Close

    Asset Allocation and Risk: How to Build Your Own Portfolio

    by James B. Cloonan

    In my past two columns, I pointed out the problems with many current approaches to asset allocation, in particular their failure to directly address the problems of risk control. I also looked at the nature of allocations over the past 10 years to see how effective the process has been. [See A Matter of Opinion, September and October 2002.]

    The primary way to control and reduce risk is through diversification. The theory behind diversification is that if you divide investment assets among different investments that have low correlations with each other, this will reduce risk.

    Asset allocation is an effort to achieve efficient diversification by dividing the individual investments among various asset classes and subclasses. However, in my previous articles, I showed how this effort is often misguided—asset allocation approaches do not necessarily lead to truly diversified portfolios. Well, then, what’s the solution?

    Diversification: Important Concepts

    I have not previously presented approaches for controlling risk through portfolio design. And I will do this. But first, I would like to point out some major concepts that underlie any approach.

    Volatility Affects Returns

    First, the variability of the return on a portfolio’s assets, which is the accepted definition of risk, affects not only the “sleep well” aspects of the portfolio, but also the real return. The example in Table 1 illustrates what I’m talking about here. The two portfolios in this example have the same average return over the four periods. The average return I am talking about here is the arithmetic mean—the individual returns over the four time periods, divided by four. This is not the same as the compounded average return (the geometric mean), which is the return that determines an actual investor’s ultimate wealth because it includes the effects of compounding—the return on the prior period’s return that an investor would receive from having his money continuously invested over all four time periods.

    Looking at each period individually, you can see Portfolio A has much less volatility than Portfolio B. The volatility for the two portfolios is measured by the standard deviation, which is the amount by which most returns varied around the arithmetic average for the whole period; the higher the standard deviation, the greater the volatility. For both Portfolio A and B, the average return (the arithmetic average) is the same. Yet an investor will actually wind up with more wealth from Portfolio A because it has less volatility. Portfolio A has a higher compound average return (geometric mean) than Portfolio B, because Portfolio B’s geometric return is reduced by volatility.

    Diversification Affects Returns

    Second, while effective diversification will reduce risk, it may reduce return as well. This is because not all potential investments have the same likelihood of gain. If you didn’t care about risk, you would put all of your assets into the single investment you thought had the likelihood of the greatest gain. But in order to diversify, you must put some assets in investments that you expect will have less potential return. At the same time, however, you anticipate that the reduced risk will be worth the reduced gain.

    Trying to reduce risk while giving up as little return as possible is the critical part of the strategy I suggest, and I’ll be emphasizing that in the approach that follows.

    Cap-Weightings Affect Returns

    Third, although I use the risk and return data of the indexes of various asset classes as a surrogate for the risk and return of the various classes themselves, portfolios chosen by astute investors from these asset classes will have higher returns and lower risk than the indexes themselves. This is because, on average over the long term, the lower the market capitalization of listed stocks, the higher will be the return. The indexes I use here all capitalization-weight their holdings—the percentage holding of each stock is based on the stocks’ market capitalization, with larger market cap stocks representing a greater percentage in the index. Because of the cap-weightings, these indexes will tend to underperform the average stock in that index asset class. On the other hand, if you use mutual funds instead of individual stocks, the risk and return may be closer to the indexes, since many mutual funds cap-weight their portfolios. In addition, the larger capitalization stocks tend not to be value stocks, yet value stocks, as defined by their lower price-to-book ratios, outperform growth stocks. As an example of this effect, over the last 10 years the Wilshire 5000 index has had an average annual return (through September 30, 2002) of 8.69%. However, the return of the Wilshire 5000 unweighted by market capitalization (in other words, the average stock in the Wilshire 5000), is 14.17% a year. And to throw in another shocker: The Nasdaq Composite index is down dramatically year to date, with a return of –39.26%, yet the average stock on the Nasdaq is up 30.54%, and the average stock has been up each year for the past three years. You can monitor this difference at www.wilshire.com at no charge (click on Indexes, then Index Return Calculator).

