The Age Pension Means Tests: Contorting Australian Retirement

Abstract

Most Australian retirees are likely to be subject to the Age Pension assets or income test at some point. Evidence is that many retirees adapt their consumption to increase Age Pension entitlements, but long-term implications are difficult to determine—even if the current rules were to remain in place. This chapter evaluates the current approach to means testing against the principles set out in a Department of Social Services discussion paper on this topic. We evaluate the implied “effective marginal tax rates” (EMTRs) on the assets of part pensioners who are subject to the assets test. We find that depending on a variety of parameters such as assumed future earnings rates, demographic status, drawdown strategy and the base level of assets held, the EMTRs are high enough to explain material distortions to savings decisions of those still in employment, and the spending and investment decisions of retirees. Optimal decisions in this context require contorted retirement strategies that do not appear to be in anyone’s interest. Some possible remedies are suggested, which should include incorporating the value of the principal residence within the assets test. The chapter therefore illustrates the application of principled analysis to policy issues of this sort.

1. Introduction

This chapter evaluates the means tests that apply in Australia against alternative principles that can be applied to evaluate the integration between the first and second pillars of a retirement income system. Central to the analysis are the Effective Marginal Tax Rates (EMTRs) that apply to the first pillar when subject to the means tests.

2. The Age Pension and means tests

The Australian retirement income system consists of two main pillars: the means tested Age Pension (AP) and Superannuation [1]. The first pillar is a welfare benefit that is means tested and funded from general government revenue. The superannuation system consists mainly of privately administered funds in which members have investment accounts into which mandated contributions are paid. Members are able to take tax free lump sums from the age of 60, but many choose to withdraw “account-based pensions”, which allow members to vary the amount drawn—subject to a regulated minimum proportion of their accounts. A third pillar consists of voluntary savings including the home.

There are currently two separate components of the means tests, the income test and the assets test, and homeowners are treated differently from non-homeowners. The means test rules are Byzantine in their complexity and change frequently—sometimes with considerable effects on some groups of pensioners. The following description omits many details on the grounds of immateriality.

2.1 The Age Pension

The AP dates back to the formation of Australia with Federation in 1909, the amounts being shown in Figure 1. It has increased roughly in line with average wages, with a particular boost under the Whitlam government of the early seventies. Relative to minimum wages, it has approximately doubled over the past century, which means that pensioners are better off relative to workers on lower incomes. This shift has partly been a response to old age poverty, but may also reflect the relative political influence of these two groups.

Figure 1.

History of AP and means tests—couples.

2.2 The means tests

Figure 1 also traces the development of the income and asset test free limits over the century. The two tests were merged between 1961 and 1985. Before 1969, pensions were not payable to those over the means test free limits, although between 1953 and 1971 assets were converted into income at a rate of 10%. From 1969, taper rates (the rate at which the AP is reduced as a proportion of the relevant means) of 50% have applied on income except for a period from 2000 to 2009 when they were reduced to 40%. The taper rate on the asset test seems to have been set at 7.8% except from 2007 to 2016 where it was only 3.9%.

2.3 Description of current means tests

The current AP for a married couple owning their own home is slightly in excess of $36,000 per annum, reduced by the greater of 50% of income over $7800 p.a. or by 7.8% of assets over a threshold of $394,500. Single pensions are a relatively generous 66% of those paid to a couple.

Those who are renting qualify for rental assistance of $3588 for singles but—strangely—only $3380 if a couple. The asset test limits for non-homeowners are also some $200,000 higher than for homeowners.

The Australian means tests appear to be particularly complex and Australian pensioners on a wider range of incomes and with a wider range of assets are affected by the taper rates under the means tests than pensioner residents of other countries [2].

The Department of Social Services (DSS) explains and justifies the means tests thus: “The means test functions to assess a person’s overall capacity for self-support and target social security expenditure according to need. … Means testing also provides incentives for self-provision in the form of participation and saving. The means test balances these objectives by the use of income and assets free areas and the tapered withdrawal of payment as a person’s assessable income and assets increase” [3].

2.4 Evaluation criteria

The DSS discussion paper [3] also lists the criteria against which the current means testing system should be judged. Those principles need to encompass the treatment of a variety of types of asset (especially annuity and deferred annuity products) as well as financial and other investments and physical assets.

