RETIREMENT INCOME ADEQUACY ARE WE STILL MAKING PROGRESS?

Transcription

1 1 John Burnett, Willis Towers Watson Burnett is an actuary with more than 35 years experience in the superannuation industry. He has particular interest in benefit projections and online calculators and their role in member decision-making. He is a senior consultant with the Willis Towers Watson Australia Benefits team. Nick Wilkinson, Willis Towers Watson Also a member of the Benefits team, Wilkinson heads up the actuarial development and review of Willis Towers Watson s online calculator offerings for clients. He is also part of the Adequacy research team. RETIREMENT INCOME ADEQUACY ARE WE STILL MAKING PROGRESS? John Burnett and Nick Wilkinson While investment markets performed strongly from 2010 to 2014, there haven t been corresponding improvements in the projected retirement income adequacy for all groups, especially for couples. A key factor causing the somewhat mixed results is the impact of changes to the Age Pension assets test announced in the 2015 Federal Budget which take effect from 1 January Are we still making progress? In most cases, yes. However the latest research highlights the significant impact that the changes to the assets test will have on the projected retirement incomes of many Australians. If we are to have a fully integrated retirement system built around the three main pillars of the age pension, superannuation and other savings, which will allow future retirees to plan effectively for their retirement, then greater certainty about the level and availability of the age pension for future retirees is vital. Ongoing research by Willis Towers Watson and The University of Melbourne is providing a more complete picture of retirement adequacy for singles and couples. Previous research, issued in July 2015, provided a look through time at the projected retirement incomes (including superannuation, other savings and the age pension) using Household, Income and Labour Dynamics in Australia (HILDA) Survey data from 2002 to In this latest research, we ve looked at the most recent HILDA data from 2014 (released in late 2015) to assess whether the adequacy of expected retirement incomes is continuing to improve. adequacy from 2010 to 2014 Using the 2014 HILDA data released in December 2015, we have projected the retirement income for a representative sample of the Australian population covering 11,815 persons aged 25 to 64. We have taken into account compulsory superannuation, voluntary superannuation, the age pension and other retirement savings (excluding the family home) in these projections. Figure 1: Percentage projected to reach or exceed ASFA Comfortable retirement income target

2 2 Figure 2a: Distribution of retirement incomes for couples in 2010 and 2014 based on current assets test Figure 2b: Distribution of retirement incomes for singles in 2010 and 2014 based on current assets test Figure 3: Percentage projected to reach or exceed ASFA Comfortable retirement income target before and after changes to the Age Pension assets test in 2017 To begin with, we consider the change in retirement adequacy from 2010 to 2014, before allowing for the impact of the recently legislated changes to the assets test applying to the age pension from 1 January 2017 onwards. In 2014 under the current asset test rules, our projections suggest 62% of couples and 38% of singles are expected to reach or exceed the comfortable level of retirement income. For couples, the proportion expected to retire comfortably rose from 53% in 2010 to 62% in 2014, representing an increase of 9%. The proportion of singles expected to retire comfortably increased even more significantly from 23% in 2010 to 38% in 2014, an upturn of 15%. Factors contributing to these results are: Relatively strong investment returns experienced during the period resulting in higher than expected superannuation assets available to fund retirement incomes; and Age pension increases of approximately 4.7% pa over the period compared to the 2.6% p.a. increase in the comfortable level of retirement income under the ASFA Standard. These factors, together with growth in the retirement income expected from superannuation as the Superannuation Guarantee system continues to build, were more significant overall for singles. As a result, in 2014 significantly more singles are expected to attain the comfortable level of retirement income than was the case in Distribution of retirement adequacy By looking deeper at the distribution of results from 2010 to 2014 we can see that changes in projected retirement income have been more uniform for couples than singles. For singles, there have been limited gains between 2010 and 2014 at the lower end of the distribution where the age pension is the main source of projected retirement income. At the upper end of the distribution, gains have been much stronger for singles with more superannuation and other savings. On the other hand, gains in projected retirement income for couples from 2010 to 2014 have been more uniformly spread, although the gains are still higher for those couples with more superannuation and other savings. Impact of Age Pension changes in 2017 We have subsequently considered the impact on projected retirement incomes of the changes to the assets test applying to age pension payments from 1 January 2017 onwards. These changes were announced in the 2015 Federal Budget and legislated in As a result of these asset test changes, we can see there will be a significant impact on the proportion of couples expected to achieve a comfortable level of retirement income, with a reduction from 62% to 51%. On the other hand, singles are expected to be affected less with a reduction from 38% to 37% in the proportion of singles expected to achieve a comfortable level of retirement income. Under the changes to the assets test, the taper rate will double from $1.50 to $3 per fortnight for each $1000 of assessable assets above a minimum threshold, meaning that eligibility for partial age pension payments phases out more rapidly and at lower amounts than previously. At the same time there were some increases legislated to the lower thresholds used in the assets test. The new thresholds expected from 1 January 2017 are shown in Figure 4 (current thresholds are shown in brackets).

