White Papers
Effective investment strategies for retirement
13/09/2013
Objectives in retirement
The principal objective of retirees is generally to maintain a particular standard of living for a given period of time. Merton^{1} offers the following general criteria for good retirement plan design for individuals:
- Offers robust, scalable, low cost investment strategies that maximise the chances of achieving the retirement income goal using all available assets.
- Manage the risks of not achieving that goal.
- Be effective for unengaged participants.
- Provides meaningful information and choices with easy implementation for those who do engage on their progress to achieving the retirement income goal.
We would add an additional objective that the solution needs to be practically deliverable to large cohorts of members (what we would term “macro-economically consistent”).
This final point is important in that a number of “solutions” proposed to pre and post retirement funding would not be macro-economically consistent if they were applied to the broader population. Such solutions may be feasible for small cohorts, however, if a solution is not macro-economically consistent then it cannot be a broad based solution for a large scheme.
Consequently, solutions that rely on substantial financial engineering (e.g. leverage) and a high degree of complexity or where underlying asset class weights (or shifts) wouldn’t be feasible for large proportions of the population are unlikely to represent sustainable long-term solutions for a fund (and in any case probably breach objectives 1 and 4 above).
In determining at what level the post retirement income stream should be set and the duration of that income stream for an individual, retirees face a number of interconnected decisions.
Generally, by the time a retiree is in a position to make decisions about their drawdown rate and investment strategy, the value of their investment capital will have largely been determined. In addition, the impact on social security payments and other assets outside of the superannuation system can be assessed. Consequently, we can consider the interconnectivity between drawdown rate, investment strategy and duration of the income stream as outlined in Chart 1.
Chart 1: Interconnectivity of retirement decisions
Clearly, a higher drawdown rate for a given investment strategy is likely to result in a shorter duration of the income stream, while for a given drawdown rate a more aggressive investment strategy will result in a wider dispersion of outcomes of the duration of the income stream, albeit with on average a longer duration (subject to the impact of sequencing risk which we address in the next section).
Why reducing investment risk is important
The nature of the accumulation and decumulation process is such that money weighted outcomes (which is what individuals receive) can and do vary quite significantly from time weighted outcomes (which are typically reported as the long term return).
To highlight this issue, Chart 2 shows the progression of a member balance for a given drawdown policy and average earning rate, but where the sequencing of the earning rate differs. Given that it is not just the post-retirement phase that matters we have conducted our analysis to also incorporate the final 5 years of the accumulation process. We have assumed a member in the final 5 years pre-retirement continues to make contributions of 12% of salary through that last five years and then sets a drawdown rate of 60% of their then final salary. The scenarios we consider are a constant earning rate of CPI+3.5% together with alternating periods of CPI+3.5% plus or minus 6.5%. Note “pos” means the first period is plus 6.5% and “neg” means the first period is minus 6.5%. We have chosen 1, 3 and 5 year periods for comparison.
Chart 2: The impact of volatility on investment outcomes
Source: Schroders. Starting balance at age 55 is 7 times salary (reasonable for a 12% accumulation rate throughout a full working career) and contributions to age 60 are at 12% of salary which is indexed with inflation + 1%. Inflation is 3%, earning rate is 6.5%. Drawdown amount from age 60 set at 60% of final salary and indexed with inflation.
Unsurprisingly, it shows that while the average earning rate across all scenarios is constant:
- the greater the “length” of the volatility period, the greater the dispersion in results; and
- the first period being a “negative” volatility period has a greater impact on the duration of drawdowns that is possible than the first period being a positive volatility period.
The variance of +/-6.5% was chosen as it results in a portfolio with broadly equivalent real return and volatility in the traditional sense (not dissimilar to a traditional balanced fund). Clearly, as this number is reduced, the differential between the outcomes will reduce and the opposite occurs as the volatility is increased. However, even a reduction in the volatility figure to 3.5% still results in a material difference between outcomes (e.g. 4 additional drawdown years for the 5 year scenarios).
In addition, Chart 3 shows the difference in the balance after 15 of drawdown from the base case (which assumes a constant earning rate with no volatility).
Chart 3: The impact of volatility on investment outcomes
Source: Schroders, Starting balance at age 55 is 7 times salary (reasonable for a 12% accumulation rate) and contributions to age 60 are at 12% of salary which is indexed with inflation + 1%. Inflation is 3%, earning rate is 6.5%. Drawdown amount from age 60 set at 60% of final salary and indexed with inflation. Results above show difference in balance at age 75 between given scenario and base case of no volatility in earning rate. A result below -100% means that the drawdown strategy had resulted in a negative balance.
