The third real-world test consists of the actual records of funds-of-funds - mutual funds that select other mutual funds for their portfolios. And these records are the most deplorable of the lot. The funds-of-funds not only lag the market - we now know that five of every six funds have done that - but they seriously lag even the style categories of the funds in which they invest, in part because of the extra layer of costs they almost universally add. For example, in the year ended June 30, 1998, of the 14 funds-of-funds investing in large-blend (value and growth) funds, 4 ranked in the 96th to 100th percentiles (one was dead last) and 5 ranked in the 90th to 95th percentiles. The champions, if that's the right term, of this undistinguished group ranked in the 65th percentile, lagging two-thirds of the peer funds from which they made their selection. In all, the 93 funds-of-funds with full-year records achieved about what might have been expected from a random selection of funds that was reduced by an added layer of costs of more than 1 percent: an average of the 68th percentile compared to their regular fund-style peers.

I present the one-year results only because so few funds-of-funds have been around very long. Over the past decade, when only nine of them existed at the outset, the record was a bit worse: they lagged 69 percent of the funds in their peer group. But excluding the single fund that did not add a layer of extra expenses (it outpaced 72 percent of its peers), the ranking quickly dropped down, with the remaining eight funds-of-funds achieving only a 75th percentile ranking among comparable regular funds. This neighborhood is hardly posh but is surely familiar, clearly reaffirming the one-year numbers presented above. To make matters even worse, managed funds-of-funds typically turn over their own fund portfolios at an average rate of about 80 percent per year, a short-term focus that inevitably impinges on the long-term returns they earn. The combination of high fund turnover and high fund costs, with two extra layers of cost - from high turnover and excessive operating expenses - has clearly proved to be a formula for failure.

Given the transitory nature of the one-year data, and the existence of a limited number of funds over the pasts ten years, perhaps the most relevant evidence is found in the three-year data. The past three years give us the opportunity to examine 35 funds-of-funds that have existed during the period, comparing each with its peer group: 11 large-cap, medium-cap, and small-cap stock funds; 4 international stock funds; 16 balanced (hybrid) funds, and 4 bond funds. The average fund-of-funds achieved a 66th percentile ranking, closely confirming the one- and 10-year findings. The average fund-of-funds returned an average of 15.5 percent for the period, a 2.4 percentage point shortfall to the average return of its peer group. Since more than half of this lag is created by the average expense ratio of 1.3 percent (1.7 percent excluding those funds that levy no additional fees) they added on, clearly the experts managing them had no particular selection ability sufficient to offset the costs of their services. Figure 9.2 shows the ranking of the 28th funds-of-funds among the 35 fund total that added on such fees in terms of their percentile rankings. It clearly reflects the powerful odds against successful fund selection for expensive funds-of-funds. It is a loser's game.

• Index Funds Explained
• Active Management and Indexing
• Advantages of Index Funds