also claim there is a lack of coordination between the regulators of the income stream market. It is argued that these issues increase the risk, and therefore cost, of developing products.

Findings

The increasing life expectancies of Australians will require a greater choice of retirement income products that can cater for the different needs of individuals in retirement.

There are not enough products that guarantee an income for the whole of a person's retirement. This is because the industry lacks tools to manage risks associated with these products, such as long-dated bonds. Also government rules restricting development of income-stream products and uncertainty about future changes in these rules inhibit product innovation.

Housing as a form of longevity insurance

Accessing the equity in the family home is another way people may choose to achieve a higher standard of living in retirement and to protect against longevity risk. Reverse mortgages allow a person to borrow against the equity in their home with the loan usually paid off when the home is sold or the home owner dies. The reverse mortgage market in Australia is still developing but has been growing over time (see Table A2–2). An alternative product has recently been introduced where a person sells a proportion of their home to an institution. The institution then has a right to claim that proportion of the proceeds on the sale of the house.

Table A2–2: The reverse mortgage market in Australia

  December 2005 December 2006 December 2007 December 2008 June 2009
Outstanding market size ($b) 0.85 1.51 2.02 2.48 2.61
Number of loans 16,584 27,898 33,741 37,530 38,048
Average loan size $51,148 $54,233 $60,000 $66,150 $68,473

Source: SEQUAL/Deloitte (2009).

Payments from a reverse mortgage on a primary residence are not treated as income for tax purposes as they are considered to be a loan. The exemption of the owner-occupied home from the income support means test may discourage people from undertaking home equity conversions as it would convert an exempt asset into an assessable one. To counter this, special means test arrangements apply. The first $40,000 of a reverse mortgage paid as a lump sum is exempt from the assets test for 90 days. Amounts over $40,000 are assessed under deeming rules if held as a financial asset. If taken as a stream of payments, the amount drawn down is not counted in the income test. The tax and means test treatment of these products is already generous and should not be made more so.

Role of longevity insurance products

Recommendation 21:

The government should support the development of a longevity insurance market within the private sector.

  1. The government should issue long-term securities, but only where this is consistent with its fiscal obligations, to help product providers manage the investment risk associated with longevity insurance.
  2. The government should make available the data needed to create and maintain a longevity index that would assist product providers to hedge longevity risk.
  3. The government should remove the prescriptive rules in the Superannuation Industry (Supervision) Regulations 1994 relating to income streams that restrict product innovation. This should be done in conjunction with the recommendation to have a uniform tax on earnings on all superannuation assets.

Recommendation 22:

The government should consider offering an immediate annuity and deferred annuity product that would allow a person to purchase a lifetime income. This should be subject to a business case that ensures the accurate pricing of the risks being taken on by the government. To limit the government's exposure to longevity risk, it should consider placing limits on how much income a person can purchase from the government.

The development of longevity insurance products is another means — along with the taxation of superannuation and the funding of health and aged care — of improving the adequacy of the retirement income system. If a person knows they can rely on a particular level of income to support them until they die, they can make better decisions on how to manage their assets over their retirement.

The retirement income report set out some issues relating to longevity insurance that the Review Panel wanted to consider in this Report. These were whether the product should be:

  • mandatory or voluntary;
  • provided by the private sector or public sector; and
  • guaranteed or non-guaranteed.

Mandatory or voluntary

As long as the Age Pension continues to provide a longevity insurance safety net, it is not necessary to impose a requirement that people invest in additional insurance.

A reasonable basis for policy design is the presumption that, having accumulated retirement savings, people are generally in the best position to determine how they use their assets during their retirement. Some people may prefer a higher standard of living at the beginning of their retirement, with high draw-downs from their superannuation during this time, before relying on the Age Pension later in their life. Other people may prefer a stable and secure income over their entire retirement. A voluntary system ensures that both these groups can insure up to the level of income they want over their retirement.

