By
Yes, Workshop screens. What, did you think I was talking about? Christmas gifts from Amazon.com <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: AMZN)") else Response.Write("(Nasdaq: AMZN)") end if %>? Sorry, that would be the Rule-Breaker Portfolio. Putting cash in a Workshop screen is more like Buying a Car. You'll want to check it out thoroughly before sinking your hard earned cash in it, and you only keep it as long as it runs well.
Below we have the "performance specs" for four of our most popular screens. Some historical return information is already available, courtesy of Elan Caspi (TMFElan). The returns below are updated to include 1998 data, and go into more detail. You'll notice that, unlike your car, your workshop screen will probably be worth a lot more money in 10 years than it is now.
Let's look first at PEG, arguably the most popular screen in the Workshop. Created by Moe Chernick, this screen combines elements of traditional growth strategies with a "value" factor. The PEG screen has been a superb performer overall since 1986. Below are the returns for the top five slots of the PEG screen (a slot is the rank of each stock as determined by the screening technique). The "A" column represents annual holds, "S" is semi-annual, "Q" is quarterly, and "M" is monthly. All returns represent the arithmetic averages of returns across all starting months.
These returns now include all start months for 1998, so they vary slightly from those published in last week's article. Also, note that the S&P 500 has returned about 17.8% from 1/1/86 to 12/31/98.
What do we see? Annual and semi-annual screens have better results from a five stock portfolio than from any individual slot. The semi-annual returns are considerably higher than annual returns, both for each individual slot and for the portfolio as a whole. This makes PEG a good candidate for a semi-annual hold, especially in a tax-advantaged account. By contrast, monthly returns are very high for a four stock portfolio, but this is mainly due to one slot (PEG3). This indicates that the monthly screens returns might not hold up over the long run. This characteristic, combined with the higher costs of monthly screens, indicates that monthly trading of PEG is probably not a good idea, in spite of the high CAGR.
Next up, let's look at the Spark screen. Created by Sparfarkle, this is a classic growth screen inspired by James O'Shaugnessy's Invest Like The Best. This was the first growth screen introduced on the Workshop board as an alternative to Robert Sheard's Unemotional Growth screen.
Sparfarkle intended for this screen to be traded annually, and the results seem to bear this out. An annual hold does almost as well as the other holds. Note also, that five stock annual and semi-annual screens do better than any one individual slot. An investor could do very well using the Spark annual screen in a taxable account, and a Peg semi-annual screen in an IRA. Also, Spark generally selects large-cap stocks, while PEG will often select smaller companies.
Speaking of large cap companies, let's look at Keystone. This screen looks at "timely" mega-cap stocks, and takes those with high Relative Strength (RS). RS simply measures the total return of the stock over the last 26 weeks (or 13 or 52 weeks, depending on which version you select), relative to all other stocks. This screen was originally designed for annual holds, and is often cited as a "safer" growth screen. Here are the average results.
Speaking of Relative Strength, the last screen we'll look at closely is RS-26. This screen simply picks the stocks with the highest relative return over the last 26 weeks that are also ranked number 1 for Timeliness by Value Line. RS screens have attracted a lot of attention in the workshop this year, mainly because they've done very well recently. Remember that the results below do not include 1999 starts.
Many mechanical investors trade RS screens monthly, but the data above shows that RS-26 works quite well as an annual. In fact, one could easily argue that quarterly trading is better, given the reduced trading costs and good returns for all 5 slots (and again, a whole five stock screen is better than the sum of its parts).
One topic that gets discussed a lot in the Workshop is "seasonality." Many of our annual or semi-annual screens tend to do better when started in January than in other months. Obviously, you don't have to worry about this with a monthly screen, since you're always rebalancing in January (and February and March and...). The last chart will show the returns for various 5 stock annual portfolios started each month.
PEG and Spark all have done significantly better in January than in other months, with PEG screens started April 1st or October 1st actually underperforming the S&P 500 from 1986 to 1999. By contrast, RS-26 has actually performed best when started in October! Keystone appears to have had the most consistent returns of all, although Keystone exhibits some seasonality around the Dec.-Jan.-Feb. period.