Chart 3 shows that not only are the differences in results potentially very large as we introduce greater periods of volatility, but interestingly where the first period was a negative variance, the difference in results was somewhat greater than when the first period was a positive variance. Bearing in mind investors have little choice in the issue of whether first year returns will be positive or negative, this is an important finding and is a result of the constant drawdown process (you have less money in later years when returns are better if the early return period is poor).
Ultimately, there are two primary drivers in reducing investment risk to enable individuals to achieve better retirement outcomes across different time periods:
- the absolute level of real returns through time; and
- controlling the level of volatility of those return streams to minimise “sequencing risk”.
In the remainder of this paper we assess the impact of alternative investment strategies and drawdown rates on the likely outcomes for retirees in terms of the duration of the income stream.
What matters more – Risk or Return?
In assessing the implications of alternative drawdown rates and investment strategies we must have some basis for determining what is “good” and what is “bad”. To this end, we would suggest that a focus on the negative outcomes is most important. Retirees (or indeed most investors) would be unconcerned with scenarios that result in better than expected outcomes. However, running out of money too early clearly represents a major risk to retirees where they are not in a position to offset that decline in income through other sources (e.g. human capital or the build up of other assets).
John C Bogle, the Founder of The Vanguard Group, phrased it quite succinctly in 2001 commenting that: “We must base our asset allocation not on the probabilities of choosing the right allocation but on the consequences of choosing the wrong allocation.”^{2}
Consequently, we focus our analysis of alternative investment and drawdown strategies on the left tail of the distribution of outcomes – what are the worst case scenarios that would impact a proportion of a retiring cohort.
Analysis
Given that the principal objective of the investment strategy in the drawdown phase is to maximise the value and the duration of the income stream, we examine the effectiveness of alternative investment strategies and drawdown rates in 2 ways:
- We look at the historical outcomes from following a given approach. This is useful in that the investment outcomes actually occurred and are arguably more realistic. The downside of this approach is that it is not forward looking and the data set is more limited. In addition, the maximum duration of a drawdown is necessarily capped at the last year of data (e.g. for a retiree in 1990, the maximum drawdown is 21 years).
- We take historical return outcomes for different asset classes in 5 year “blocks” and then randomize these blocks to generate any number of potential investment outcomes of the length necessary to examine the drawdown. The key points to note here are that we are sampling from a period in history that has been generally quite good (longer term returns of CPI+5% from a balanced approach) and there is no natural “mean reversion” where periods of strong returns are followed by periods of poor returns as sample periods are chosen randomly in 5 year groups.
Traditional investment strategies
In this section we compare how a traditional Conservative Balanced or Balanced strategy^{3} would have fared on both a historical basis and our “sampled” approach.
For reference, the Chart 4 shows the rolling 10 year real return outcomes from a Conservative Balanced and a Balanced investment approach (on a time-weighted basis).
Chart 4: Rolling 10 year real returns
Source: Schroders, Global Financial Data.
The average real return since 1930 from a Conservative Balanced fund has been 2.8% p.a. while a Balanced fund has had an average real return of 4.7% p.a. (or 3.2% p.a. and 5.2% p.a. respectively since 1900). However a conservative balanced fund has generally suffered from greater negative outcomes in real terms, with a minimum of -3.8% p.a. real over the period versus only -3.0% p.a. real for a Balanced fund. This is worth noting in terms of understanding the true risk of an investment strategy as in a more traditional risk framework, the annualised volatility of the Conservative Balanced fund has been 35% less than a Balanced fund over the same period.
Consequently, in thinking about approaches which notionally represent a “lower risk” outcome for the individual and where that is taken to represent a greater exposure to bonds over equities, this does not necessarily mean that the strategy is lower risk in terms of potential outcomes for the individual.
We now turn our analysis to drawdowns. In Chart 5 we determine the length of time in years a different drawdown approach lasts based on a either a Conservative Balanced or a Balanced investment approach. We calculate the number of years until the account balance falls below zero assuming the drawdown starts in any year from 1930 until 1990 and where the drawdown amount is initially set at a percentage of the capital balance and then indexed with inflation each year. Investment returns are calculated to January 2012 (albeit for drawdowns that commence in the 1980’s these may still not have run out by the end of the calculation period). The range shows the minimum length of drawdowns up to the 25th percentile with the orange dot representing the 10th percentile. That is, we are only looking at the bottom 25% of observations in these charts, not the full distribution of outcomes.
