Thursday, November 06, 1997
The Keystone Portfolio
by Robert Sheard (TMF Sheard)
In recent months, I've been tracking a portfolio screen in our Foolish Workshop geared towards Blue Chip stocks with the best earnings prospects and price strength. But there's no permanent home for an explanation of the Keystone Portfolio Approach yet, so I'll Fribblize it today.
The process. Using the Value Line for Windows software, I screen for the 100 largest American stocks. To narrow that field of 100 quickly, I only accept those stocks with Value Line Timeliness rankings of 1 or 2. This typically filters out a group of 25 to 30 stocks. Then as the final ordering screen, I rank these stocks by their previous Six-Months' Total Return and select the top ten (or however many you may want). The holding period for this approach is one year.
The theory: I'm attracted to large-caps for several reasons. These are the Bluest of Blue Chips. They're typically market leaders, financially stable, long-term success stories. Not too many companies make it into the 100 largest stocks by screwing up regularly. It's not a guarantee of a great investment, of course, but it's a quality and stability screen that affords somewhat conservative investors a measure of comfort. You know these companies and they aren't likely to tank overnight.
But within that group, I want the stocks with the best and most consistent earnings-growth trends. That's where the Value Line Timeliness rankings come in. The stocks ranked 1 and 2 have thrashed the market returns since Value Line began this system in 1965. That's a real-time track record I take comfort in.
But I also want the stocks performing well today. So within that select list, I focus on the stocks with the highest Relative Strength. I opt for a Six-Month Total Return screen over the RS ranking from Investor's Business Daily for two reasons. First, I think the last six months provides a more sensitive test than does IBD's twelve-month time-weighted ranking. And second, for purely practical reasons, I can perform the Total Return screen in seconds using Value Line for Windows where using the IBD numbers requires a manual check of each of the thirty stocks in the newspaper. (Above all else, I like approaches that are simple and quick. Each Friday when I get the latest Value Line data update, the entire screening process can be performed in less than a minute.)
The Results: Let me begin with a big disclaimer; I have very little data for this approach so far and am looking to back-test the approach in the near future. So take this for what it's worth -- preliminary experiments only. That said, I like what I see. I've done a bit of testing on large-cap stocks in general and found that when combined with the Timeliness rankings, they perform quite well over the last decade.
This particular screen I've only been able to track recently, however. Beginning with the end of June 1996, I began following the screens and in each of the five completed tests so far (a new portfolio each month), the entire field of stocks with rankings of 1 and 2 out-performed the S&P 500. The stocks with the highest Relative Strength, however, did so even more impressively. Here are the returns for the completed tests so far (neither the Keystone Portfolio nor the S&P 500 Index numbers here include dividends). This result is for a ten-stock portfolio.
6/30/96 to 6/30/97 Keystone 48% versus 32% for the S&P 500 Index.
7/31/96 to 7/31/97 Keystone 72% versus 49% for the S&P 500 Index.
8/31/96 to 8/31/97 Keystone 44% versus 38% for the S&P 500 Index.
9/30/96 to 9/30/97 Keystone 51% versus 38% for the S&P 500 Index.
10/31/96 to 10/31/97 Keystone 46% versus 30% for the S&P 500 Index.
After 10/31/96, there's a gap in my data I'm trying to fill. The next monthly starting point I was able to test begins on 4/30/97. And I've been following each new month since then. Obviously, this is too short a test to make any definitive claims, yet I'm pleased that the theories seem to be borne out in other similar tests I've performed. As I can extend this history by back-testing, I'll keep you posted.
One final question that comes up frequently is about the holding period. I checked the very limited data I have and found that an 18-month holding period hurt the annualized gains. So even with a lower tax rate, this may not be a good option. My theory is that since the choice of stocks is partly based on recent price strength, the extra six months waters down the effectiveness of that element. I'll continue to check this as I compile more data.