• “Neutrality—the means test assessment of investments should not advantage a particular type of product or provide an incentive for people to invest in a particular asset …

• Equity—the rules should treat people with similar means in a consistent way (horizontal equity) and those who have a greater capacity to self-provide for their retirement should receive lower income support (vertical equity).

• Resilience—the rules should be able to apply to a range of products, including new products, without diminishing neutrality and equity …

• Integrity—the rules should ensure the social security system remains targeted to assisting those people who need support and that people cannot maximise their Age Pension by engaging in strategies that minimise the extent to which their own income or assets are counted in means test assessments.

• Fiscal sustainability—the means test treatment of new retirement income stream products should have regard to the cost of the social security system.

• Simplicity—the rules should be easy to understand for income support recipients, financial advisors and income stream providers. Complicated rules can result in people making poor financial decisions. Simple rules support people to make good decisions.”

We have no argument with these principles. This chapter, however, argues that the asset test is especially inequitable to the cautious who retain assets for precautionary purposes; makes planning extremely complex; and significantly distorts the investment and drawdown behaviour of many retirees. It therefore fails all the policy principles except possibly that of fiscal sustainability. There are alternative approaches that could better meet these principles including fiscal sustainability.

3. Housing

While not the main point of this chapter, the difference in treatment between home-owners and renters is egregiously inequitable, and needs to be discussed first as it would not be complex to address, and the adjustment could be performed in a manner which would be revenue-neutral.

3.1 Poorer renters

The pressing issue is the dire position of those who neither own their own home nor have access to social housing. Given that only 10% (or about 200,000) of such properties had rents of less than $10,000 annually in 2011 [4], rent assistance of $3588 p.a. places considerable financial stress on those non-homeowners with limited other resources, many of whom are reduced to poverty or homelessness. Unpublished data from the Australian Department of Social Services shows 32,000 homes, where people are aged 65 and above and receiving rent assistance, are paying unaffordable rents in NSW [5]. It is a 50% rise in the past five years and includes 9000 people, 65 and over, who are paying more than half their income on rent.

Given that there are some 1.2 million people over 65 in NSW, this implies that about 2.5% of the aged population face the problem—90,000 country-wide. We make no detailed calculations as to precisely what would be fair, but suggest that the amount should be more or less doubled. While $6000 p.a. is not enough to fully cover rent, homeowners do face significant costs that would have to be considered in detailed modelling. Doubling rent assistance to pensioners not in social housing would cost less than $200 m, which is close to 0.25% of the cost of support to the aged.

3.2 Richer homeowners

At the other end of the financial spectrum, the exemption of the home from the asset tests means that there may well be as many pensioners receiving the AP while living in multi-million dollar houses. Particularly given that dwellings and land represent over 50% of the household balance sheet, an exemption seems myopic and distorts financial planning, spending, investment and the price of housing [6, 7]. The DSS discussion paper justifies it so:

“The exemption of the principal residence from the assets test recognises the greater financial security that pensioners have if they live in their own home and importance placed on the family home in Australian society.”

While these assertions are clearly true, the rational conclusion would be to include homes in the assets test rather than exclude them precisely for these reasons. Home ownership does provide financial security so those who rent need to be given more, while those who obviously have secure lodgings need less social support. The “importance placed on the family home” emphasises this. For many, however, the “family home” has three empty bedrooms and is crumbling around the surviving parent, who is increasingly unable to manage the activities of daily living, let alone maintain the house. But the homeowner is not willing to downsize to a less expensive and more manageable property, sometimes for psychological reasons but often also because of the negative impact on their main source of income, the AP.

Each of the criteria mentioned in Section 2.4, except perhaps simplicity, demand the inclusion of the principal residence. The arguments against ignore the disadvantages faced by renters. Homes provide an imputable rent and the AP should be adjusted if pensioners decide to spend more on it.

The need to include the value of the house in the asset test has been recommended by almost all who have considered the question. There have been eight different recommendations in as many years: the Harmer Pension Review, the Henry Review, a Grattan Report, three Productivity Commission reports, Rice-Warner and the National Commission of Audit [8]. It seems overwhelmingly clear that affordability issues should be addressed by including the value of the home.

3.3 The right to housing security and fungibility

There are two reasons to distinguish one source of wealth from another. The first is the basic human right of housing security; the second is fungibility. These two can explain much of the visceral response of some people when suggestions are made to include the home in the means tests. The right to housing has a simple logic: if people do not have their own place, they must occupy someone else’s. This right must include some security of tenure to be meaningful, and without some security and protections for both renters and owners the incentive to care for and improve the place is absent. Addressing the lack of security enjoyed by renters is the subject of another debate. Those interested in some of the theoretical and political controversies might like to refer to Hayward [9].

