Chasing Top-Performing Funds May Be a Winning Strategy After All

It seems there is no end to strategies for investing in mutual funds.

The most obvious among them include market-timing (all in or all out), lump-sum investing, dollar-cost averaging , classic asset allocation, intuitive opportunism (switching around when things seem to warrant), buy and hold and sector rotation.

My own approach has always been to pick great managers who know how to be opportunistic within a disciplined strategy. I want managers who can play defense in bad times and offense in good times. Since it is not a perfect world, and consistent perfect timing doesn't exist, it is next to impossible to find that kind of manager. The few who come close: Bill Fries ((TVAFX Quote)Thornburg Value), Bill Miller ((LMVTX Quote)Legg Mason Value Trust), Glen Bickerstaff ((ENEAX Quote)Enterprise Equity), Chris Davis ((SLASX Quote)Selected American Shares), Jim Oelschlager ((WOGSX Quote)White Oak Growth), Spiras "Sig" Segalas ((HACAX Quote)Harbor Capital Appreciation), Helen Hayes ((JAWWX Quote)Janus Worldwide) and Mark Yokey ((ARTIX Quote)Artisan International).

Of course, there are others, but as my handball coach says, "You only need to execute enough of the right shots to win the game."

You can probably win the mutual-fund investing game with only four or five managers. Since some of these great managers will not stay great forever, I look at the job of managing a mutual-fund portfolio as twofold: First, select the best managers, and second, follow them like a bloodhound. When they don't perform well for a few months, I generally pull out and use another manager who does.

There is another unique strategy that may have some merit. It is articulated in great detail in an article in the August Journal of Financial Planning entitled, "Timing Patterns in Effective Periodic Investment Strategies." If you understand the title you can understand the rest. The authors, two University of Colorado professors, did an in-depth risk/return analysis of a number of mutual-fund portfolio strategies, looking for better long-term performance with a reasonable degree of risk.

The three strategies they analyzed:

• Passive strategy, in which an equal amount of money is invested each period in each of six fund categories. This is most closely associated with dollar-cost averaging.

• Winner-take-all strategy, in which money is continually reallocated to the one fund offering the highest return over the most recent period, be it monthly, quarterly or annually. This is an active strategy that cites a body of evidence that says past performance and rank are reliable predictors of future performance.

• Low-cost strategy, in which all funds are invested in the one fund category with the lowest return for the most recent period. This strategy is based on the assumption that low-return funds will cost less.

Returns for these strategies were compared for a period from January 1977 through December 1992, and are summarized in the chart below. For each strategy, money was invested or reallocated either monthly, quarterly or annually, based on the previous month's mutual-fund returns. So, for example, the winner-take-all strategy was tried three ways: In the first, money was reinvested monthly in the previous month's best-performing fund. In the second, money was reinvested quarterly in the previous month's best-performing fund. In the third, money was reinvested annually in the previous month's best-performing fund.

	Monthly	Quarterly	Annually
Passive Strategy	14.23%	14.23%	14.23%
Winner-Take-All	17.45	18.20	16.85
Low-Cost Strategy	9.19	6.32	8.59

The study concluded that the winner-take-all strategy worked the best, especially when moves were made on a quarterly basis. "The evidence suggests that a simple portfolio strategy of allocating funds on a quarterly basis in the fund category with the highest prior month's return is effective in providing mutual-fund investors with superior wealth accumulation with only moderate increase in risk," the authors conclude.

The study did not take into account transaction costs and taxes -- which would eat away a significant portion of the gains if this strategy was tested outside of a tax-deferred retirement account like an IRA.

My take on this strategy is that it is an interesting synthesis of some sector rotation and momentum investing. Especially enticing is the fact that it imposes a well-defined discipline to an otherwise intuitive process. It is hard to go against the numbers, and I am curious how they would play out with real money and a real investor.