    Note, too, that the risk in the weighted indexes is increased because the stocks that are most heavily weighted tend to be correlated, and thus diversification among these larger stocks has less impact. Small- and micro-cap indexes also are capitalization-weighted, and equally weighted individual stock portfolios chosen from these asset subclasses will have higher returns than the indexes. Table 2 shows correlations among the various indexes. Numbers approaching 1.0 indicate greater correlation in return performance; negative correlations indicate that returns tend to move inversely. When one index has a positive return, the other tends to be down.

    Whatever the risk and return levels of the indexes we are about to talk about, you can expect that your portfolio, if not cap-weighted, will have a lower risk and a higher return.

    Your Portfolio Risk Level

    How can you develop a portfolio that establishes a risk level suitable for you and offers the highest returns while maintaining that level of risk?

    There are two steps:

    • The first is to establish a level of risk you can live with, and
    • The second is to choose the individual assets and asset classes that will provide the highest return within that risk level. Table 2, along with Table 3, which shows asset class returns and volatility, can provide some guidance.
    Setting a Suitable Risk Level

    There are a number of ways you can approach setting a suitable risk level. However, I suggest an approach that is relatively intuitive.

    I start with the assumption that if your holding period is less than three years, you should not be in stocks, but in low-risk debt instruments with a due date corresponding to when you need the money.

    Assuming you are a long-term investor, let’s examine the history of the stock market since World War II. I think the world’s economic system changes gradually, and that 1946 is far enough to go back for historical guidance.

    Since 1946, what is the worst that could have happened to your portfolio if all you held was the S&P 500 index?

    During the 1973–74 period (based on calendar years), you would have lost 37.4% of your investment in the S&P 500.

    If we assume that year-end 2002 will see the S&P 500 where it is now (September 30), the 2000 through 2002 period will result in a loss of over 40%.

    For our purposes here, it is good that we are in one of the down periods. When discussing risk in past periods, investors tend to say they would have ridden it through because they know it turned out all right. But we’re still in a bad period, and the S&P 500 could go down more from where it is now. Did you ride through this downturn or did you reduce your equity holdings, indicating your portfolio risk was inappropriately high?

    Let’s say the S&P 500 will have several periods of a 40% drop during your investment lifetime. The S&P has a long-term standard deviation of 18%, and a RiskGrade of 90, so this should provide an indication of what standard deviation or RiskGrade you are willing to assume. [RiskGrades provide a measure of portfolio risk, with a RiskGrade of 90 implying a risk 90% as high as the average for world equities—click here for more on RiskGrades.]

    Contemplating occasional losses in the range of 40% and knowing how you felt during this past year should enable you to set an appropriate risk level. For example:

    • If you can face a roughly 40% drop every 25 to 30 years and believe that in the long run stocks will prosper, you don’t need to be in any asset class outside of U.S. stocks, although you will still want to diversify to keep your risk as low as possible while obtaining the long-term high return of stocks.

    • If, on the other hand, you can only face a maximum of a 20% loss, you will need a portfolio with a standard deviation of about 9%, or a RiskGrade of 45. It will be difficult to reduce risk much below this level without moving out of stocks almost completely.
    Let’s assume for this discussion that you want to keep portfolio decreases below 25%, which implies a portfolio standard deviation of below 11.2%, or a RiskGrade below 56. You may recall from my previous column [A Matter of Opinion, October 2002] that an allocation of 60% stocks, 40% bonds would give us this level of risk if we used the existing indexes of those two asset classes. This would have provided a return of 8.73% per year over the past 10 years.

    With any effort at all, you should be able to beat that return with the same level of risk. In fact, in the October 2002 column, I gave examples of many allocations that would have done better using other indexes. But as I just discussed, selected portfolios in the various asset classes will usually outperform cap-weighted indexes.