The problem of fungibility is that the home cannot easily be turned into income. Conventional reverse mortgages are relatively expensive and, given the normal operation of compound interest, may lead to significant reductions in the net residual value of the home which might constrain future choices if the borrower wishes to move into other accommodation. There is also a government scheme (the “Pension Loans Scheme”) that lends the AP shortfall to part pensioners. The interest rates charged are a little lower than commercial reverse mortgages but more restrictions apply to eligibility and the manner of drawdown of funds.

If the value of the home is to be included in the means tests, then the lack of fungibility means that some pensioners at least will be worse off. Johnson and colleagues [8] refer to a growing consensus for effective instruments to address the issue of fungibility, and refers to some commercial schemes including those that share capital appreciation.

Careful consideration would need to be devoted to transitional arrangements so that inclusion of the home (or the excess of the value of the home above some threshold value) does not bear unreasonably harshly on those pensioners who have already retired, in the expectation that the current means testing rules will continue to apply (perhaps with slight adjustment to parameters) throughout their retirements.

4. Integrating the superannuation and AP systems

The key issue addressed by this chapter is the way by which the AP is reduced by financial assets—including those in superannuation funds. As will become clear in this section, the issue is not simply the “taper rate” (the rate at which the AP is reduced in relation to assets held above the assets test threshold), because the long-term effects are very different for different approaches to the drawdown of the superannuation balance.

We take it as given that the objective of the superannuation system is to provide for consumption in retirement, and that the AP supplements this for those who do not have the means to support themselves. One important consequence of this is that superannuation benefits are to be seen as a mechanism for supporting the lifestyle of retired Australians, not as a mechanism by which the real capital value of a pensioner’s estate is maintained or grown, to be passed to non-dependents. The current dominance of account-based pensions does not achieve this result, and we could argue that the pooling of longevity risk should be encouraged by the means tests. This is however not the main thrust of this chapter, which is the unfairness and distortions created by the asset test.

4.1 Concept of effective marginal tax rates (EMTRs)

In each case, we consider the marginal losses to AP entitlements that arise from having assets over the thresholds. We refer to these as EMTRs. Legally, the tapers in the AP are not “taxes”, but given that money is fungible, the loss of social security entitlements caused by having additional assets is effectively the same as a tax.

4.2 Those who do not draw down

Figure 2 illustrates EMTRs for those who retain their assets and spend their income. It applies at any one time, and would apply over a lifetime to a single person and show the results of earning a real rate of interest of 3% and another earning 5%. The assets test creates a trough in income between about $300,000 and $700,000 in assessable assets. Within the assets testing range where the assets test produces a lower AP entitlement than the income test, annual income declines with increasing assets because the income from the marginal assets (whether returning 3% p.a. or 5% p.a.) is less than the AP foregone, which is 7.8% p.a. The calculations do not consider any investment volatility.

Figure 2.

The asset trough for those who retain assets.

4.3 Definition of “lifetime EMTRs” and approach to calculation

A more complete picture may be obtained by considering the impact of holding additional assets over the course of a retiree’s lifetime. We do this for a variety of asset levels and drawdown strategies. We also vary the following:

• Assumed future real rates of investment return on assets

• Drawdown strategy adopted

• Demographic status (single male/single female/couple)

• Initial level of assets held

For purpose of our calculations it may be helpful to define “lifetime EMTR” as follows:

$$\text{Lifetime EMTR}=\left\{\mathrm{PV}\left(\mathrm{AP};\text{\$}\mathrm{X};\mathrm{DS}\right)–\mathrm{PV}\left(\mathrm{AP};\text{\$}\mathrm{X}+\mathrm{\delta };\mathrm{DS}\right)\right\}/\mathrm{\delta }$$E1

where $X is the base level of assets held when the pensioner first becomes eligible for AP; $δ is an incremental amount of assets in addition to $X hypothetically held when the pensioner first becomes eligible for AP; DS refers to a specified Drawdown Strategy; PV (AP; $X; DS) is the present value of AP received over the lifetime of the single of couple pensioner assuming initial assets of $X and that the pensioner follows drawdown strategy DS.

