PRODUCT DIFFERENTIATION, SEARCH COSTS, AND COMPETITION IN THE MUTUAL FUND INDUSTRY: A CASE STUDY OF S&P 500 INDEX FUNDS by Ali Hortaçsu and Chad Syverson
I. Introduction
An investor seeking to hold assets in a mutual fund is a consumer with many choices: in
2001, there were 8307 U.S. mutual funds in operation. If one counts different share classes for a
common portfolio as separate options available to an investor, the implied total number of funds
to choose from exceeds 13,000. Note in comparison that there were a total of 7600 companies
listed that year on the NYSE, AMEX, and Nasdaq combined. A mutual fund investor’s choice
set has also been growing robustly over time: while there were 834 mutual funds in operation in
1980, this nearly quadrupled to 3100 by 1990, and almost tripled again by 2001.
An additional, less documented feature of the mutual fund marketplace is the enormous
dispersion in the fees investors pay to hold assets in funds, a dispersion that persists despite the
competition among large number of industry firms. These fee differences are not simply a result
of variation across fund sectors; price dispersion within (even narrowly defined) sectors is large.
Table 1 summarizes this within-sector dispersion. The table shows fund fee dispersion
moments—the coefficient of variation, the interquartile price ratio, and the ratio of the 90th to the
10th percentile price—for each of 22 fund objective sectors in 2000.
As is evident in the table, the 75th-percentile-price fund in a sector-year cell typically has
investor costs about twice those of the 25th percentile fund. The 90th-10th percentile price ratios
indicate between three- and seven-fold fee differences within sector-years. The extrema of the
distribution (not shown) can exhibit vast dispersion; the minimum-price aggressive growth fund,
for example, imposed annualized fees of only 14 basis points (i.e., 0.14% of the value of an
investor’s assets in the fund), whereas the highest-price fund charged a whopping 1670 basis
points. While some of these sectors are fairly broad and may include funds with very different
portfolios, dispersion remains even within what are plausibly quite specialized sectors. The
table’s second panel shows the same dispersion measures for four randomly selected specialized
sectors (these are based on Strategic Insight’s classification system, which is considerably finer
than the above taxonomy). There is a modest drop in price dispersion compared to the broader
sectors, but considerable differences still remain.
Of course, fund portfolios can vary considerably even within narrow asset classes. Fund
price dispersion might then reflect within-sector differences in demand or cost structures across
portfolios. On the demand side, certain portfolios will outperform their sector cohorts; higher
prices may just reflect investors’ willingness to pay for better performance. As for costs, fund
managers choose different securities with which to comprise their portfolio, and some of these
securities may be more expensive to analyze or trade than others. Fund prices may reflect this
fact.
Portfolio differentiation, too, may explain in part the other salient fact discussed above: the
large number of industry funds. Investors may differ in their ideal portfolios and their current
asset compositions. Perhaps thousands of funds (and the several hundred new funds each year)
are necessary to meet the demand for the many risk-return profiles sought by investors.
However, a look at the retail (i.e., non-institutional) S&P 500 index fund sector strongly
suggests that the composition and financial performance of funds’ portfolios are not the only
factors explaining fund proliferation and fee dispersion. All funds in this $164 billion (in 2000)