    TABLE 3. Asset Class Return and Volatility
    Asset Class Index Ticker 3-Yr Avg
    Annual
    Return
    (%)
    Standard
    Deviation
    (%)
    10-
    Yr Avg
    Annual
    Return
    (%)
    Total Stock Market Vanguard Total Stock VTSMX -9.63 16.74 9.82
    Foreign Vanguard Total International Stock VGTSX -10.45 15.22 5.68
    Large Cap Vanguard 500 Index VFINX -10.67 15.75 10.31
    REITs Vanguard REIT Index VGSIX 12.78 13.76 12.69
    Gold Vanguard Precious Metals VGPMX 11.84 35.38 4.03
    Large-Cap Growth Vanguard Growth Index VIGRX -13.78 18.51 8.93
    Large-Cap Value Vanguard Value Index VIVAX -8.34 15.9 13.01
    Small-Cap Growth Vanguard Small-Cap Growth Index VISGX -0.41 27.99 6.23
    Small-Cap Value Vanguard Small-Cap Value Index VISVX 5.59 22.27 14.82
    Total Bond Market Vanguard Total Bond Index VBMFX 8.05 3.35 7.05
    Short-Term Bond Vanguard Short-Term Bond Index VBISX 7.2 2.11 6.52*
    Micro-Cap DFS U.S. Micro Cap DFSCX 6.53 33.02 13.54

    Getting the Highest Return at a Given Risk Level

    Once you have selected an appropriate level of risk, the next step is to create a portfolio that provides the highest return for that risk. Of course no one knows what the future holds, so the selection of individual assets must be based on your own estimates. But the information here should give you some idea of which classes and subclasses you should examine. In studying the data in Tables 2 and 3, remember that results for the past 10 years are somewhat different than over the past 50, and the last three years have been very atypical. The correlations in Table 2 are close to those for the very long run even though there may be short-term variations.

    Let’s look at how you might create your own portfolio.

    On paper, you will usually get the best return at a given risk level by combining the highest return asset class with the safest class in the proportion that gives you the desired risk level. For instance, looking at the tables, if this were 10 years ago, you could have selected a portfolio consisting of 90% REITs and 10% short-term bonds, maintained your RiskGrade at 60 and had a compound return of 12% a year. That combination also would have been best for the last three years.

    It would have been hard to predict that result in advance, however, and asset class returns can change in a hurry. Even if you monitor your portfolio closely and adjust frequently, things can change too quickly to be in only one equity sub-class. Remember how quickly energy stocks collapsed and tech stocks before that?

    So how do you go about selecting which classes to be in?

    The first clue is the correlation between classes as shown in Table 2. The only negative correlations are between debt and equity classes. So to reduce risk significantly, there will need to be debt issues in your portfolio. The percentage that needs to be in debt depends on what else you choose.

    In the Table 2 correlation table and the Table 3 risk and return table, short-term bonds dominate long-term bonds in terms of their risk/reward trade-off. The small loss of return is well worth the larger reduction in risk. In fact, if your investment horizon is three years or longer, short-term government bonds have no risk. In my opinion, there is no reason for individual investors to ever own long-term or intermediate-term bonds in a balanced portfolio.

    In selecting which other asset classes in which to invest, I would look at those that have had the highest returns in the long run, with additional weight given to those also doing well in the short term. From the top classes look at the correlations and try to choose those with low correlations to each other. This is not a pure numbers game, nor is there one best way because the returns of each class for the future are not known. You can then use the RiskGrades Web site to check out the risk for your chosen portfolio.

    My Simplified Approach

    Let me go through a more simple thought process for selecting asset classes, simply based on my own preferences and biases.

    For the reasons given above, I would eliminate bonds, except for short-term bonds. Based on results from over the last 70 years, from this year, and from all long-term periods in between, I would eliminate all stocks on the S&P 500 index—along with any index that includes those stocks. I realize there will occasionally be a great opportunity in an S&P 500 stock, but there are more opportunities elsewhere. If you are screening for individual stocks, you can simply eliminate stocks with over $1 billion market cap.