In other words, the lifetime EMTR is the present value of the future loss of AP due to holding additional assets $δ at age pension eligibility age, as a percentage of those additional assets. A lifetime EMTR of zero means that the pensioner has incurred no loss of AP as a result of holding an additional $δ of assets, while a lifetime EMTR of 100% would mean that the pensioner has effectively received no net value from the additional $δ of assets.

For some of the results in this chapter (especially in Section 5.1 below) we have taken $X to be the lower assets testing threshold and $δ to be the width of the asset testing range (i.e. from the threshold to the level of assets where the AP is zero). The EMTR is therefore the impact of holding additional assets equal to the width of the asset testing range, relative to holding assets of an amount equal to the lower assets test threshold. We have also performed a set of calculations for smaller tranches of incremental assets, for drawdown strategies in the following specific forms:

1. Asset drawdown each year is the regulated minimum required as a proportion of the beginning of year assets. We have also calculated EMTRs for annual consumption expenditure from 1% to 5% higher than the minimum—in 1% increments.

2. Asset drawdown each year is a fixed percentage of the beginning of year assets. The fixed percentages for which we evaluated EMTRs were varied from 5% to 15% in multiples of 2.5%.

The asset testing range for a single pensioner is approximately $300,000 in total width (from $253,750 to approximately $552,000). We have taken $δ for single pensioner calculations to be $100,000, and have therefore calculated three separate EMTRs for the asset testing range. Similarly, the asset testing range for pensioner couples is very nearly $450,000 in total width and we have taken $δ for calculations involving couples to be $150,000.

In total we have done six computations of EMTRs for each drawdown strategy, demographic status and rate of real investment return considered, reflecting marginal lifetime losses of AP from holding an additional $100,000 of assets (for singles, and $150,000 for couples) at the date of becoming eligible for the AP. Three of the calculated EMTRs apply within the asset testing range and three to assets in excess of the asset testing range but where the additional assets held will result in the loss of AP not immediately, but after a period of time when the assets are depleted to an amount within the asset testing range.

Table 1 shows the parameters for each of the six sets of calculations.

Table 1.

Asset ranges.

The results of these detailed EMTR evaluations are reported in Section 5.2 below.

4.4 Assumptions

Table 2 summarises the assumed AP parameters that we have used for the purposes of the detailed calculations in this chapter. (The parameters in Table 2 were those that applied between September 2017 and March 2018; however we note that most of the parameters below change only gradually over time, due to indexation).

Table 2.

AP projection assumptions.

Also implicitly we have assumed, by using a real rate of investment return (where by “real” we mean relative to average weekly ordinary time earnings—AWE) that future rates of AP payment will be indexed to AWE and also that future lower and upper thresholds for the assets test and income test will also be indexed to AWE. Current practice is for AP payment amounts to be indexed normally to AWE but threshold amounts are indexed to CPI; nevertheless for simplicity of calculation, both payment amounts and thresholds have been assumed to be indexed to AWE. If we had used CPI, the EMTRs reported would still apply, but at slightly different levels of assets.

5. Results

The results below show first the extent to which alternative strategies can contribute to significantly higher AP benefits over a lifetime. Then, in Section 5.2, we show the EMTRs that apply to retirees subject to different circumstances when they use the alternative strategies.

5.1 Annual income and division of benefits

Figure 3 shows the lifetime effects (until age 109) of the various strategies for a single pensioner. We have assumed assets exactly equal to the top of the asset test level, which produces the largest impact—for illustration purposes.

Figure 3.

Alternative retirement consumption patterns.

The first strategy is for the pensioner to draw down the minimum permitted by superannuation regulations. Her spending is shown as the unbroken Minimum drawdown line in Figure 3. This is the strategy we understand is adopted by approximately half of pensioners, more than any other strategy [10]. Its presence in regulations as a minimum gives it an implicit authority and we have taken it as an effective benchmark. Her spending pattern is however totally inappropriate—starting at $28,000 p.a. at 67 and rising to $48,000 at 85 before declining again. We have also calculated the actuarial present value of the benefits to the member, her dependents and as a reduction in AP. The value of her assets at the beginning of retirement is shared into roughly equal thirds: one to herself, one to her heirs, and one in fiscal savings. That means tests are structured to permit the heirs to inherit a third of the superannuation benefit represents a lack of targeting of superannuation and the means tests.

