Strategic asset allocation – set and forget or dynamic?
This edition of Investment Insights discusses strategic asset allocation (SAA) and the pros and cons of the two commonly applied SAA methodologies - dynamic versus static.
Tuesday, November 25th 2003, 12:00AM
The three-year bear market in global stocks has naturally caused investors to consider how they invest their money.
One issue subject to debate is how to manage the strategic asset allocation (SAA) for investment portfolios. Given SAA is the dominant driver of performance, the rough experience of the past few years has raised the question as to whether it should be set on a fixed (or static) basis over time or whether it should be varied dynamically to reflect clear mis-valuations between asset classes.
Textbook definitions
In a classical textbook definition, strategic asset allocation refers to the long term or benchmark asset allocation to each asset class in an investment portfolio. Tactical asset allocation (TAA) refers to any deviations from this long-term benchmark position. As such, SAA is primarily concerned with the long term return and risk objectives of the investor, while TAA is concerned with adding value to this objective.
Difference of view
The two schools of thought in relation to how SAA should be set are characterised as follows:
Static SAA - this approach assumes that there is some constant long-term return differential between each asset class. In other words, there is a constant risk premium between bonds and cash, between shares and bonds, and so on. In essence this approach implies that markets are always efficient and that the real return expectation from any level for share prices, bond yields, etc, is always the same. As a result, the SAA can be set on a once-only basis and there is nothing to be gained by varying it. This concept is epitomised by the 'time in not timing' catch-cry.
Dynamic SAA - this approach starts from the assumption that markets are not efficient, and as a result, the starting point valuation or price at which you invest in any asset class will dramatically affect the return achieved over relevant time horizons, both on an absolute and a relative basis. In essence it assumes that investors cannot wait 20 years to get a long term target rate of return but rather the SAA should be set on something like a three to 10 year horizon, using currently available valuations as a starting point.
The difference between static SAA, dynamic SAA and TAA can be explained in the following chart, which shows the 95% confidence band for actual US share returns since 1900 for various investment horizons. Over this period, 95% of six month annualised returns have varied between -45% and +70%. As the investment horizon lengthens, the confidence interval steadily narrows and beyond 10 years it is relatively stable. While there is no precision, TAA is generally trying to capture return variations over six to 12 month periods, while static SAA is working on the fact that over long investment horizons the confidence interval for returns is fairly stable. However, between the extreme volatility of the six to 12 month horizon and the stability of the investment horizons far beyond 10 years, there is scope for dynamic SAA.
US S&P 500 Total Returns - 1900-2003
Source: Global Financial Data and AMP Henderson Global Investors
Static SAA - pros and cons
The main advantages of the static approach are that it makes no pretensions as to the ability or skill required to allocate between the different asset classes over time and it is relatively easy - basically set and forget. The only drivers of change are changes in the investor's objectives or risk tolerance.
The main disadvantages though, flow from its underlying assumption that markets are efficient. If markets are efficient then asset prices will always reflect all currently available information and price changes will be random as they reflect new information which is, by definition, unforecastable. As a result the price you pay for an asset should have no impact on the return you can expect both in an absolute sense or relative to other assets. However, there is a mounting array of evidence to suggest that markets are not efficient - that the price you pay for an asset can have a massive impact on the return achieved, particularly over realistic return horizons, and that future medium term returns are predictable to some degree. Key findings in violation of the ‘efficient market hypothesis’ are that long periods of high returns tend to be followed by long periods of low returns and that high valuation multiples signal low future returns ahead and vice versa. These observations hold true at medium to long term horizons, eg, five to 10 years.
The bottom line is that markets can go through periods of bear markets and bull markets that extend well beyond the investment time horizon of many investors. This means that in reality there is no constant risk premium between asset classes, but rather that it fluctuates dramatically over the medium term - as evident in the next chart for US shares and bonds.
US rolling 10yr return differential - equities less bonds
Source: Global Financial Data and AMP Henderson Global Investors
None of this would be a problem if investors' time horizons were 20 years or more - but they are not.
Dynamic SAA - pros and cons
The main advantages of the dynamic approach to SAA are that it is consistent with the evidence that markets are not efficient and that they go through long-term bull and bear phases, and it recognises that investors' horizons are not really long term.
That said it still possesses several draw backs. Firstly, even over long periods, market timing can be hard. Secondly, significant skill and effort is required to consistently get markets right even on a medium term basis. Finally, the dynamic approach may result in significant changes to SAA over time, as relative valuations and prospective medium term return projections move in response to events like share market crashes. The key to implementing a successful dynamic approach therefore requires a disciplined framework to predicting future medium term returns.
An approach to dynamic SAA
Under a dynamic approach to SAA, expectations of future medium term returns need to be formed for key assets and incorporated. Our analysis indicates that, for shares, a simple model of current dividend yields plus nominal GDP growth does a good job of predicting medium term returns. For property, a similar approach works well using current rental yields and inflation as a proxy for rental and hence capital growth. For bonds the best predictor of future medium term returns is the current bond yield. Our analysis indicates that using this approach over the past 100 years to drive a dynamic SAA process produces a superior result to a static SAA approach - although not necessarily over all periods.
Conclusion
The decision by investors regarding how they conduct their strategic asset allocation is an important one. The key is to be aware of the pros and cons of both the static and dynamic approaches to SAA. While evidence that markets are not efficient and our own empirical research indicating that dynamic SAA can add significant value over time provide a strong case for the dynamic approach, it must be recognised that this approach is not without risk and that to be successful it requires significant skill and resources.
Changes in markets, investor liabilities and risk tolerance, and new asset classes mean that strategic asset allocation should not be a set and forget decision. We believe that a fund or an individual investor's SAA strategy should be reviewed annually.
This article was written by Dr Shane Oliver, Chief Economist and Head of Investment Strategy of AMP Henderson, Australia.
| « Market Review : The Challenges and Opportunities in China | China - The Roaring Dragon » |
Special Offers
Commenting is closed
|
|
Printable version |
|
Email to a friend |