Some submissions suggest that people should be required to use part of their superannuation to purchase longevity insurance. Such a requirement could help to overcome one source of market failure in longevity insurance markets related to access to information. These markets may fail to yield efficient outcomes because a person may have more information on how long they are likely to live than insurers do. This may mean that the only people who purchase these products are those who consider they are likely to live longer than average. Insurers can react to this 'information asymmetry' either by not selling the products or by pricing them at a level that discourages most people from purchasing them. This is one of the reasons for the unpopularity of life annuities.

A mandatory system would remove this market failure by ensuring that the people in the insurance pool reflect the average life expectancy within the community as a whole. This would allow insurers to sell these products at a lower price because the capital of the people who die early in their retirement supports those people who live for longer. Research by the University of New South Wales confirms this effect (Sherris & Evans 2009). Table A2–3 shows the first annual payment of an annuity in a mandatory and voluntary system.

Table A2–3: Income differences between mandatory and voluntary annuitisation

Type of annuity First annual payment (a)
($)
Compulsory annuitisation, immediate annuity (b) 8,965
Voluntary annuitisation, immediate annuity 7,942
Compulsory annuitisation, deferred annuity (c) 71,408
Voluntary annuitisation, deferred annuity 44,181
  1. The payment is based on an annuity purchased by a male from the private sector with $100,000. Longevity is based on the improvements to mortality that have occurred in the past 5 years. It assumes no indexation and is valued using the end June 2009 yield curve from government bonds quoted on Bloomberg. The values are in nominal dollars.
  2. The immediate annuity commences at age 67.
  3. The deferred annuity is purchased at age 67 and commences at age 85.

Source: Sherris and Evans (2009).

However, overcoming this market failure by mandating the purchase of longevity insurance can come at a cost to people who are in poor health or have lower life expectancies, such as Indigenous Australians. Such people would be disadvantaged by a mandatory system as they would effectively be subsidising people who live longer than average. It may be possible to provide a different pricing structure for these people to accommodate their lower life expectancies. However, even if these arrangements existed, a mandatory system is likely still to be seen as punitive to these people.

Another argument for a mandatory system is that it would reduce the risk that people exhaust their assets quickly in order to receive an Age Pension. However, the research by Lim-Applegate et al. (2005) suggests that people in retirement are conservative in how they draw down their assets. This may be as a result of them attempting to self-insure against longevity risk.

On balance, the Review does not recommend a mandatory system for longevity insurance. A mandatory system would constrain the ability of people to make their own decisions on how they use their superannuation to fund their retirement. The evidence suggests that people make conservative decisions on how they use their assets in retirement. Also, the existence of the Age Pension already provides longevity insurance for a significant proportion of the population.

Findings

There are some arguments for requiring people to purchase longevity insurance. These include addressing information asymmetries that exist between the purchaser and provider and ensuring that at least part of a person's superannuation is used for retirement.

However, a mandatory system would have a detrimental effect on people with lower than average life expectancies. The Age Pension will continue to provide longevity insurance for the majority of retirees.

Public or private sector

The Review Panel considers that the development of a voluntary market for longevity insurance will require input from the private and public sectors. The private sector is in a better position to develop products that best meet the preferences of individuals. However, the public sector may be in a better position to deal with the significant counterparty risk associated with longevity insurance. The public sector could also provide more tools to assist in the development of longevity insurance products.

International experience

In other countries, new products are being developed that provide an alternative to typical annuity products where a person purchases an income for life. In the United States, a person can purchase a guaranteed income based on the value of an investment account similar to an allocated pension from a set time. These products differ from typical annuity products in that the amount of income and the start date are not fixed. While the minimum amount of income is fixed, it can increase depending on investment returns. The income only commences when the value of the account falls below the income guarantee. The level of guarantee also tends to be supported by hedging arrangements by the provider. This differs from typical annuities, which are supported by capital that the provider must hold under prudential regulation. A similar product has recently been introduced into the Australian market.

