Mirror mirror on the wall, who's the fairest of them all?
This quarter's analysis uses a new set of indexes developed by Ron Surz, president of PPCA, San Clemente, Calif., instead of the Russell indexes. There are three reasons for the switch:
* The Russell indexes only consider half of the stocks in the market, the largest 3,000;
* Of that 3000, 30% are included in both the value and growth indexes; and
* Russell rebalances once a year.
The Surz indexes are exhaustive, mutually exclusive and rebalanced quarterly. Mr. Surz puts that 30% of stocks in both the growth and value indexes into a separate "core" category.
The Frank Russell Co. constructed its indexes as a product to sell, and not primarily for use in style analysis. Therefore, the decision to put some stocks in both the value and growth indexes was quite reasonable. Mr. Surz, on the other hand, constructed his indexes so they met the statistical requirements of the returns-based style analysis procedure. They are exhaustive in that they include all stocks in the Standard & Poor's Compustat database and mutually exclusive in that no stock is in two categories.
Accuracy key
Beauty may be in the eye of the beholder, but fairness is not. It is not fair to fund managers or those who hire them to misrepresent their risk-return characteristics. I rely on the indexes to identify the inherent risk and upside potential in a fund's style. The more accurately I can identify the fund's style, the better estimates I can get of the upside potential and downside risk characteristics of that fund. While many of the funds ranked tops by my criteria using the Surz indexes also rank near the topusing the Russell indexes, there are too many contradictions.
For example, the Fidelity Dividend Growth fund is ranked high with the Surz indexes but doesn't make the cut using the Russell indexes. This is due in large part to Surz having a core category and Russell putting those stocks into both value and growth. The graph titled "Style Composition" shows the average fund in this analysis has 12% large core and 12% small core. T. Rowe Price Blue Chip is 51% core. Also, the four Russell indexes together account for less than 50% of the variance in returns of the Surz core index.
There is no similar problem with large-cap growth and large-cap value, as these Surz indexes are highly positively correlated with the Russell indexes.
All funds in the ranking table are first ranked by upside potential ratio (U-P). Those funds in the top quartile with a positive Omega excess are marked with an asterisk. These are the funds I use in my asset allocation model. They are the finalists in the beauty contest. The fairest of them all is ... in the eye of the allocation model.
Frank Sortino is president of Pension Research Institute, Menlo Park, Calif. The fund data were provided by LCG Associates in Atlanta.
