[The following article is based on The Mutual Fund Bible, * 1996, 1997 by Mark Edward Newsome and Thomas Holyoake Box]
Asset allocation is based on the fact that both the probable return and the probable volatility of each asset class is different. By combining asset classes in different proportions, it is supposed to be possible to modify a portfolio’s overall volatility and return. Our view is that its appeal is based on false premises, and that it is in fact a misguided approach to portfolio management.
The Appeal of Academic Legitimacy
Much of asset allocation’s appeal lies in its theoretical credentials. While a great deal of investment wisdom is anecdotal, asset allocation has its basis in scholarly research. In research which earned him a Nobel Prize in Economics in 1990, Professor Markowitz demonstrated as early as 1959 that by combining investments which were negatively correlated, it was possible to reduce a portfolio’s overall volatility.
Correlation is the degree to which investment returns vary in tandem. When two investments are perfectly correlated, changes in the return of one investment are exactly equaled by changes in the return of the other. When there is no relationship between the returns of the two investments, there is no correlation. However, when changes in the return of one investment are matched by equal but opposite changes in the return of the other, there is perfect negative correlation.
In practice of course identifying investments with perfect negative correlations is impossible, because the relationships between investments are never completely stable. As it turns out however, the most durable negative correlations are found between asset classes. It is therefore easiest to manage portfolio volatility by varying the allocation of assets among the different asset classes. Research on portfolio performance conducted in the 1980s appeared to corroborate these conclusions. In a 1986 study of 91 large pension funds, Gary Brinson, Randolph Hood and Gilbert Beebower concluded that, on average, more than 90% of the variation in portfolio performance could be attributed to the portfolio’s asset allocation, while only about 7% of portfolio performance was related to the selection of specific securities. Yet despite the apparently impressive academic pedigree, the actual evidence in support of asset allocation is considerably less compelling.
First, notwithstanding the claims of its adherents, asset allocation has an inherent short-term bias. The objective of the process is to minimize volatility, but because volatility is conventionally measured as the variation in monthly returns, the process of allocation places disproportionate emphasis on minimizing short-term fluctuations in return. In fact however, volatility is time-frame dependent. As the length of the investment period increases, the volatility of investments typically decreases. What’s more, the volatility of stocks decreases more rapidly than the volatility of either bonds or T-bills. For holding periods longer than 20 years, a fully diversified stock portfolio is on average less volatile than either a bond portfolio or a T-bill portfolio. Moreover, as Professor Jeremy Seigel points out in an exhaustive review of historical data, this is even more true when the results are adjusted for inflation.
The implication of this fact is that for long-term investors, asset allocation is counterproductive. On the one hand, far from decreasing volatility, asset allocations combining stocks and bonds increase the portfolio’s long-term volatility. On the other hand, since the assets with low short-term volatility also have low long-term returns, the process of asset allocation can actually reduce an investor’s long-term returns. Second, the conclusions of the ‘Brinson’ study have been widely misconstrued. Most observers have understood the study to demonstrate that asset allocation explains 90% of portfolio returns. This is wrong. The actual conclusion was only that most of the variations in return were attributable to the portfolio’s asset allocation. In other words, the asset classes in which a portfolio invests will have a greater effect on variations in return than the specific securities chosen from those asset classes. The study does not address the issue of absolute portfolio returns. Since it is well known that the short-term volatility of stocks is greater than that of bonds, this conclusion is rather unremarkable.
Furthermore, there is a world of difference between identifying after the fact the ‘best’ asset allocation, and identifying the best mix in advance. The fact that asset allocation can explain market performance does not mean that it is possible to predict what the best allocation will be for the future: indeed, it is extremely unlikely that anyone will be consistently right about the best asset allocations.
The Appeal of Precision
Another reason for the appeal of asset allocation is that it seems simultaneously simple and precise. An investor determines the required level of return, and then mixes and matches assets until the desired return is achieved with the lowest possible level of volatility. Investors are thus able to tailor a portfolio which precisely meets their needs. In a most common scenario, the high volatility of stocks is ‘moderated’ or tempered by the incorporation of less volatile bonds in the portfolio. Mutual funds are ideal for this purpose because they are frequently perceived as convenient proxies for different kinds of asset classes, making the process of mixing and matching both convenient and (relatively) inexpensive.