Chart 5: Historically, Balanced has produced longer drawdowns than Conservative
Source: Schroders, Global Financial Data. Assumes initial drawdown of stated percent of capital at end of year and indexed thereafter with inflation. Investment return based on stylised conservative balanced fund and stylised balanced fund with fixed strategic asset allocation as set out in appendix. Drawdowns commence in year stated and go until 1 January 2012 or when balance reaches $0. We can observe that:
- As expected, the balanced strategy generally delivers better minimum outcomes than a Conservative Balanced strategy.
- For all of the drawdown amounts, there appears to be no benefit to choosing a more conservative strategy as the observed outcomes were generally lower for a Conservative Balanced strategy than a Balanced strategy.
This is again an interesting observation in the context that many would see a Conservative Balanced strategy as “lower risk” than a “balanced” approach. This is particularly the case for more traditional “lifecycle” funds.
Chart 6 considers the same range of outcomes but utilising our “sampled” returns history and consequently having no constraints in terms of the maximum time for which drawdowns could run.
Chart 6: Sampled distribution also shows a better outcome for riskier strategies
Note that the scale on this chart is set to match the historical chart above to enable better comparability. However this does mean that our 10th percentile on the 4% balanced scenario is above the chart.
Source: Schroders, Global Financial Data. Assumes initial drawdown of stated percent of capital at end of year and indexed thereafter with inflation. Investment return based on stylised conservative balanced fund and stylised balanced fund with fixed strategic asset allocation as set out in appendix. Results are based on 500 random simulations for up to 600 years each.
Similar to the analysis based on actual historical data, the Balanced strategy generally results in better (minimum) outcomes than a Conservative Balanced strategy across all drawdown approaches. As expected, the use of a longer data set results in a higher 90% probability (10th percentile) across all approaches, with a bias towards the balanced strategy achieving better overall outcomes.
The principal result from the above analysis is that:
- Balanced strategies are generally better in terms of minimum likely outcomes for a variety of different drawdown levels as the higher return has usually been more than adequate to compensate for the greater sequencing risk given the higher volatility of these approaches.
- “Lifecycle” type strategies that target a more conservative balanced style outcome at retirement are unlikely to result in better outcomes during the drawdown phase than simple balanced approaches that do not change asset allocation in a glide-path. The observation that these are “lower risk” portfolios is clearly false (and has been historically).
However, as can be observed in Chart 4, the real issue with Conservative Balanced or Balanced approaches is that they have considerable medium-term volatility of results, with outcomes that have ranged from less than -3% p.a. real to almost +12% p.a. real. Likewise, in Charts 5 and 6, while we are only focussed on the bottom 25% of potential outcomes, even that range can be necessarily quite wide and result in poor predictive power of any given investment strategy for a retiree looking at potential drawdown outcomes.
Clearly, a better approach would be one in which the range of potential outcomes was significantly reduced and where the minimum outcomes were improved.
Alternative investment strategies
In this section we compare how an alternative outcome orientated approach would have fared on both a historical basis and our “sampled” approach. We compare a CPI+3.5% outcome with a Conservative Balanced strategy and a CPI+5% outcome with a Balanced strategy over the historical data set and using the sampled data. The principal point of this analysis is to examine the impact of reducing the potential from sequencing risk while targeting an outcome, versus a traditional strategic asset allocation (SAA) orientated approach.
We have chosen CPI+3.5% and CPI+5% to be broadly representative of what a Conservative Balanced and Balanced fund would typically aim to achieve over time. For the purpose of this analysis the return series used is exactly CPI+3.5% and CPI+5%. Clearly, any outcome based strategy that targets these returns is likely to have some variability around the target however this is not possible to model in an historical framework as there are few such strategies that have been around for any length of time. We would however expect an outcome based strategy to have significantly lower variability around these targets than an equivalent SAA based strategy (and indeed that is the point of an outcome based approach).
Charts 7 and 8: Outcome orientated strategies provide greater certainty and longer duration
Source: Schroders, Global Financial Data. Assumes initial drawdown of stated percent of capital at end of year and indexed thereafter with inflation. Investment return based on stylised conservative balanced fund and stylised balanced fund with fixed strategic asset allocation as set out in appendix. Drawdowns commence in year stated and go until 1 January 2012 or when balance reaches $0.