Deferred annuities, which provide an income from a certain age, are also becoming more prevalent. These annuities allow a person to lock up part of their retirement savings to generate an income when they are entering the latter stages of their retirement. This provides a person with more certainty in how they manage the rest of their assets before the commencement of the deferred annuity.

The preferences of retirees, advancements in technology and risk management techniques will continue to affect the development of longevity risk products into the future. The private sector is likely to be able to respond to these factors more quickly than the public sector.

The role of the government

The government can assist in creating an environment where the industry has greater flexibility and confidence to develop longevity insurance products.

Removing restrictive rules

A product must comply with certain rules to be treated as a superannuation pension or annuity. The prescriptive nature of these rules, such as a requirement for specific annual payments and limits on indexation, can constrain product development.

The recommendation to tax all superannuation fund earnings (whether or not they support an income stream) at a uniform lower tax rate removes the concession these rules were protecting (see Recommendation 19). Therefore, implementing the earnings tax recommendation would provide the opportunity to remove these rules. The removal of these rules would provide greater scope for innovation in the income-stream market and enable product providers to get products into the market more quickly. The current rule that caps the amount of payment from a transition to retirement pension should continue, however, as this protects the integrity of the preservation rules.

The removal of these rules would also make redundant the current income test assessment of superannuation pensions and strengthens the case for deeming income on account-based income streams, so they are treated like other financial assets. The proposal for means testing superannuation pensions is at Section F2 Means testing.

In many cases, people may choose not to purchase longevity insurance at their retirement age. As they grow older they may be in a better position to judge their potential longevity. However, after a person retires they may be unable to make further contributions into a superannuation fund due to the work test rules. These restrictions should not apply to contributions made to a prudentially regulated superannuation fund or life insurance company for the purpose of purchasing a longevity product (see Recommendation 20).

The government should also consider removing other legislative constraints that may inhibit the development of longevity products. However, this should not be at the cost of necessary prudential or consumer protection. Given the nature of these products, they should only be provided by prudentially regulated entities. Products that provide a guaranteed income should follow consistent prudential requirements to reduce the risk that a provider is unable to meet their obligations as they fall due.

Coordination between the regulators

Another concern raised with the Review is that there is little coordination between the various regulators of income-stream products. These regulators are: the Australian Taxation Office (ATO); the Australian Prudential Regulation Authority (APRA); the Department of Families, Housing, Community Services and Indigenous Affairs (FaHCSIA); and the Australian Securities and Investments Commission (ASIC).

Most concern was raised about the interaction between the ATO and APRA, especially in relation to whether a product meets the definition of a pension or annuity and is therefore eligible for tax concessions. While the ATO administers these concessions, it is unable to advise product providers whether their product meets the definition of a pension or annuity as this definition is in the Superannuation Industry (Supervision) Regulations 1994, which are administered by APRA. This definition would no longer exist if all superannuation fund earnings are taxed at a uniform rate. This would also remove the need for the ATO to be involved in the regulation of income streams.

Better management of risks

The providers of longevity insurance must deal with a number of risks in bringing a product to the market. These include investment risk, inflation risk and longevity risk.

Investment risk relates to the long-term nature of these products. In guaranteeing a future income, the provider assumes an average rate of return over the period of the guarantee. The more closely a provider can match their long-term liabilities with long-term assets, the lower this risk and the lower the price at which the product can be sold.

Inflation risk affects both the purchaser and the product provider. Inflation reduces the spending power of a person's retirement income over time. People can purchase income streams indexed to inflation to counter this risk but this increases the price of the product. Inflation risk also increases the costs of providing long-term products.

Longevity risk relates to the likelihood that the provider will have to meet pension obligations for longer than expected due to an unanticipated increase in life expectancy in the community.

Government policy can help product providers manage these risks and reduce the price of longevity insurance. In particular the government could issue a broader range of debt instruments, such as indexed and long-dated bonds. The availability of these assets would greatly assist in the development of a longevity insurance market by reducing investment and inflation risk.

The Australian Office of Financial Management announced on 7 August 2009 that it would resume issuing Treasury indexed bonds. The government should also consider issuing longer-dated bonds where this is consistent with its fiscal obligations.

