By
That's fine, but why go to all that trouble? Of the choices listed above, U.S. stocks have historically performed the best. If I just want stable returns, I'll go out and buy U.S. Treasury Notes or deposit my savings in one of Vanguard's money market funds. As for valuable paintings and commemorative coins, they make perfectly good hobbies, but not good investments.
Collectibles have huge transaction costs associated with purchase and sale, not to mention the insurance cost while you own them. Also, unlike money market funds, you can't write a check on the value of your Monet. Real estate has similar issues (if you've ever bought or sold a home, you know that it's not quick or easy). And gold has dropped about 50% in value since 1980. So much for being "safe."
Here at The Motley Fool we believe that all your long-term savings are best invested in the U.S. stock market. Over the long term, carefully chosen and diversified stocks have been as safe as fixed-income investments, and the return has been much higher.
Bonds or money market funds are the right place for your shorter-term savings (money you'll need in less than five years), because over the short term, stocks can be extremely volatile. At the Foolish Workshop, we believe that mechanical stock strategies are an excellent way to select stocks. But they are not the only way to invest, which is why we also have Rule Breaker and Rule Maker portfolios, among other Stock Strategies. As an investor, you have to find a style of investing that is right for you.
Although we believe that mechanical investing is an excellent way to build a stock portfolio, it's not perfect. As we mentioned last week, our screen strategies have performed substantially better than the S&P 500 since 1986 (and in the case of the Relative Strength screens which have been tested longer, since 1969), but with much greater volatility than the S&P 500 index.
So here's where we get back to diversification. One way to potentially reduce the volatility of a mechanical portfolio is by using several different screens. For example, let's say that Strategy A returns 40% in odd-numbered years, but returns -5% in even-numbered years. By contrast, Strategy B returns -10% in odd-numbered years, and 45% in even-numbered years. Obviously, investing in one of these strategies is a wild roller-coaster ride. However, if you were to split your funds between the two screens, your results would look something like this:
Year StratA StratB Total Return
1 40% -10% 15% (40%/2 + -10%/2)
2 -5% 45% 20% (-5%/2 + 45%/2)
3 40% -10% 15% (40%/2 + -10%/2)
4 -5% 45% 20% (-5%/2 + 45%/2)
5 40% -10% 15% (40%/2 + -10%/2)
6 -5% 45% 20% (-5%/2 + 45%/2)
7 40% -10% 15% (40%/2 + -10%/2)
8 -5% 45% 20% (-5%/2 + 45%/2)
And voila, your wild ride has become a smooth series of 15 to 20% annual returns. If you're the type who thinks in terms of pictures, imagine two sine waves, each with offsetting peaks, superimposed over each other to make a much straighter line.
Of course, real-life isn't nearly this perfect. Our strategies aren't nearly this predictable. However, the basic idea (using multiple screens to smooth out the bumps) should be a central tenet of any mechanical investor's portfolio. This "combination" of screen strategies is often referred to in the Workshop as a Blend. You can backtest various blends for yourself at (where else?) Jamie Gritton's Backtest Engine. Also, Moe Chernick recently showed a real-life example of a blend.
So, what blends are likely to work well? First, you should be careful not to choose screens that are substantially the same. For example, if you were building a portfolio of individual stocks, you would be careful to choose stocks in different, unrelated industries. You would hardly consider yourself diversified if you owned three stocks, and they were Ford (NYSE: F), General Motors (NYSE: GM), and DaimlerChrysler (NYSE: DCX). Along the same lines, a mechanical portfolio which uses an RS13 strategy and an RS26 strategy would not be very diversified.
A good workshop portfolio uses a mix of screens that choose different types of stocks. How can you tell if the screens "track" each other too closely? One way is to look at the selection criteria of the screen. Moe's Real-Money Portfolio is an excellent example of this type of diversification. Moe uses a mix of "value," "growth," and "momentum" screens which often go in different directions. However, all of the screens have historically outperformed the market.
A more precise, "mathematical" way of comparing screens is by taking the returns of the various screens, and figuring the correlation coefficient between the two. What, you may ask, is a correlation coefficient? Well, correlation is a measure of the relation between two or more variables (like our screen returns).
The correlation coefficient shows the degree to which the two variables are related. A correlation coefficient is a value between -1 and +1. A value of +1 would reflect a perfect positive correlation (Screen A goes up 10% when Screen B goes up 10%). A value of -1 would indicate a perfect negative correlation (Screen A goes down 10% when screen B goes up 10%). A coefficient of zero would mean that there is no apparent relation between the two (Screen B goes up 10%, and Screen A does something completely random).
Just for fun, I calculated the correlation coefficients for the Keystone, PEG26, Plow, RS26, and Spark strategies (annual holds, averages for all start months). Here's a chart showing the results:
Key PEG Plow RS26 Spark
Key XXX 0.21 0.74 0.58 0.46
PEG XXX 0.29 0.39 0.48
Plow XXX 0.57 0.53
RS26 XXX 0.45
Spark XXX
As you can see, all of the correlations were positive, not surprising since all the strategies are invested in the same market. But the PEG screen is the most independent, with coefficients that range from 0.21 to 0.48.
One surprise is the close relation between Keystone and Plowback. The correlation coefficient between those two strategies is 0.74. That's a lot higher than we might expect, since the two screens use much different factors to select stocks. This suggests that that checking correlation coefficients may be much more useful than simply looking at the screen criteria.
I hope to do more with correlation coefficients and post the results on our message boards. If you want to learn more about correlation coefficients (as well as all the other statistical jargon we toss around in the Workshop), then check out this site. (Thanks to Russ DeGarmo, rwde on the boards, for that link.)
Until next time, Fool On!
RSS Headlines
Fool UK
