Why SAA is flawed
Most investors have as their investment objective the desire to achieve a real return outcome over a defined time period. The industry approach to meeting the investment objective has been to set a broadly fixed strategic asset allocation (SAA, eg: 60/40) and implement this along single asset class lines (typically in a multi-manager structure for each asset class).
In this paper we examine the portfolio outcomes from a traditional strategic asset allocation based investment approach and investigate what is required to achieve a typical balanced/growth investment objective of a real return of 4% to 5% p.a. over a 5 to 10 year time frame. The historical analysis in this paper supports the proposition that real returns from capital markets consistent with this objective can be achieved over the longer term – ie. the level of the return objective is broadly right in the very long term. However, the historical evidence does not provide the same level of support for real returns to be delivered consistently over 5 to 10 year time frames.
Given the impact of the accumulation (and decumulation) process of most savers, achieving return objectives over very long term time horizons is simply not enough to meet their requirements. Individuals need greater comfort that their investment return objectives will be delivered over more shorter horizons. This will require a rethink of the strategic asset allocation dominated approach to investing.
Most investors have as their investment objective the desire to achieve a real return outcome over a defined period. For example, a typical investment objective for a “balanced” or “growth” investment option may be outlined as:
“To outperform the Consumer Price Index (CPI) by 4% per annum over rolling 5 year periods after tax.”
After stating the relevant investment objective, the typical industry approach would then be to define the method to achieving these objectives as a broadly static allocation to a range of asset classes. At a high level, a typical balanced fund with a CPI + 4[(-5) was not found] objective is likely to have an allocation to growth assets of 60-70%.
Conceptually, investors are trying to balance three competing objectives as shown graphically below. We outline in this paper how it is not possible to achieve these at the same time. One must be relaxed. Given that the fixed SAA is the means to an end and therefore the one investors should be indifferent to, the common sense approach would be to drop it and so improve the ability to meet individual’s investment objectives.
The purpose of this paper is to consider the extent to which the fixed strategic asset allocation approach (which is the mainstay of the investment portfolio construction of most investors) will be able to meet the typical balanced fund investment objective and to understand the real risks introduced from such an approach. To do this we analysed a ‘stylised’ traditional balanced fund using asset class data commencing in 1900. Such long term modelling was able to provide a foundation to suggest whether the ability to achieve a real return objective is realistic over long periods of time.
Importantly we sought to shed light on the frequency and duration of failure of the traditional strategic asset allocation approach to meet this objective.
Interestingly, we note that some industry participants have elected to broaden their investment objectives to include an additional objective (or sometimes the sole objective) of outperforming the “median fund”. In our view, this is perhaps partly a reaction to the last decade’s inability of consistently achieving rolling real returns and a reflection on how most of the industry really manages investment options. That is, with an eye on what everyone else is doing and the ubiquitous league tables! Unfortunately, from an individual’s perspective the “objective to outperform the average of others” is relatively pointless, conveying little real meaning and generally resulting in poorer outcomes (through herding). From a risk management perspective we would argue that it should provide little confidence to individuals (and indeed the regulator) to then define investment risk relative to others.
Analysis and Results
To test the characteristics of a traditional strategic asset allocation based portfolio and validity of a target real return objective of 4% to 5% p.a. analysis was conducted on a sample of asset class returns over the last 110 years. The index data used in our analysis is summarised in Appendix A. The data period is seen as a useful representative sample as it covers such economic events as depression and boom, environments of rising inflation and deflationary forces.
The representative portfolio used to conduct the analysis we describe as a ‘stylised’ balanced portfolio. It contains an exposure to growth assets of 60% (half Australian and half international equities), and defensive assets of 40% (one quarter in cash and three quarters in bonds). We chose this structure as it is broadly representative of long term investment portfolios with real return objectives of 4% to 5% p.a.
It is fair to observe that the portfolio is perhaps not as well diversified across asset classes as many current long term portfolios. However, we favoured the long term availability of data to model long term outcomes over a more diversified portfolio (with exposures to property, overseas bonds and alternative asset classes) which is more constrained in terms of sourcing long term data series. In any case, most new “asset classes” are not in fact asset classes but rather derivatives of more mainstream listed asset classes above.
The purpose of this analysis was to understand whether long run return drivers over the last century would have given rise to a satisfactory real rate of return. Given that our modelling used monthly return series, we made no further assumptions about asset allocation positioning and therefore rebalanced the portfolio at its strategic weights at monthly intervals.