A second strategy is to give away enough of her assets to escape the asset test altogether, which the regulations require her to do before 62—five years before becoming entitled to the pension. She then draws down at the minimum rate. This dashed line shows somewhat bizarrely that, at 67, her income is $6000 p.a. (20%) higher than if she had not given the money away. This amount falls slightly after her mid-eighties, but it is reasonable level and appropriate. Her share of the benefits drops only slightly to 29%, because the savings to the fiscus is minimal, with its proportion dropping to 2% (via deemed income caught by the income test). The heirs’ share more than doubles to 69%—$260,000 of which they receive before the pensioner reaches the age of 62. This is by far the most fiscally efficient for the pensioner and family. Rather than be given away, the money could be invested in her residence or spent before the pension age. The alternatives, singly or in combination, would allow the pensioner to directly enjoy 98% of her savings.

A third alternative is for the pensioner to spend the money after retirement in order to maintain the “comfortable lifestyle” determined by the Association of Superannuation Funds of Australia (ASFA) [11] until she falls below the asset test threshold, after which she reverts to the minimum. The income is shown in the Comfortable line in Figure 3. She enjoys a 50% higher standard of living than the benchmark, and remains above it until her early eighties. Her share of her savings rises to a little over 50%, with her heirs and fiscal savings sharing the balance.

A final alternative therefore is a level lifetime annuity. The total income and the amount provided by the annuity plus Age Pension are shown by the dotted and dashed lines. The present value is shared 70% to the pensioner and the balance to the fiscus.

5.2 Analysis of EMTRs for small tranches of incremental assets

In addition to the broad calculations of EMTRs referred to in Section 5.1 above, we have computed EMTRs for smaller values of δ ($100,000 for singles and $150,000 for couples). The objective of these calculations was to gain additional insight into the extent to which asset-testing rules impacted EMTRs as a function of base asset levels, investment returns and drawdown strategy combinations.

For these more granular calculations a slightly different approach was taken to mortality assumptions than in the calculations underlying Appendix 1 and Section 5.1. In particular it was assumed that a single female would receive AP for 25 years and a single male would receive AP for 22 years. These durations are slightly higher than the life expectancies for a 65-year-old according to Australian Life Tables (ALT) 2010–2012, which are 22.05 and 19.22 years for females and males respectively. Ongoing improvement to mortality means that current age-specific mortality in Australia is lower than mortality rates that applied during the 2010–2012 triennium and is likely to continue to decline in future.

For couples, it was assumed that the couple rate of pension would be payable for 22 years and the single rate for a further 5 years. This is because the expected age at death of the last to die from a couple is greater than the expected age at death of either considered separately.

The drawdown strategies allowed for are those detailed in Section 4.3 above.

Figure 4 shows EMTRs for a single pensioner using the minimum drawdown strategy we use as a benchmark.

Figure 4.

EMTRs based on drawing down the regulated minimum—single pensioners.

The curves show the EMTR of a marginal additional of $100,000 of assets at the eligibility age for the AP, valued over the lifetime of the pensioner. We note from Figure 4 that female EMTRs are slightly higher than the corresponding male EMTRs, due to the longer life expectancy of females. This means that they have more time to lose AP to the means tests.

The effect of different investment earnings is significantly greater than those for the two genders. For example, the 5% real interest curves fall away more steeply for base asset level of $553,000 or more (that is, at or above the upper assets test threshold). This occurs because the 5% real interest rate will tend to mean that, if assets are at or above the upper asset test threshold when the pensioner attains eligibility age, then the value of assets will remain above the upper threshold over the entire lifetime, and no further AP is foregone. However if assets are within the asset testing range at AP eligibility age, they will remain between the thresholds for longer so increasing the loss of AP. At lower rates of real interest, the asset values are depleted more quickly over time with the result that less AP is lost at lower starting balances. Assets over the upper threshold depreciate faster, so more is lost from an incremental $100,000.

Finally, the most significant observation is just how high the EMTRs are, throughout the asset testing range (especially in the middle and upper thirds of the asset testing range). Top marginal income tax rates in Australia are currently only 47%, so any EMTR in excess of this appears unfair and provides an incentive to adopt means test avoidance strategies. EMTRs in the range of 80%, 90%, 100%, or even higher represent a very powerful incentive to:

• Give money away 5 years prior to attaining eligibility age;

• Spend down rapidly in the first few years on pension;

• Spend money on house renovations, or even purchase a more expensive property; or finally,

• Not save in the first place. This can involve higher borrowing at younger ages to be repaid by superannuation lump sums, or retire early and draw them down before qualifying for the AP.