3 3 Figure 4: New assets test thresholds expected from 1 January 2017 Figure 5b: Distribution of projected retirement incomes for singles in 2014 before and after the 2017 Age Pension asset test changes Figure 5a: Distribution of projected retirement incomes for couples in 2014 before and after the 2017 Age Pension asset test changes Figure 6a: Composition of projected retirement income for selected percentiles of couples Distribution of Impact of Age Pension Changes in 2017 Looking at Figure 5a, we can see that couples in the 15th to 85th percentile will be the most affected by the assets test changes in Couples on lower retirement incomes than this will still generally receive the full age pension, while the age pension plays only a small part for those couples with the highest levels of retirement income. In contrast, more singles are expected to receive the full age pension and have generally lower levels of superannuation and other assets being assessed under the assets test. While there is some impact from the asset test changes in 2017 on singles at higher levels of retirement income shown in Figure 4b, the impact is minor in comparison with the impact on couples. Composition of projected retirement incomes By examining the projected retirement income for those at or near the 25th, 50th and 75th percentiles, the relative contribution of the age pension, superannuation and other savings can be assessed. In examining the breakdown of results from 2010 to 2014 (prior to the asset test changes), there was growth in projected retirement incomes arising from each of the age pension, superannuation and other assets. At the 25th percentile for singles, the limited growth from superannuation and other assets was due to coming off a low base in The breakdown of results in 2014 before and after the asset test changes is also worth noting: For couples the age pension payments over the period in retirement are expected to reduce by 7% at the 25th percentile, 29% at the 50th percentile and 37% at the 75th percentile. In dollar terms

4 4 Figure 6b: Composition of projected retirement income for selected percentiles of singles Figure 7a: Percentage of couples projected to reach or exceed ASFA Comfortable retirement income target at each measurement date Figure 7b: Percentage of singles projected to reach or exceed ASFA Comfortable retirement income target at each measurement date * Old rules reflects the current Age Pension assets test whereas New rules allows for the 2017 changes to the assets test the annual reduction in pension payments exceeds $6000 at both 50th and 75th percentiles as a result of the asset test changes commencing in 2017; For singles the reduction in the age pension payments as a result of the asset test changes is less severe with a 6% reduction at the 75th percentile and little or no reduction at the 25th and 50th percentiles. The asset test changes have less impact here due to the lower superannuation and other savings held by singles at these levels. Are we still making progress with retirement income adequacy? By combining these 2014 results with our earlier results covering 2002 to 2010 issued in July 2015, we are able to track the proportion of our population projected to achieve a comfortable level of retirement income based on the ASFA Standard. Investment markets delivered strong results to superannuation savings and other assets over the period from 2010 to 2014 and this was reflected in the improving retirement income adequacy we initially measured in However, once we allowed for the asset test changes in 2017 which will reduce the part pension payable in future to those with relatively more superannuation savings and other assets, this has eroded much of the progress made, particularly for couples. While it is encouraging to see the progress with more singles likely to reach a comfortable level of retirement income in 2014 than previously, there is still more to be done here as reliance on the age pension as the primary source of retirement income remains high. Whether the 2017 asset test changes are necessary to create a more sustainable age pension system is a separate question which we have not addressed here. We simply note that a single policy change to the age pension assets test has reasonably significant implications for the retirement incomes of both current and future retirees. If Australians are to be encouraged to plan for retirement in advance, they need greater certainty on the level and availability of the age pension that will be payable to them once they retire. income targets While there are many ways to measure retirement adequacy, in this research we have adopted a fixed dollar amount as the target retirement income, based on the ASFA Standard Comfortable level. For couples this is $58,128 p.a. and $42,433 p.a. for singles in mid We have indexed these targets in future years to allow for wage inflation so the same targets apply in real terms. We note that alternative targets may be applied in other circumstances, such as basing the target on a percentage of pre-retirement income. Information from the HILDA Survey Our research uses data from the Household, Income and Labour Dynamics in Australia (HILDA) Survey collected in 2010 and We also refer to earlier results using 2002 and 2006 data. The HILDA Survey is funded by the Australian Government Department of Social Services (DSS) and is managed by the Melbourne Institute of Applied Economic and Social Research (Melbourne Institute). The findings and views reported in this paper, however, are those of the authors, and should not be attributed to either DSS or the Melbourne Institute.

5 5 While the HILDA Survey collects information from surveyed households each year, there is detailed information on household assets and debts collected every four years (Wave 14 in 2014 being the most recent, released in late 2015). This includes information about superannuation balances, voluntary superannuation contributions, other savings and investments. How the projections are conducted The projections are based on the methodology adopted in the Willis Towers Watson Planner engine which powers a number of retirement income projector tools issued by super funds as well as a number of calculators available for public use through ASIC s MoneySmart website. This methodology has been extended in this research to incorporate the additional information on individual circumstances that is available from the HILDA Survey. Using this model, we project the retirement savings to the assumed retirement age of 65 and then the age pension eligibility is calculated in that year and in each subsequent year in line with both income and assets means-testing requirements. During this postretirement period, superannuation and other retirement savings are then drawn down so that this wealth is exhausted by age 90. We calculate the level of retirement income that is maintained in real terms over the period from age 65 to age 90. The current status as either a home owner or renter is assumed to continue into retirement and is allowed for in the age pension means-testing. Where home ownership applies, we assume this continues to at least age 90 and do not draw on this asset when projecting retirement income in this research. Further research ahead Projections on which this research is based depend on a large set of assumptions about individual, household and market factors, including wage and salary progression, future investment returns, employment status, marital status, home ownership status and retirement age. In our research to date, we hold many of these factors steady but there is an opportunity to introduce more variability into these assumptions, to assess the sensitivity of the adequacy measures to each contributing factor. For example, this could include stochastic modelling of investment returns or investigating the impact of changes in retirement age. The HILDA data provides a wide range of information and in future research this will continue to contribute to a deeper understanding of the characteristics of each sub-group within the surveyed population and how this influences their projected retirement incomes. fs