The chart on the following page shows the rolling 5 and 10-year real returns of the stylised balanced portfolio.
Stylised balanced fund - rolling 5 and 10 year real returns1
Source: Schroders, Global Financial Data, Shaded areas represent periods below 5% p.a. objective
1US equity returns are used as a proxy for global equities up until 1925 and then Global Financial Data’s global equity series is used.
The annualised compound real return since the start of the 1900s has been 5.3%p.a. above inflation which places it comfortably above the 4% to 5%p.a. real return objective of most investors before any implementation costs, or active management outcomes are taken into account. However, another critical observation is that the performance is delivered in long term regimes, with the portfolio underperforming a 5% real return objective 49% of the time on a rolling 5 year basis and 47% of the time on a rolling 10 year basis. The worst rolling five year period provided a -10.8% p.a. real return and the worst rolling 10 year period -3.7% p.a. real return.
The analysis above shows the rolling 5 and 10 year returns through time. We can see from these charts that for some of these 5 and 10 year rolling periods there are periods of time, potentially quite long, when the real rolling return can be below zero or, put another way, real returns have been negative for the prior 5 or 10 years.
Given that an individual may start investing at any point in time, the use of rolling period returns does not reflect the initial experience of that investor. Historical returns may have been strong and so rolling period returns look relatively good, however the immediate future return may turn out to be quite poor. In this situation, rolling period returns can remain strong and positive for a period of time however the new investor does not have this prior experience and so starts with a negative outcome.
In order to examine the potential negative return experience faced by a new investor we can look at a “drawdown” analysis. This analysis shows the cumulative maximum loss faced by an investor who starts investing at the previous market peak. Essentially we consider the experience of a theoretical investor who commences at the point real returns from the strategic asset allocation turn negative. In this way we ignore prior periods of outperformance and just examine how much the investor would have lost in real terms until such point when the market regains its previous high and that investor has recovered the real value of their savings.
What the drawdown analysis does is examine the maximum loss that would have been made historically for any investor unlucky enough to have started saving at the market peak. Recognising that at any point in time a typical approach will have new individuals starting an accumulation process or starting the decumulation phase there is a strong potential that at least some cohort of the investor base will experience these drawdowns.
Stylised balanced fund – Cumulative Real Drawdown
Source: Schroders, Global Financial Data
Relative to inflation, it is clear from the above that a traditional strategic asset allocation can result in substantial loss of purchasing power for long periods of time. While eventually such drawdowns are recouped, the time horizon for this is in some cases over a decade and in most cases many years. We can see in the figure below the top 5 drawdowns in real terms over the last 110 years and the time in years taken to recover from these. However we note that this is just recovery in real terms, it takes no account of the need to then deliver a positive real return. In any case, the magnitude of these drawdowns has been significant along with the time taken to recover.
Stylised balanced fund – Top 5 Real Drawdowns in Last 110 Years
Source: Schroders, Global Financial Data
For individuals where the underlying value of the capital may change given contributions or pension payments (that is, for pretty much everyone), such long time periods until recovery can have a substantial negative impact on accumulated balances (or account based pension payments).
Looked at another way, we show below the decade by decade real returns of the traditional strategic asset allocation approach together with the respective returns for the defensive (bonds and cash) and growth (equities) parts of the portfolio. We can see from the chart that the “lumpiness” of equity returns effectively creates intergenerational inequity. For example the members of the 1910’s, 40’s, 70’s and 2000’s (to date) have effectively subsidised the equity returns for members of other generations. This is not a framework for arguing the consistent achievement of individual objectives!
We also note from the chart that the 1930’s depression is “hidden” in Australia (it wasn’t in many other countries) and that the nature of the monthly rebalancing process meant that the fixed strategic asset allocation gave a slightly better return than either of its defensive or growth components in this period.
Stylised balanced fund – Decade by Decade Real Returns
Source: Schroders, Global Financial Data
We can observe from the above chart that:
- The dispersion in returns between growth and defensive assets has at times been quite large;
- The 1930’s and the 1990’s were the two periods where asset allocation mattered less, albeit for the 1930’s we also know that while the decade returns were similar the path taken to achieve that was very volatile for equities; and
- In only 6 of the last 11 decades did the strategic asset allocation portfolio meet or exceed objectives (not exactly a strong hit rate when we are considering decade long periods).
How long do I really need to be sure of meeting my objectives?