All of these are obviously distortions to the economic decisions that the pre-retiree would otherwise have made. Those who adopt such strategies will also be an increased burden on the public purse.

6. Evaluation

While the results will vary significantly depending on the underlying assumptions, one can draw three conclusions from these calculations:

1. Pensioners are penalised, and greater fiscal savings accrue, the longer pensioners retain their asset-testable assets.

2. The penalties can be reduced by giving money directly to heirs, by spending the money faster, and especially by investing in the family home. These actions are therefore incentivised.

3. There are very significant gains that can be made by adopting different drawdown strategies. Determining an optimal strategy is however complex for pensioners, made more difficult by ongoing changes in the means test rules. Forecasting is also difficult for government as pensioners change their strategies in response to the incentives.

6.1 Evaluation against the criteria

We now consider whether the system meets the criteria set by the DSS.

6.1.1 Neutrality

The current structure of the asset test is not neutral when the rules are applied over a lifetime. Annuities do enjoy relatively favourable treatment relative to applying the minimum drawdown to an account-based pension, but the latter allows for flexibility that can be used to manipulate the rules to the point of avoiding the asset test entirely.

6.1.2 Equity

The current system creates significant inequities.

• Horizontal inequity arises for people with the same level of wealth (which is at one level fungible). Non-homeowners with $1 million in assets will get no AP, while homeowners with a house of that value will draw the whole AP.

• Vertical inequity arises when the differences in treatment applied to people are not proportional to the differences in their characteristics. The dramatically higher taper applied by the asset test does not seem proportionate, neither does the failure to differentiate between those with valuable houses and those with small apartments.

• The asset test fails to provide vertical equity for those with different life expectancies. Most important is that they penalise younger retirees at the expense of older ones. A couple of 67 with $1 million in assets, which could generate investment income of about $50,000 p.a., cannot afford the ASFA comfortable retirement standard. A couple of 90 in the same position could afford to spend $100,000 p.a., well above the comfortable level.

• Women also need more than men to maintain the same spending given their greater expected longevity. Arguably the assets tests are discriminatory as they advantage men (who require less) than women.

• Regional differences in rent and the cost of living also create vertical inequities, although in this case people do have the ability to move to less expensive areas.

• Those who can obtain appropriate advice and are able to act on the advice can avoid the worst consequences of the tests. Their complexity however makes it very difficult to make the right decisions. This means that the costs are borne by the less well informed, the cautious, and possibly by those who regard it as anti-social or degrading to be supported by welfare.

6.1.3 Integrity

The current system completely fails the integrity test as defined by the DSS, quoted in Section 2.4 above. There is a clear failure to adequately address some of the most indigent whose rental costs leave little else to live on. It also fails to limit access to the AP by pensioners with significant housing assets.

It is also far too easy for retirees to adopt a range of strategies that mean that they will obtain a much greater share of the AP than those who are not so well advised or who do not wish to pursue the strategies we identified.

6.1.4 Resilience and fiscal sustainability

The desired level of fiscal expenditures could be maintained by incorporating the principal residence into the assets test and adapting the quantum of the AP and the parameters governing the means tests. In times of fiscal constraints, all Australians should expect to bear some of the burden and pensioners as much as anyone should expect smallish reductions in their standard of living at times. Transitional arrangements should apply for some limited period to mitigate the impact of the changes on those most adversely affected by the changes.

6.1.5 Simplicity

Benefits cannot be targeted precisely at those that need them. Needs are complex and fluctuating; evaluation methods inevitably somewhat simplistic, limited to a particular time and prone to error. Bradshaw and Finch [12] looked at three different approaches to measuring poverty; “lacking socially perceived necessities; being subjectively poor and having a relatively low income… we have found that there is little overlap in the group of people defined as poor by these dimensions.”

The different levels of AP and associated supplements, and permutations of the means tests create an entirely spurious impression of targeting needs. The trivial nature of some of the supplements and concessions, such as the lower deeming rate for the income test, are not only wasteful of energy but bring the whole system into disrepute. This is particularly true when material differences such as housing are not taken into account.

