Popular Allocation Approaches Put the Cart Before the Horse

    by James B. Cloonan

    I am amazed at all of the articles and Web pages devoted to “asset allocation.”

    While they all take the same general approach, they disagree strongly on the details. The general approach is to mandate that at a certain age, you should have a certain percentage of your wealth in each of stocks, bonds, cash, and sometimes other asset classes such as real estate. Sometimes they further divide each class into subclasses. For example, stocks might be divided into: large vs. small, domestic vs. foreign, or value vs. growth. They will then briefly mention that this reduces risk, but there will be no analysis of how it does this or proof that it works. The emphasis will be on the debate over whether a 50-year-old should have 50%, 60%, or 70% in equities.

    Most of the articles treat the topic as if asset allocation were an end in itself, rather than one possible tool in the creation of a portfolio that provides the highest return for a given level of risk. This is akin to arguing about the best road to take before you decide where it is you want to go.

    The Wrong Emphasis

    Those of you close to my age remember Adele Davis, the famous nutritionist. When she would lecture, invariably she would be asked something like, “Are tomatoes good for me?”

    Her response would be, “Which tomatoes, grown where?”

    My response to an author advocating a certain percentage in bonds vs. stocks is “Which bonds, which stocks?”

    Take a portfolio that is 70% invested in an S&P 500 index fund and 30% in a bond index fund, and calculate its risk. Well, I can find a different 100% stock portfolio that is less risky than the 70%/30% indexed portfolio, and I can also find a 100% bond portfolio that is more risky.

    Allocation may be helpful, but the key to risk control is the selection of the individual assets in the portfolio. I firmly believe that diversification within asset classes is more important than the allocation between classes. Choose individual investments first, allocate weightings later.

    More Problems

    In addition to this general problem of mistaken emphasis, there are some special problems with the usual approach to asset allocation. These include:

    • Stocks and bonds sometimes move in the same direction, sometimes in opposite directions, and sometimes independently (over the last 70 years, their correlations have run from –0.3 to +0.6, where 1.0 means perfect positive correlation and –1.0 means perfect negative correlation). Any allocation between them must adjust to this reality.

    • Foreign stocks in a portfolio may reduce short-term risk, but this effect dissipates over a few years as economic conditions travel. And there is a significant return reduction for the temporary benefit.

    • Diversifying between value and growth stocks would make sense if we are talking about true value and true growth stocks. However, the current approach of pigeonholing all stocks into either value or growth doesn’t make sense. I am a value investor, and I can’t find even 15% of stocks that qualify as value by any reasonable definition. And I know growth/momentum investors, and they certainly would not be investing in stocks with a relative strength below the 85th percentile (which would mean the stock’s price performance was not among the top 15% of all stocks over that period). By that rough calculation, we have 15% value stocks, 15% growth stocks and 70% “something else” stocks. If you allocate between a value index fund and a growth index fund, you are primarily investing in “something else” stocks, which may not have the risk-reducing effect you want.

    • In any approach to controlling risk, investment horizon and risk aversion should be important factors. The use of age as a surrogate for investment horizon is too simplified. Some people retire at 62, while many others work into their 80s. Many individuals do not plan on using up all assets during retirement, but want to pass their wealth on to heirs. So the level of risk any investor should assume at any point in time depends on more complex factors than age.

    • The allocation of wealth across stocks of different sizes as a method of controlling risk has limitations. Typically, the size of a stock is measured in terms of market capitalization (share price times number of shares outstanding), and classified as micro, small, medium or large. Since over the long run micro-cap stocks outperform other stocks, particularly large caps, it may make more sense to look for portfolio diversification within the small-cap, micro-cap and perhaps mid-cap markets. For individuals who use mutual funds, there is a real danger of over-investing in large-capitalization stocks and of not being adequately diversified because most funds weight their holdings based on market capitalizations (the larger the company, the more of the stock they buy). An S&P 500 index fund will have 23% of its assets invested in the largest 10 stocks. Not only is this too much concentration, but the largest stocks often have high correlations. So, you may think you are getting diversification, but you aren’t. AAII’s 18-stock Beginner’s Portfolio has less volatility (risk) than an S&P 500 index fund. [For an explanation of the experimental Beginner’s Portfolio, see “The AAII Beginner’s Portfolio: An Annual Performance Review,” August 2001 AAII Journal; available on AAII.com.]
    Allocation between certain asset classes can be a tool to reduce risk. We must, however, get away from simplistic approaches and look at all the factors that influence the risk of a portfolio. And, most importantly, we must consider the nature of the individual assets in each class before we decide on the allocation between classes.

    Table 1. Asset Allocation—Calculation of Benchmark
    Micro-Cap Stocks*** 12.6 75
    T-Bills 4.5 0
    S&P 500 Index 10.2 90
    Desired Risk Level na 60

    Allocation Benchmarking

    Let me discuss one process for measuring the effectiveness of any asset allocation approach. This analysis can be carried out using the RiskGrades Web site.

    Take the highest return asset class, which is micro-cap stocks. You can use the DFA U.S. Micro Cap stock fund (ticker symbol: DFSCX) as a surrogate. Combine this fund with enough Treasury bills to reduce the risk to your appropriate risk level, where risk is measured by the RiskGrade (based on standard deviation, the amount by which most actual returns over a given period varied around the average return). You can combine the return of the micro-caps and the return of the T-bills in the chosen proportions to determine the overall return of this portfolio. You will then know the highest return per unit of risk that you can achieve at your desired risk level. This is your benchmark, and any potential asset allocation program you choose should be able to beat this benchmark to be acceptable (see Table 1). Note that the RiskGrade for this particular benchmark is based on performance as of June 30, 2002.

    This is a method of measuring allocation across general asset classes, not of particular portfolios. You can certainly find individual portfolios with a higher ratio of return per unit of risk, but it will be difficult to find an asset class diversification program that is more effective over the long run.

    A Step Beyond

    I have really only tried to point out the problems with simple asset allocation approaches. The means of finding the highest returns for a given level of risk and adjusting allocation and diversification over time is more than enough material for a book.

    I will take asset allocation analysis one step further, however, in Part Two of this series, with examples of how you can evaluate various asset allocation methods using the RiskGrade methodology and the wide assortment of index funds.

       Analyzing Your Portfolio Risk On-Line
    To measure the overall risk of your portfolio and the effectiveness of your diversification, go to the RiskGrades Web site (www.riskgrades.com). You can enter your entire portfolio—including stocks, bonds and mutual funds—and determine its risk and its diversification efficiency. You can also compare it to several indexes in terms of performance and risk. And you can determine the amount of return per unit of risk, to see if you are being compensated enough for the level of risk you have taken on.

    RiskGrades uses standard deviation as the basis for its risk measurement, but makes it more meaningful. It standardizes the somewhat meaningless levels of standard deviation by taking the average of all the world’s equities, and assigning it a standard deviation value of 100. All other standard deviations are expressed as a percentage of that figure. For example, if the RiskGrade of a portfolio is 77, it implies it has a risk 77% as high as the average risk of all equities in the world.

    The Web site provides the mathematical details of the approach, and it is free for individual investors.

    James B. Cloonan is chairman of AAII.

→ James B. Cloonan