    The correlation between foreign and U.S. stocks is too great for there to be meaningful diversification and the loss of return is too great a price for such little risk reduction, which I believe is temporary anyway. So that eliminates foreign stocks.

    I would personally not choose gold mining stocks even though the correlations with other asset classes are low. I feel that the risk reduction effect is very short term and the long-term return reduction is severe. However, I can understand why it might be appropriate under some circumstances.

    I would not choose small-cap growth stocks because of their performance and risk relative to small value stocks, and their high correlation with micro-cap stocks.

    Basically then, I would select a portfolio made up of REITs, small-cap value, micro-caps, and value stocks with market capitalizations below $1 billion but larger than small caps.

    One caution: When selecting value stocks, choose a real measure of value rather than the classification that splits all stocks into either growth or value. One criteria would be that the price-to-book ratio must be below 1.5, and lower is better.

    In terms of the percentages to invest in each category, I would invest an approximately equal amount in each, and then after having examined the risk of such a portfolio, I would invest enough in short-term bonds to reduce the risk level to one appropriate for me.

    Diversifying by Style

    An entirely different approach to allocation and diversification among stocks is to diversify among strategies for selecting stocks. You can use different established stock selection approaches or you can design your own. The focus should be on choosing approaches that emphasize different criteria for selection. Three to five approaches should provide excellent risk reduction, but you should analyze the risk to see if the diversification is effective.

    Specific Investments

    The topic I have not discussed is how to choose the specific investments within each class.

    The first step, that of choosing a pool of potential stocks in each asset class, has been the subject of thousands of books. You can design your own approach or you can rely on the advice of others.

    However, there is a second step that is quite important, and that is diversifying the individual issues within each class as much as is practical. The degree to which you are able to do this will reduce the percentage of low return, safe investments you will have to add to the portfolio to maintain your required risk level—some sacrifice of return in diversifying is worthwhile.

    One effective approach is to select, by whatever means you have chosen, more stocks in an asset class than you actually want. For example, if you want 10 micro-cap stocks, choose 16 and then pick the 10 that have the lowest portfolio risk. You can do this by going to the RiskGrade Web site and using “what if” processes to find the lowest risk group of 10 [click here for more information on RiskGrades].

    You should also review the entire portfolio after you finish choosing individual investments in each class. Say you have four asset classes (not including your bond portfolio) and have selected 10 stocks in each of them. You can set up a RiskGrade portfolio and see if all of the individual issues are necessary. You may well find that many of the individual stocks are highly correlated even if they are in different classes, and you can eliminate the ones you feel will produce the lowest returns.

    Portfolio Monitoring

    Once a suitable portfolio has been established, it is necessary to monitor and adjust it. With three or four equity asset classes, it will not be necessary to check risk levels constantly and, in fact, short-term variations should be ignored. If the risk of your portfolio starts to increase and continues to increase over a month or more, you should look at the components, see where the additional risk is coming from, and make adjustments as necessary. And, although in this column I am not discussing ways to choose stocks, such a strategy must also be monitored. Monitoring can be done by looking at the components of your portfolio on RiskGrades.com, or by looking at the risk levels of the mutual fund surrogates listed in the tables here. The risk of the surrogate mutual funds can be found at Morningstar.com or other sites.

    Bottom Line: Risk Control

    I hope this and my preceding two columns have at least indicated the problems with current approaches to asset allocation and pointed the way to better methods of controlling risk. We have not covered every aspect of the problem of risk control, but I hope the material has provided some insight into how effective approaches can be developed.

    Analyizing 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 through standardization—it takes the average of all the world’s equities, and assigns it a standard deviation of 100. All other standard deviations are expressed as a percentage of that figure. For example, a portfolio RiskGrade of 77 implies it has a risk 77% as high as the average risk of all equities in the world.

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


    James B. Cloonan is chairman of AAII.


→ James B. Cloonan