Part 2: Mandatory, Standardised Retirement Estimates

In Part 1, we argued for mandatory retirement estimates that allow for the Age Pension and include a total retirement income. To fill the data gaps that exist, we argued for some level of standardisation with flexibility that allows retirement estimates to incorporate:

• As much member data as is available
• Product data that represents the provider’s product range
• Reasonable and consistent economic assumptions
• Indications or reasonable assumptions about member retirement circumstances and preferences
• Terminology and disclosure that aligns with consumer and industry expectations

In this article, we’ll explore standardisation in more detail.

What aspects of retirement estimates should be standardised?

Member and product data

Super providers hold account information essential for administration purposes, such as age, balance, contributions, and investment choice, and this should, of course, be used in retirement estimates.

Over time, funds may gather more information about their members that can assist in providing retirement estimates. Relationship status, employment income, risk profile, retirement age, health status, homeowner status, and various other gaps may be filled as members interact with their provider through helplines, calculators, and transactions. If there is sufficient credibility attached to this information, it makes sense to use it to provide a more personalised retirement estimate, provided it is disclosed at the same point.

Similarly, when considering super and retirement products, using the actual product features and fees/deductions associated with the member’s account makes more sense than a generic industry average or no fees at all. Where the member has options, such as a menu of retirement products, then presenting a couple of representative choices with appropriate disclosure might be better than a one-size-fits-all product assumption. For example, a retirement estimate could include a scenario to illustrate 50% of the retirement balance being converted to a life annuity or similar product offered by the provider.

It should be noted that guaranteed income streams are not perfectly comparable with market-linked income streams on a return-only basis. Appropriate scenario comparison should address volatility, longevity and inflation risk in ways that members can understand.

Economic assumptions

One dilemma that gives actuaries cause for hesitation is the selection of economic assumptions. This often comes down to the trade-off between inter- and intra-provider comparisons of economic assumptions. A high level of standardisation for retirement estimates may generate inconsistencies between the retirement estimates and other superannuation forecasts, advice documents, investment objectives, and research publications from the same provider. This is due to the necessity of showing retirement estimates in today’s dollars, and the role of inflation in both setting investment return assumptions and deflating future values.

For comparisons between providers, it makes sense to standardise economic assumptions, so that when members change providers or have accounts with more than one provider, their retirement estimates are consistent. Whether I hold my balanced portfolio with provider A or provider B, I should expect to get a similar retirement estimate.

However, when thinking about comparisons between the retirement estimate from provider A and the investment objectives of provider A, it makes sense not to necessarily standardise the economic assumptions. We might find that the retirement estimate shows the balanced portfolio earning an assumed 5.5% p.a. above assumed price inflation of 2.5% p.a. over the projection period, but the investment objective targets 4% p.a. above price inflation of 3% p.a. expected over the next 10 years.

There is no perfect level of standardisation in economic assumptions that delivers an acceptable level of inter-fund consistency without compromising intra-fund consistency. Having standardised approaches for when to use price and wage deflators but letting providers set values for those variables according to their modelling, compromises inter-fund consistency. Alternatively, if numerical values for price and wage deflators are prescribed, then requiring providers to use investment returns that make allowance for these will likely give rise to intra-fund discrepancies.

Retirement preferences

A member’s retirement preferences are among the most personalised aspects of a retirement estimate, but for members there is often an interdependency between setting preferences and seeing an estimate in the first place.

At the time of providing an estimate to a member in their 20s or 30s, they may not have considered their retirement preferences at all, whereas a member in their 50s may have a strong view as to when they want to retire and the kind of lifestyle they want. Given these inputs are preferences and are often a long way into the future, standardisation of retirement age and retirement period is appropriate if otherwise unknown.  A reasonable alternative to setting a retirement period might be to set a target retirement income relative to a member's working income and have the estimate solve for how long this may last. However, if a member expresses particular preferences for retirement, providers should be permitted to use them in an estimate.

Presentation

Considering the educational potential of a retirement estimate, standardisation of the presentation permits an excellent financial literacy opportunity. To that end, it is important to use standard terminology and present standard results on all retirement estimates. Some of the key presentation aspects are:

• that a retirement estimate is presented in inflation-adjusted terms (including disclosure on which measure of inflation)
• that both a lump sum and a retirement income are presented
• that all data and assumptions used are disclosed
• that options available to the member are highlighted

The inclusion of ‘benchmark’ retirement metrics (e.g., such as the lump sum required to provide a ‘comfortable’ retirement income) can be helpful to frame a retirement estimate, but this is beyond the scope of standardisation, and care should be taken to disclose the differences between data and assumptions used in the benchmark versus the retirement estimate.

The layout and cosmetic aspects of retirement estimates are best left flexible to align with the branding and tone of the provider.

Selection

Accepting our preferred definition of mandatory retirement estimates at the provider level, a crucial aspect of the debate is in relation to which members are selected for retirement estimates. There are good reasons not to provide retirement estimates to some members where their circumstances suggest that the data held or assumptions applied could lead to unhelpful or misleading retirement estimates. Some of these examples are:

• no regular contributions have been received
• direct investments
• under ongoing advice
• where both accumulation and pension accounts exist
• defined benefit accounts
• uncertainty about future voluntary contributions

It is important to allow providers to withhold retirement estimates from some members, which would allow us to argue more strongly for mandatory retirement estimates at the provider level.

Coming in Part 3

We'll compare our proposed approach with the current regulatory framework under ASIC RG276, identify what's working well, and recommend specific changes that would improve both the quality and availability of retirement estimates for Australian superannuation members. 

The authors would like to thank Colin Grenfell, who peer-reviewed this series.