We can observe from Charts 7 and 8, that in both cases, there is a material improvement in outcomes across all drawdown scenarios (with the exception of the 4% drawdown in the Balanced/CPI+5% case) from a smoother outcome orientated approach versus a more traditional SAA based approach. While we would expect that an actual outcome orientated investment approach would generate more volatility than shown here, there is such a significant difference in the results that even a modest increase in volatility would still lead to significantly greater certainty and overall higher outcomes from the CPI+ approaches.
We now consider the same analysis for the sampled data set.
Charts 9 and 10: Outcome orientated strategies provide greater certainty and longer duration
Source: Schroders, Global Financial Data. Assumes initial drawdown of stated percent of capital at end of year and indexed thereafter with inflation. Investment return based on stylised conservative balanced fund and stylised balanced fund with fixed strategic asset allocation as set out in appendix. Results are based on 500 random simulations for up to 600 years each. Note the scale mans that some of the data is missing off the top of the chart.
Once again we can observe that, in general, the outcome orientated approaches resulted in substantially better outcomes and a significantly lower range of potential outcomes, i.e. greater certainty.
Macro-economic consistency
The point that any solution should be macro economically consistent is predicated on the concept that for an approach to be feasible for a large number of members, scaling it up to represent a large part of the population should not cause it to fail. Clearly, some solutions can be applicable for smaller cohorts without destabilising the system, however these can only ever apply to smaller groups. To this end, we would question the ability of a number of commonly suggested strategies to be feasible for post-retirement, such as:
- Income strategies - if we all pursued them then these stocks become progressively over-priced, significantly more risky and gradually fall in yield. In addition, corporates are likely to respond to a yield demand by generating greater yield through financial engineering. This is unlikely to be optimal.^{4}
- Leverage and synthetic replication – strategies that involve undue amounts of leverage or synthetic exposure to smaller asset categories cannot be scaled across large cohorts.
- Strategic asset allocation strategies –strategies which allocate capital to asset classes irrespective of price suffer from significant longer term volatility and cause individual asset classes to become over-priced.
- Structured products that require capital backing – structured products are not investments in their own right, rather a route to an underlying investment. To the extent that these require significant “capital” to back them, the drag on returns by definition makes them expensive. We would put lifetime annuities in this category.
- Protection strategies – be they CPPI, option replication or actual exposure to options, these cannot be implemented in significant size.
- Unrealistic return targets – long term returns should be consistent with limits imposed by aggregate real growth and sustainable risk premiums.
By addressing the broader question of macro-economic consistency we can focus our attention on less complicated solutions that are likely to deliver acceptable outcomes over time. To this end we prefer strategies that are anchored in fundamental valuation, make only limited use of leverage, exotics or synthetic instruments and are fully invested in capital markets (equity, bonds and cash) versus hidden in a structure that requires capital backing.
Conclusion
The process of decumulation and the investment strategy adopted will have a large bearing on the duration of drawdowns, the size of those drawdowns and the consistency with which different member cohorts will achieve similar outcomes. These outcomes can vary significantly even when the chosen investment strategy is relatively stable.
For investors looking for a greater degree of certainty in their drawdown outcomes it is important to control sequencing risk as well as achieving an absolute real return through time.
The decumulation process, and in particular, the structure of the offerings available to those approaching or already in retirement, is a very important part of the overall superannuation process. Traditional Conservative Balanced or Balanced approaches have considerable medium-term volatility of results, and result in quite low predictability of drawdown outcomes. In addition, a Balanced strategy which would generally be assumed to be of “higher risk” than a Conservative Balanced strategy, resulted in better minimum outcomes than the lower risk strategies. Given that many lifecycle approaches target a more bond heavy allocation towards retirement, it is clear that more conservative approaches have not resulted in greater certainty of outcomes for the retiree.
Clearly, a better approach would be one in which the range of potential outcomes was significantly reduced and where the minimum outcomes were improved. In general, the outcome orientated approaches result in substantially better outcomes and a significantly lower range of potential outcomes, i.e. greater certainty.
In particular, a range of outcome orientated approaches provides:
- A better series of strategic choices for retirees that is easier to understand and implement in their retirement planning process.
- Can be applied simply to disengaged as well as engaged members and reduce the incidence of poor aggregate outcomes.
