Excerpted from Common Sense on Mutual Funds by John C. Bogle, pages 124-127

Given these variations between the composition of the capitalizations of the stock market (75 percent of the value of which is represented by large-cap stocks) and the number of funds in the equity fund universe (50 percent of which is represented by large-cap funds), how do we arrive at a fair basis for comparison? One simple, rudimentary way is to compare the results of the Wilshire 5000 (all-market) Equity Index with only those funds whose portfolios have weightings similar to the total market. As it turns out, mutual funds emphasizing stocks with large market capitalizations meet that standard. Their performance can be approximated by combining all diversified growth funds and growth-and-income funds, and excluding small-cap and aggressive growth funds. The net result is about as fair as it can possibly be: funds emphasizing large-cap stocks, but not to the exclusion of all others, and an index that also emphasizes large-cap stocks, but, again, not to the exclusion of all others.

During the past 15 years, the average return of these large-cap-oriented funds, which I'll refer to as growth funds and value funds (rather than growth and income funds), has averaged 14.1 percent, compared to 16.0 percent for the Wilshire 5000 Equity Index. Cumulated over the period, this 1.9 percent difference, applied to an initial investment of $10,000, results in a final value of $72,600 for the average fund, a shortfall of more than $20,000 to the $92,700 that would have been accumulated in the Wilshire 5000 Index. (The return on the Standard & Poor's 500 Index, a much tougher standard during that time period, averaged 17.2 percent, for a final value of $107,800.)

Only 33 of the 200 growth and value funds that survived the 15-year period outpaced the Wilshire 500 Index during this period; the remaining 167 funds fell short. The odds of fund superiority were thus one in six. Even more interesting and, I think, more significant, is the variation of fund returns around this average. We can get some sense of the significance of the differences among funds by arraying their fund returns around the returns of the Wilshire 5000 Index, as shown in Figure 5.5.

If we adjust our thinking and agree that a spread of less than one percentage point above or below the Index return is merely statistical noise, the odds shift: 15 funds toppled the Index by more than that margin, compared to 125 funds that lost to it by the same margin. The odds of achieving that modest level of outperformance more than double: one in thirteen. Change the spread to three percentage points, a margin that has proven difficult for funds to achieve over the long run, and only one fund was a victor, with 43 funds among the vanquished. The odds of a given mutual fund's providing a three-point margin above the Index were just 1 in 200. These odds give you some idea of the Herculean challenge represented by the search to select the fund's big winners of tomorrow. (I explore this challenge more fully in Chapter 9.)

Another important lesson emerges here: The principal reason for the mutual fund shortfall is the heavy burden of fund expenses. The fund returns, relative to those of the Wilshire 5000 Equity Index shown above, are calculated in the basic manner - that is, after the deduction of all mutual fund operating expenses, which are explicit (they averaged about 1.4 percent per year during this period), and portfolio transaction costs, which are implicit (during the period, they appear to have averaged at least 0.5 percent per year for these growth and value funds). The Index was cost-free, incurring neither operating expenses nor transaction costs. If we adjust each fund's return for its approximate costs, we see a far different pattern of returns. Looking at fund returns on a gross (rather than a net) basis shifts the odds in a way that makes the industry profile look considerably better. (Fund shareholders, of course, earned only the net return.)

An examination of fund returns on a precost basis, presented in Figure 5.6, confirms the fundamental theory of indexing. Managers as a group must, be definition, provide gross returns equal to the market, just as a representative index does; therefore, the net returns earned by managers as a group will provide below-average returns once their investment costs are deducted. That result is not astonishing, nor even counterintuitive. Indeed, over the past 15 years, the 16.0 percent return on the Wilshire 5000 Index exceeded the 14.1 percent net return on the average growth and value fund by 1.9 percentage points, precisely what we might have expected based on total estimated fund costs of 1.9 percent.

In Figure 5.6, we present the returns of mutual funds before expenses. We simply redraw the previous chart and adjust for each fund's expense ratio and estimated transaction costs. In Figure 5.6, the fund distribution shifts to the right. Now, nearly half of the funds (93 of 200) outpace the Index, and 49 do so by one percentage point or more, compared to 107 that fail to match the Index. Still not great odds, but much improved over the after-cost pattern.

When we compare these gross returns with what would be a normal distribution of results - say, based on the random results of a coin-flipping contest - something interesting, or even astonishing, happens. When we fit the dotted line in the chart against the funds' gross returns, the result is almost perfect chance - similar to flipping a coin, albeit with the funds demonstrating a somewhat greater likelihood of falling in the middle of the distributions. Yes, one participant in 100 may flip heads ten times in a row, but 50 participants will flip five heads and five tails. The skill of portfolio managers, then, would appear to be largely a matter of luck, a game of chance. For, as Figure 5.6 shows, relative gross returns of mutual funds have followed a random pattern. For managers in the aggregate, the heavy handicap of costs is simply too heavy to overcome.