There are two problems with this point of view. The first is the idea that the portfolio should be understood as a undifferentiated whole for the purpose of managing “risk”. The reality is that there is an important difference between an investor’s objective tolerance for risk, and his or her subjective comfort level. By distinguishing between them, it is possible to meet multiple objectives without reducing the performance of the portfolio to the level of its lowest common denominator. The second problem is that asset allocation isn’t nearly as precise as its disciples pretend. The reality is that when asset allocators project a rate of return for a portfolio, what they really mean is that the projected rate of return will be somewhere around the anticipated rate, most of the time. If for example the anticipated rate of return for a portfolio is 10% per year (0.85% per month) with a standard deviation of 2.71 the reality is that the rate of return will be somewhere between 3.56% and -1.86% per month two-thirds of the time, and between 6.27% and -4.57% per month one-third of the time. There is a one-in-six chance that the monthly return will be below -1.86%.
Proponents of asset allocation argue that on average the actual return of the portfolio will correspond to its projected return, but this argument misses the point. If the exercise were about averages, all investors would buy stock portfolios, since stocks outperform all other assets on average. In fact, it is not averages with which the investor is usually concerned, but with the potential downside, and it is clear that the protection provided by asset allocation is less than certain.
Asset Allocation as Marketing Tool
The truth is that the preoccupation with asset allocation is motivated far more by marketing interests than by its investment fundamentals. The apparent simplicity of the process has a powerfully seductive appeal, and has been enormously successful in attracting additional assets. The most obvious evidence of this marketing bias is the proliferation of new ‘asset classes.’ A portfolio composed of stocks, bonds and t-bills is now regarded as inadequate. To be ‘properly’ diversified, investors need more elaborate portfolios, including:
- large and small cap domestic stock funds
- US stock funds
- European stock funds
- Pacific Rim stock funds
- Latin American or emerging markets funds
- domestic bond funds
- global bond funds
- high yield bond funds
- real estate funds
- commodity futures funds
To the average investor however this multiplication of asset classes is utterly impractical. To manage this burgeoning list of funds, many fund companies now offer asset allocation “services” which assist investors in identifying the most suitable proportion for each asset class. While this is, on the surface, a useful addition, the vast majority of these services require an investor to commit all their resources to a single fund group.
The services are therefore extremely effective in increasing the assets on which fund management is able to charge fees, but limit the investor’s ability to seek out the most suitable funds. Even more telling is the fact that most of these services require investors to choose from a (relatively) limited number of different combinations of ‘packages’. This greatly simplifies the exercise for the fund companies which sponsor such services, but renders the process inherently impersonal.
A third example of the emphasis on marketing is the extension of the asset allocation principle to the field of investment strategies. Many fund groups now offer investors the opportunity to diversify management styles. Mackenzie Financial was the first to offer this feature for its equity funds, by organizing their funds into three groups: the Industrial group, which uses a “top-down” investment approach; the Ivy group, which uses “bottom-up” approach; and the Universal group, which provides access to a variety of external managers.
More recently, CI Mutual Funds and Spectrum United have joined the party. CI has introduced a fund using a bottom-up approach managed by Jonathan Baird to complement their existing top-down based fund managed by John Zechner. Spectrum United has three different portfolio managers for its Canadian equity funds. Sun Life of Canada (which owns Spectrum United, and is the “in-house” manager) combines top down and bottom-up investment styles. AMI Partners uses a bottom-up value approach. C.A. Delaney combines the bottom-up styles of growth and value.
A variation on this theme is the so-called “multi-managed” fund, in which two (or more) different managers share responsibility for a single portfolio. Both Global Strategy and 20/20 Financial Group (now part of AGF) offer funds managed in this way.
While it is true that every manager’s “style” produces better results in some market conditions than in others, these differences largely offset each other over long time periods. There is no good evidence that “diversification of management styles” actually reduces investor risk, but there is also no doubt that the idea helps sell funds.