- Increases significantly the certainty of outcomes and so provides a much better opportunity to plan and in particular decide on a more robust drawdown strategy.
- Is macro-economically consistent and so can be applied to the general population and consequently large cohorts of members.
Appendix
Bibliography of prior research pieces on Objective Based Investment Strategies (and related topics) from Schroder Investment Management Australia Limited referred to in this paper. Copies available on request:
- 2007, 2008, 2009, “CPI+5 White Paper”
- January 2009, “It’s about risk, not return”
- April 2009, “What price complexity”
- August 2009, “Keeping it simple, back to the future for Asset Allocation”
- February 2011, “Complexity Adding Value”
- August 2011, “Post Retirement – Time to Focus on the Endgame”
- September 2011, “Life Cycle Funds – Just Marketing Spin”
- March 2012, “Why SAA is Flawed”
- April 2012, “Asset Allocation - How flexible do we need to be?”
- May 2012, “Understanding the Journey to Retirement”
- October 2012, “Risk Parity – No Free Lunch”
- November 2012, “Avoiding the valuation traps in Strategic Asset Allocation”
- February 2013, “Searching for the Holy Grail in Asset Allocation”
- May 2013, “It’s not all about Income”
- August 2013, “CPI+3.5% white paper”
Data sources used in this paper
Data sourced from Global Financial Data:
- Australia Consumer Price Index
- Australia Total Return Bills Index
- Australia 10-year Government Bond Return Index
- GFD World Return Index
- Australia S&P/ASX 200 Accumulation Index
- S&P 500 Total Return Index (w/GFD extension)
Asset Allocations
Balanced | Conservative Balanced | |
Australian equities | 30.0% | 12.5% |
Global equities | 30.0% | 12.5% |
Fixed Income | 30.0% | 40.0% |
Cash | 10.0% | 35.0% |
Disclaimer
Opinions, estimates and projections in this article constitute the current judgement of the author as of the date of this article. They do not necessarily reflect the opinions of Schroder Investment Management Australia Limited, ABN 22 000 443 274, AFS Licence 226473 ("Schroders") or any member of the Schroders Group and are subject to change without notice. In preparing this document, we have relied upon and assumed, without independent verification, the accuracy and completeness of all information available from public sources or which was otherwise reviewed by us. Schroders does not give any warranty as to the accuracy, reliability or completeness of information which is contained in this article. Except insofar as liability under any statute cannot be excluded, Schroders and its directors, employees, consultants or any company in the Schroders Group do not accept any liability (whether arising in contract, in tort or negligence or otherwise) for any error or omission in this article or for any resulting loss or damage (whether direct, indirect, consequential or otherwise) suffered by the recipient of this article or any other person. This document does not contain, and should not be relied on as containing any investment, accounting, legal or tax advice.
^{1}Robert C Merton, “The future of retirement planning”, CFA Institute 2007 and 2008
^{2}“Risk and Risk Control in an Era of Confidence (or is it Greed?)” Remarks by John C. Bogle, Founder, The Vanguard Group in a speech to the New England Pension Consultants’ Client Conference, April 6, 2000
^{3}The terms “Conservative Balanced Fund” and “Balanced Fund” are commonly used in the industry. Assumed asset allocations for these strategies are shown in the Appendix and based on typical strategic asset allocations of such strategies. Conservative Balanced is also known as Capital Stable.
^{4}For a broader discussion on this point see “Its not all about income.”, Schroders, May 2013.
Important Information:
Opinions, estimates and projections in this article constitute the current judgement of the author as of the date of this article. They do not necessarily reflect the opinions of Schroder Investment Management Australia Limited, ABN 22 000 443 274, AFS Licence 226473 ("Schroders") or any member of the Schroders Group and are subject to change without notice. In preparing this document, we have relied upon and assumed, without independent verification, the accuracy and completeness of all information available from public sources or which was otherwise reviewed by us. Schroders does not give any warranty as to the accuracy, reliability or completeness of information which is contained in this article. Except insofar as liability under any statute cannot be excluded, Schroders and its directors, employees, consultants or any company in the Schroders Group do not accept any liability (whether arising in contract, in tort or negligence or otherwise) for any error or omission in this article or for any resulting loss or damage (whether direct, indirect, consequential or otherwise) suffered by the recipient of this article or any other person. This document does not contain, and should not be relied on as containing any investment, accounting, legal or tax advice. Schroders may record and monitor telephone calls for security, training and compliance purposes.