Retirement Adequacy: A Case for a Hybrid Measure

The  Retirement Income Review  (RIR) that has been commissioned by the Government is undertaken by a panel chaired by Mr Michael Callaghan, with Ms Carolyn Kay and Dr Deborah Ralston as panel members.

The objective of the RIR is to establish a fact base of the retirement income system that could potentially improve the understanding of the system and the outcomes it is delivering for Australians. In particular, the terms of reference states that “it is important that the system allows Australians to achieve adequate retirement incomes, is fiscally sustainable and provides appropriate incentives for self-provision in retirement”.

What is meant by adequacy?

The  Consultation Paper  sets the tone by stating that no consensus exists on what the best metric is to measure an adequate retirement income and that it will vary between individuals.

It introduces two distinct measures (which should be well known to actuaries working in superannuation):

• Relative measures – replacement rates based on an individual’s income prior to retirement. A replacement rate of between 60% and 70% is typically considered appropriate for most people ^[1]  and is consistent with many defined benefit pension designs;
• Absolute measures – placing a dollar value on the income required. The  ASFA standards  are examples of this. 

The paper explores the weaknesses and strengths of both types of measures individually. The discussion shows that both measures, in isolation, are inadequate. It also shows that many of the weaknesses of each tend to be the strengths of the other, suggesting that the most reasonable measure might lie somewhere between the two.

The paper poses a consultation question of what measures the Panel should use in their assessment whether the retirement income system allows Australians to achieve an adequate retirement income (Page 15).

Why not adopt a hybrid measure?

Following a number of discussions on this topic, the Actuaries Institute’s Superannuation Projections and Disclosure (SPD) subcommittee suggests that a better measure would be a hybrid measure.  

Intuitively, a hybrid measure makes sense. Most low income individuals will not see a significant drop in their expenses post retirement, as the majority of their income will be used to meet their needs. At the extreme, an absolute measure of the poverty level should be the minimum. As income levels rise, a greater proportion will be spent on discretionary items which may reduce once in retirement.

Implicitly, the Australian system already adopts a hybrid approach.  For retirees on low incomes, an absolute measure applies – the Age Pension.  Similarly, an absolute measure also applies to retirees with high incomes – the $1.6 million superannuation cap imposes an upper dollar limit that can be transferred into the retirement phase. For other retirees, the superannuation guarantee (SG) theoretically provides a relative measure based on lifetime salary.

Continuing with the measures as they are introduced and defined in the consultation paper, the hybrid measure would be:

            A% x a Relative measure + (1 – A%) x an Absolute measure

where:

• 0% < A% < 100%; and
• the Relative measure can take the form of an income replacement model, i.e. B% x Income prior to retirement where B% tends to fall between 60% and 80%. 

There are many ways that this hybrid measure can be calibrated, simplified and re-expressed to assist understanding for a broader audience. To that end, we propose a hybrid model based on essential and discretionary expenditure ^[2]  as:

$X + Y% x Income

To calibrate this against the ASFA Retirement Standard benchmarks:

• For simplicity, we note the annual budget for a “modest” retirement lifestyle is approximately $28K, which could be assumed to represent essential expenditure, or $X.
• The annual budget for a “comfortable” retirement lifestyle, which could be assumed to include discretionary expenditure for a “typical” worker, is approximately $44K. If we assume that discretionary expenditure varies proportionately to income this suggests that Y% would be in the order of 22.22% based on a median income of $72K, noting that ($44K – $28K)/$72K = 22.22%. This could be rounded down to say 20% because some of the discretionary income may not vary proportionately to income.

This is likely to make more sense than the original form because it directly captures the essence of a hybrid measure – what is considered to be adequate should differ (to a certain degree and extent) for different individuals but a minimum level should be factored-in for those at the less fortunate end of the income distribution.

The proposed hybrid measure is to spark further discussion as a hybrid measure requires considerations such as:

• How do we set the values of $X and Y%? As a guide, the above analysis suggests that $X = ASFA Modest and Y% = 20%.   
• Should there be an upper limit?
• What period of pre-retirement income should be used to determine the “Income”?
• How do we address the question of individual versus family income? Should the RIR panel follow the Age Pension and ASFA leads by defining four standards (Couple / Single, Home-owner / Renter)?

Like all difficult questions in life, the answer is “it depends”. It depends on the purpose of the Australian retirement income system and the role each pillar should play for different Australians.

Superannuation Projections and Disclosure Sub-Committee

Estelle Liu, Convenor
Bill Buttler
David Carruthers
Esther Conway
Colin Grenfell
Ian Fryer
Jean Nette Koay
David Orford
Richard Starkey
Young Tan
Rein Van Rooyen
Brnic Van Wyk