The Review Panel does not consider that the government should issue longevity-indexed bonds to encourage the development of a longevity insurance market. The government already takes on the overwhelming majority of longevity risk through the Age Pension. Longevity bonds would increase the government's exposure to this risk. This is consistent with the findings of the OECD (2007b), which found the prospects for a successful, large-scale market in longevity-indexed bonds did not seem favourable due to the already high level of longevity risk already on government balance sheets.

A longevity index shows the number of years that, on average, a member of the population at a particular age is expected to live. The index can be used to establish a market in which providers can hedge part of their longevity risk. Such an index has been set up by J.P. Morgan for the United Kingdom market (known as the LifeMetrics index). The government could help to develop a longevity index by making available the data necessary to create and maintain one (see Recommendation 21b). This is consistent with an OECD (2007b) finding that governments, through their national statistical institutes, could help private market participants produce longevity indices. Sherris and Evans (2009) suggest that the government would be in the best position to produce such an index.

Tax and means test concessions

There have also been calls for the government to provide tax or social security concessions to encourage people to purchase longevity insurance. Specific concessions for longevity risk products are not supported as they could distort the market.

If specific concessions did exist, there would need to be rules setting out the characteristics of these products. The government would effectively need to approve new products that may fall outside these rules by making legislative changes. The longevity insurance market is likely to be very innovative. Placing legislative restraints on product design would be an unnecessary and costly constraint on innovation.

However, given the unique nature of deferred annuities, there is a case that they should only be means tested when they start to pay an income, unless a person can access the capital before this time. Further details on the proposed means test treatment of superannuation pensions and annuities are in Section F2 Means testing.

Government-provided products

In purchasing a longevity risk product, the purchaser is taking on the risk that the provider will be unable to meet their obligations. Prudential regulation provides some protection against this risk. However, other factors could affect this counterparty risk.

One of these risks is systemic longevity risk, where advances in health research or changes in lifestyle result in unexpected increases in life expectancy for the entire community. Significant changes in life expectancy may go beyond what is catered for in prudential regulation. This might affect a provider's ability to meet its current and future obligations.

The government should consider whether these risks are such that it should enter the market and sell products that provide a guaranteed income stream (see Recommendation 22). For example, the government could use the existing Age Pension infrastructure to allow a person to purchase an immediate annuity. The government should also consider selling a deferred annuity that, if purchased, would give retirees greater certainty over the period they have to draw down their assets. People should be able to purchase these products with superannuation as well as non-superannuation money.

Several submissions state that the government should not sell these products. They argue that the government may not provide these products at an actuarially fair price. This might result in low-income households who do not purchase this product subsidising higher-income households who are more likely to do so. This may also result in the private sector leaving the market (which could result in less product innovation).

As the government already takes on the majority of longevity risk through the Age Pension, if it were to offer these products it should limit the amount of additional longevity risk it takes on. It could, for example, limit the value of the annuity and place a cap on the amount a person could invest in a deferred annuity. The government would need to develop an appropriate business model that would ensure the products are sold at a price that accurately reflects the risk the government would be taking on.

Findings

The development of a longevity insurance market will require involvement by both the private and public sectors. The private sector is better placed to develop products that meet the needs of retirees. The public sector can assist in developing these products by providing more tools that the private sector could use to limit the risks associated with the products.

The public sector may be better placed to deal with the counterparty risks that exist with these products. However, the government would be taking on more longevity risk by entering this market.

Guaranteed or non-guaranteed

Products can either be guaranteed by the provider or non-guaranteed. The income from a non-guaranteed product would depend on the investment returns on the assets supporting the pool and the mortality experience of the people in the pool.

The government should not restrict the types of products that could be sold. Longevity insurance should form part of a portfolio of products people can use to finance their retirement. Placing restrictions on products, such as requiring them to be guaranteed or non-guaranteed, reduces the potential range of products that could be included in this portfolio.