One argument commonly adopted by the industry is that despite the objective of generating more consistent medium term outcomes, in reality it is the long term that counts. As we outline above, over the very long term the strategic asset allocation process has been able to generate satisfactory real return outcomes, so why should we be concerned about this medium term variability of returns? In particular, the long term nature of retirement savings means that shorter term time horizons are just not that relevant.
To this end, it is worthwhile analysing the historical probability of achieving certain real return outcomes over longer periods of time.
The chart below shows what level of real return would have been achieved 90% of the time or 80% of the time for a given time horizon.
Historical probability of achieving real return outcomes
Source: Schroders, Global Financial Data
The above chart shows that even with a 40 year time horizon we would have achieved a rolling real return outcome of CPI+4.3.% only 80% of the time. If our time horizon was 20 years and we required a 90% probability the return target was only CPI+2.2%. That is, one in every 10 individuals over 20 years would have achieved less than CPI+2.2%.
Historically if our target was CPI+4% we would have required a time horizon of 53 years to achieve this with 90% probability. For a CPI+5% target with 90% probability required 85 years. This also suggests that in fact the achievement of these returns is not so much a statistical probability that will occur given enough time, but rather a statistical anomaly in that we have only achieved these returns from this asset allocation because the return on equities (particularly Australian equities) has been very high.
While the 110 year return from a 60/40 portfolio has been circa 5.3% p.a., to have a reasonable degree of certainty of achieving our desired outcome the time horizons required exceed most investors lifetime accumulation periods. Controlling medium term volatility becomes even more important when we take into consideration the accumulation and decumulation process and the resultant impact on money weighted returns. We address this in a forthcoming paper.
A significant observation in respect of the analysis conducted in this paper is that we are utilising for a large part Australian biased assets. As such we should note:
“Whether it is down to luck or good economic management, Australia has been the best-performing equity market over the 112 years since 1900, with a real return of 7.2% per year.”
Long Term Real Returns of Equities for Developed Markets
Credit Suisse Global Returns Yearbook, 2012, Real returns from 1 January 1900 to 31 December 2011.
As such, we would suggest that the results of this study are biased upwards in favour of equities. Were future years not to be as good as history (which, on the balance of probability is likely) then very Australian equity biased portfolios are unlikely to perform as well as outlined above even over the very long term.
Consequently, to consider a more normalised view we consider a similar analysis to that shown above from the perspective of a US investor with 60% US equities and 40% US bonds the results of which are set out below. Recognising that US equity returns have also been relatively good over the last 110 years, just not as good as Australia, we can see that the potential underperformance of the traditional strategic asset allocation portfolio is quite large.
Stylised balanced fund – Decade by Decade Real Returns US Investor
Source: Deutsche Bank, “A third generation in Asset Allocation”, Brad Jones, Jan 2012
While this simple study shows that a 4% to 5%p.a. real return has been achievable over very long periods, it also shows that it has been much more difficult to achieve over shorter periods, including the 5+ year time horizon typically included in many investor’s objectives. In order to achieve objectives consistently, substantially greater flexibility in asset allocation would have been required historically.
The purpose of this paper was to investigate the suitability of the fixed strategic asset allocation approach to achieving individual investment objectives.
The historical analysis in our paper suggests evidence exists to support the proposition that real returns from capital markets are consistent with the objectives of the typical approach of up to 5% p.a.
However, the historical evidence does not provide as much support for real returns consistent with many individuals objectives to be delivered reliably within the required timeframes for fixed asset allocation portfolios. Instead fixed asset allocation portfolios require a very long term time horizon, given that equity markets in particular have delivered real returns in long term cycles or ‘regimes’. This also has important implications for the money weighted returns earned by individuals.
In summary if the industry is to achieve the investment objectives it communicates to individuals, a substantial rethink of the approach is required. Fixed strategic asset allocations generate significant medium term volatility of outcomes, making them unsuitable for consistently achieving objectives. In the trade-off between delivering real returns over constant time frames with a fixed strategic asset allocation, something has to give. It is in this context that we have demonstrated above that fixed strategic asset allocations don’t work. Given the unsustainability of fixed SAA portfolios it follows that the lifecycle type approaches that follow a pre-defined asset allocation glidepath are also unlikely to be any more useful in achieving investor objectives. We address the issues of lifecycle funds and money path dependency (money weighted returns) in more detail in a forthcoming 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)
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.