The Daily Workshop
Report
by Robert Sheard
(TMF Sheard)
LEXINGTON, KY. (Sept. 11, 1997) -- In our Workshop description of the Formula90 approach, we've listed a ten-year annualized return of 36% for a five-stock portfolio and a 35% return for a ten-stock version. As a reminder, this is a screen where you begin with the top 100 stocks in the Value Line ranking system (Timeliness ranking of 1). Then, only the stocks with EPS (earnings per share) and RS (relative strength) rankings of 90 or better from Investor's Business Daily are considered. This group is sorted by RS ranking, using EPS as a tie-breaker where needed.
A reader asked, however, whether I had run a similar test reversing the order of the screens. In other words, select from among the Formula90 stocks based on EPS first, using RS for tie-breakers. I couldn't remember whether I had done that test or not, but ran the numbers again today. As I thought I had remembered from a year or so ago, the results weren't as strong this way.
Using the EPS screen first generates an annualized return for a five-stock approach of 29% and for a ten-stock approach of 23% per year. I also ran a ten-stock version using the sum of EPS and RS rather than either one by itself. The returns for that test equaled 27%.
So of the obvious options with this approach for an annual holding period, using the RS rankings seems the best based on the past decade. As I've said on a number of occasions, with only ten data points, it's hard to make a very strong claim for these results, regardless of what they say. I haven't tested this screen for other holding periods, however, and the "best" factor might change as one changes the holding period.
What it might suggest is that EPS screens lose their effectiveness the longer you push out the holding period. As Jim O'Shaughnessy has suggested in his research, the market, while not the most efficient in the short run, is ultimately efficient enough over the course of a year that the EPS screen loses its effectiveness. Good earnings numbers are priced into the stocks quickly enough to ruin this screen for long-term holding periods. Over an annual holding period like the one I tested for Formula90, the relative strength screen appears more useful.
Monthly Growth Screens (Jan. 3 to present) 78.37% Relative Strength 34.06% Investing for Growth 25.23% YPEG Potential 22.51% EPS Plus RS 22.28% Low Price/Sales 22.01% S&P 500 Index 15.18% Unemotional Growth 13.28% Formula 90 Annual Value Screens (Jan. 1 to present) 21.72% Dogs of the Dow 19.41% Dow Combo 18.81% Dow Jones Ind Avg 18.43% Unemotional Value 18.43% Beating the Dow 17.14% Beating the S&P 13.48% Foolish Four