Common Sense on Mutual Funds by John C. Bogle, pages 101-102
A single ready-made balanced index fund - holding 65 percent stocks and 35 percent bonds, as shown in my earlier example - can meet the needs of many investors. A pair of stock and bond index funds with a tailor-made balance - a higher or lower ratio of stocks - can meet the needs of many more. But what is the optimal number of funds for investors who elect to use actively managed funds? I truly believe that it is generally unnecessary to go much beyond four or five equity funds. Too large a number can easily result in overdiversification. The net result: a portfolio whose performance inevitably comes to resemble that of an index fund. However, because of the higher costs of the non-index-fund portfolio, as well as its broadened diversification, its return will almost inevitably fall short. What is more, even though it may be overdiversified, such a portfolio (for example, one with two large-cap blend funds and two small-cap growth funds) may exhibit much more short-term variation around the market return. Therefore, according to the common definition of risk, it will be riskier than the index. To what avail?
A recent study by Morningstar Mutual Funds - to its credit, one of the few publications that systematically tackles issues like this one - concluded essentially that owning more than four randomly chosen equity funds didn't reduce risk appreciably. Around that number, risk remains fairly constant, all the way out to 30 funds (an unbelievable number!), at which point Morningstar apparently stopped counting. Figure 4.6 shows the extent to which the standard deviation of the various fund portfolios declined as more funds were added.
Morningstar noted that owning only a single large blend fund could provide a lower risk than any of the multiple-fund portfolios. I've added such a fund to Figure 4.6. But Morningstar did not note, though it might have, that a single all-market index fund provided as low a risk as did the 30-fund portfolio. I've added that too. Morningstar also failed to mention, though perhaps it should have, that the assumed initial investment of some $50,000, which would have grown to final values ranging from $85,000 to $116,000 in various fund combinations, would have grown to $113,000 in a single all-market index fund - right at the top of the range. The alleged virtues of multifund diversification and risk control hardly appear compelling.
I would add that I am not persuaded that international funds are a necessary component of an investor's portfolio. Foreign funds may reduce a portfolio's volatility, but their economic and currency risks may reduce returns by a still larger amount. The idea that a theoretically optimal portfolio must hold each geographical component as its market weight simply pushes me further than I would dream of being pushed. (I explore the pros and cons of global investing in Chapter 8.) My best judgment is that international holdings should comprise 20 percent of equities at a maximum, and that a zero weight is fully acceptable in most portfolios.
What is the point of having as many as 20 diversified funds in a portfolio (i.e., 5 percent of assets in each fund), and thus - given the inevitable overlap of their holdings - owning as many as 1,000 individual common stocks? I'm not at all sure. Perhaps a simple five-fund portfolio like the one shown in Table 4.4 would suit the needs of investors seeking active equity management.