Two final cautions for would-be Workshop investors. First, please remember that all of these returns have occurred during a limited period when the market has done very well. Pre-'86 results on some screens are beginning to be presented, thanks to huge collaborative effort by many Workshop regulars. The results appear promising, but aren't nearly in the 25%+ range. Not surprising, since the S&P only returned 8.9% from January 1, 1969 through the end of 1985.
Second, the results above do not include measures of volatility. One PEG5 annual screen started in October of 1997 lost nearly 60% over the next year, and RS26 has occasionally lost 20% (or more) in only one month! Even the best screens can have very bad times.
So, how can one use all this information? We'll give you an example next week, when I choose the screens for rebalancing my (pitifully small) IRA. Unlike Moe's Real Money Port, this will only have two screens.
Until then, a very happy Winter Solstice to everyone this Holiday season.
For more on the screens above see:
Rank A S Q M
1 11.36% 21.23% 13.03% 13.28%
2 20.87% 32.38% 41.43% 23.25%
3 24.58% 31.86% 40.39% 73.06%
4 13.58% 14.97% 12.37% 32.82%
5 14.72% 17.27% 17.65% -9.08%
1 to 4 28.69% 35.75% 36.07% 43.80%
1 to 5 28.67% 34.59% 35.16% 34.32%
2 to 5 29.78% 35.13% 37.61% 36.88%
Rank A S Q M
1 22.31% 22.08% 18.86% 21.04%
2 18.16% 18.92% 19.44% 14.99%
3 20.55% 26.80% 22.88% 20.03%
4 17.02% 17.70% 28.57% 29.12%
5 20.52% 27.22% 37.96% 25.02%
1 to 4 24.08% 25.67% 26.17% 24.88%
1 to 5 25.32% 27.29% 29.73% 26.37%
2 to 5 24.73% 27.29% 31.36% 26.56%
Rank A S Q M
1 34.13% 33.90% 32.37% 39.86%
2 25.30% 23.04% 25.61% 17.69%
3 19.80% 18.03% 18.46% 20.84%
4 21.95% 19.84% 18.73% 16.19%
5 17.13% 16.24% 14.41% 15.36%
1 to 4 29.59% 27.72% 27.46% 26.20%
1 to 5 28.02% 26.20% 25.45% 24.53%
2 to 5 23.80% 22.14% 22.16% 19.66%
Keystone's annual returns have been slightly better than Spark's. Keystone is also the first screen we've seen that does best as an annual hold. This is a little surprising for a screen that relies heavily on RS.
Rank A S Q M
1 20.70% 24.07% 24.43% 33.57%
2 8.97% 24.53% 32.83% 51.55%
3 21.80% 25.61% 22.54% 5.25%
4 21.13% 28.58% 32.11% 40.82%
5 17.97% 22.82% 33.46% 28.39%
1 to 4 30.10% 36.40% 38.13% 42.07%
1 to 5 29.92% 35.47% 39.09% 41.12%
2 to 5 27.15% 34.36% 39.15% 39.74%
Screen Jan Feb Mar
PEG 41.19% 34.59% 28.39%
Spark 36.89% 29.09% 18.86%
Keystone 27.69% 31.24% 19.75%
RS-26 26.99% 34.87% 31.02%
Screen Apr May Jun
PEG 14.62% 19.95% 33.14%
Spark 20.27% 26.81% 23.99%
Keystone 31.17% 27.16% 26.47%
RS-26 20.51% 32.68% 25.73%
Screen Jul Aug Sept
PEG 35.47% 30.70% 27.50%
Spark 27.51% 18.01% 25.80%
Keystone 26.20% 30.49% 26.64%
RS-26 28.49% 19.46% 28.70%
Screen Oct Nov Dec
PEG 15.71% 27.06% 35.73%
Spark 25.59% 24.63% 26.39%
Keystone 24.15% 29.05% 36.29%
RS-26 43.09% 28.64% 38.85%