Maeda [13] suggests that the three laws of simplification are reduce, organise and save time. Reduction and saving time are obvious. Organisation may not be. For instance, Schedule 7 of the Superannuation Industry (Supervision) Regulations 1994 (SIS Regulations), which has 7 categories for minimum drawdown amounts for different age bands, can be seen as more complex than Schedule 1A, which has 80 categories for each age from 20 to 100. This is because one cannot remember all 7 categories so that they both have to be looked up, but changing the factor each year is more intuitive and so easier to understand, to administer and creates a smoother cash flow.

The suggestions made below in Section 7, would not only be simpler, but more clearly meet the criteria that have been mentioned: neutrality, equity, integrity and resilience.

6.2 Perversity

Failure to meet these principles creates perverse outcomes. We have shown that there is a strong incentive to draw down quickly or move assets into the family home. It appears retirees do both. While only 19% of couple pensioners are subject to the asset test, rough calculations suggest that over a third of those over 65 should be in this category. An analysis of Centrelink data suggested that those pensioners subject to the asset test draw down their assets up to 10% faster [14].

7. Options for reform

We suggest the following remedies for the problems identified.

7.1 Revert to former asset test taper rate

One policy option would be to revert to the former 3.9% asset test taper rate that applied between 2007 and 2016. Such a solution would mitigate the extremely high EMTRs that currently apply especially in the upper parts of the asset testing range and therefore reduce the attractiveness of strategies designed to increase age pension entitlements.

If the taper rate were reverted to 3.9%, there would need to be changes in the upper and/or lower assets test thresholds. For example the thresholds could revert to something like the pre-2016 asset testing thresholds (but with indexation adjustments). However if the opportunity were taken to include the principal residence in the means tests, then the lower threshold should be increased markedly, by something like the value of a basic home.

7.2 Abolish the asset test and convert superannuation into an income stream

A simpler solution is to abolish the asset test, which so contorts the retirement system and the lives of most Australian pensioners. Assets should be converted into a lifetime income (for the couple and other dependents if relevant). This can be done by dividing annually the value of assets by the Pension Valuation Factors (PVFs) in Column 4 of Schedule 1A of the SIS Regulations for singles, and using higher PVFs for couples.

This would lead to higher AP costs in the short run, but there would be no increase in AP payments as people age and the thresholds could be calibrated for the change to be fiscally neutral.

The additional fiscal costs in the short term can also be covered by including the value of the principal residence at an imputed rent. Rent assistance would be payable to everyone, further simplifying the system. The net impact would be felt most by people with low financial assets and expensive houses. Assuming an annuity rate of 7.5% at age 65, a deeming rate of 4% and a doubling of rental assistance, Table 3 shows the annual gains and losses.

Table 3.

Potential gains and losses due to proposed changes.

Single renters of moderate means would be the main winners. There are however very few of them currently, but removing the incentive to hold onto a home might well increase numbers.

The main losers would be homeowners with minimal financial assets. Their loss of spending power could be addressed by an expanded Pension Loans Scheme. Under the existing scheme, part pensioners can be lent the difference between 150% of the full AP and the part AP for which they qualify. Centrelink would take ownership of a proportion of the value of the home that this represents. As the loss of pension is only 2% of the value of the house (under the second proposed scenario), they could borrow for many years before exhausting the value of the house. A “no negative equity” guarantee would have a trivial cost.

7.3 Personal responsibility and political games

The problems with the existing system are clear. It cannot be justified, but the dire state of much current political debate means that the problems remain running sores. There is a clear need for bi-partisan agreement not to resort to irrational sound bites. Such an agreement needs to include the major industry players and media commentators. Those who exploit the uncertainty that will naturally be felt by many pensioners need to be openly and quickly rebuked. The current awareness of the damage caused by manipulated outrage may offer an opportunity for sensible discussion.

At the very least, if change is to be made, politicians and commentators will need to be courageous and to make a stand for what it right rather than attempt to score points.

8. Conclusion

The means tests in their current format make it exceptionally difficult to plan for any pattern of spending in retirement—as the Age Pension increases with reductions in assets and income in a way that is difficult to understand and model over a retired lifetime. They also provide very strong incentives for pre-retirees to engage in a range of strategies to maximise their age pension. Apart from the unnecessary fiscal burden that such strategies impose on taxpayers, they also distort savings and consumption decisions and asset prices (especially of the principal residence) and cry out for change.

Acknowledgments

This chapter is an extension of [15], a public submission in response to the Department of Social Services discussion paper [3]. Quotations from the earlier paper are therefore not marked.

Conflict of interest

The authors declare no conflict of interest.