Division of Investment Management:
Before looking at the expense ratio numbers, it is useful to identify in greater detail the costs that are included in a fund's expense ratio and the costs that are excluded.
A fund's expense ratio is its total expenses divided by average net assets. Form N-1A, the mutual fund registration form, divides total expenses into three categories: management fees, rule 12b-1 fees, and other expenses. Management fees include investment advisory fees and administrative or other fees paid to the investment adviser or its affiliates for services.60 Rule 12b-1 fees include all distribution or other expenses incurred under a plan adopted pursuant to rule 12b-1.61 Other expenses include all expenses not included in the first two categories that are deducted from fund assets or charged to all shareholder accounts. Typical "other expenses" include payments to transfer agents, securities custodians, providers of shareholder accounting services, attorneys, auditors, and fund independent directors. A mutual fund's expense ratio does not include the sales load (if any) or the cost that the fund incurs when it buys or sells portfolio securities, such as brokerage commissions. As described in the following section, fund marketing and distribution expenses are increasingly paid out of 12b-1 fees rather than out of sales loads -- a change that has had a large impact on expense ratios.
2. The Changing Role of Distribution Expenses
The past two decades have seen significant changes in the way that investors pay for the marketing and distribution of fund shares. Any analysis of mutual fund expenses must take into account the effect of these changes.62
Prior to 1980, most mutual funds were load funds, so-named because they were marketed by a sales force of brokers who received a commission (load) when shares were sold.63 The remaining funds (no-load funds or directly marketed funds) were sold by investment advisory firms directly to the public without a sales load. The more limited sales expenses of no-load funds (primarily advertising) were paid by the funds' investment advisers or underwriters, out of their own profits. In other words, prior to 1980, irrespective of whether a fund was a load or a no-load fund, distribution expenses were not included in the fund's expense ratio.
In 1980, after much debate, the Commission adopted rule 12b-1, which permits funds to pay for marketing and distribution expenses directly out of their assets.64 Since 1980, marketing and distribution expenses paid under rule 12b-1 have been included in a fund's expense ratio in the same manner as any other fund expense. Sales loads, on the other hand, continue to be excluded from a fund's expense ratio because loads are paid directly by investors and not by the fund.
Although initially few funds adopted 12b-1 plans, the number of funds with plans increased during the mid-1980s 65 as sponsors of load-funds developed a new pricing arrangement in which the combination of a 12b-1 fee and a contingent deferred sales load (CDSL) replaced the traditional front-end load.66 A CDSL is a sales load that is paid, if at all, at the time of redemption. A CDSL is "contingent" because the sales load is paid only if the shares are redeemed before a specified period of time (often 5-8 years). These CDSL funds are sold by the same brokers who sell traditional load funds, but the investor does not pay a sales load at the time that shares are purchased. Instead, the investor pays an annual 12b-1 fee or a contingent deferred sales load if shares are redeemed within a specified period of time.67 The 12b-1 payments made by CDSL funds are included in their expense ratios.
As CDSL funds became more popular, the NASD, with the approval of the Commission, determined that 12b-1 fees should be governed by the rules that apply to sales loads.68 After careful consideration, the NASD determined that funds should pay no more than 100 basis points in 12b-1 fees, 75 basis points of which could be for distribution expenses and 25 basis points for service fees annually.69 In addition, the NASD determined that a fund with no sales load and a 12b-1 fee of 25 basis points or less could identify itself as a no-load fund.
In view of the changes described above, some observers of the fund industry, including the industry's largest trade association, argue that any overall evaluation of the fees and expenses borne by fund shareholders should consider trends in total shareholder cost -- a measure that includes the cost of services paid for separately by the shareholder (most notably, distribution costs paid via sales loads) as well as the costs included in a fund's expense ratio. Although we believe that the total shareholder cost approach has considerable merit, we focus primarily on expense ratios in this study for two reasons. First, our goal is to analyze trends in fees and expenses that are incurred at the fund level and paid directly out of fund assets. Second, two data items that play a key role in total shareholder cost analysis-- actual sales loads paid by fund investors, and the actual length of time that investors hold their shares - are not publicly available.70
3. Data Sources/Explanation of Data Items
Expense ratio and other data were collected for all stock and bond funds in our database at the end of 1979, 1992, 1995, 1996, 1997, 1998, and 1999.71 Data for 1979 were taken from Weisenberger's Investment Company Services, 1980; data for 1992 and 1995 through 1999 were taken from Morningstar Mutual Funds OnDisc and Morningstar Principia Pro.72 Money market funds, another major segment of the mutual fund industry, were excluded from this study because of their different cost structure. Therefore, in this report, the terms "mutual fund" and "fund" include all mutual funds that are not money market funds. Also excluded from this study are the underlying mutual funds of insurance company separate accounts, closed-end investment companies, and face amount certificate companies.
A key issue is whether to evaluate the expense data at the level of the fund or at the level of the class. As previously indicated73 some funds issue only one share class (single-class funds or stand-alone funds); other funds issue two or more classes (multi-class funds). The multi-class form of organization is designed to provide investors with more choices. For example, different share classes may offer varying levels of service or alternative ways to pay for the cost of distributing the fund's shares.74 Because of the differences among the classes, each class has its own fee structure and expense ratio, and shareholders investing in different classes pay different expenses for an undivided interest in the same portfolio of securities. Consequently, the data reported for a multi-class fund is not the fund's expenses and assets, but rather the expense ratio of each separate class and its related assets.
A multi-class fund actually incurs most of its operating expenses at the fund level, and then allocates these expenses among the fund classes, often based on the relative asset-size of each class. The magnitude of these expenses tends to be influenced by the asset-size of the fund and not the asset size of the various classes.75
We believe, therefore, that when an expense analysis includes the relationship between funds' expense ratios and their asset sizes, it is appropriate to evaluate the asset-size of multi-class funds at the fund level. We use this approach in Section III.D (A Model for Estimating a Fund's Expense Ratio).76 In contrast, when the expense analysis focuses on the amount of expenses paid by fund shareholders, we believe it is more appropriate to perform the analysis at the class level. Accordingly, in Section III.C (Factors That Affect Fees: Descriptive Statistics) we evaluate multiple class funds at the class level -- i.e., we consider each class to be a separate data item, with its own assets and its own expense ratio.
In most cases, our study analyzes expense data for all funds or classes in existence at the end of the year.77 In three cases, because the relevant information had to be collected by hand, we limited the analysis to a sample of large classes.
1. Mutual Fund Growth
The mutual fund industry grew at an extraordinary rate during the 20 years covered by our study ("study period"). The number of stock and bond classes in the study went from 517 in 1979 to 8,901 in 1999 -- an increase of 1,622% (Table 1). Assets under management soared from $51.7 billion in 1979 to $4,456.6 billion in 1999 -- an increase of 8,520%. In terms of both number of classes and total assets, the greatest portion of the growth took place between 1992 and 1999.
Mutual Fund Growth
2. Expense Ratio Trends: All Classes
During the study period, the expense ratio of the average class ("equally weighted average") rose from 1.14% in 1979 to 1.36% in 1999 (Table 2). However, because investment dollars are spread unevenly among classes -- the largest 100 classes account for 42% of all assets and the largest 1,000 classes account for 82% of all assets -- an equally weighted average may not be the best indicator of what the typical investor is being charged. The computation of an equally weighted average gives the same importance to a small class (net assets $100,000) as it does to the largest class (net assets $92 billion).80
Expense Ratio Trends: All Classes
We believe that evaluations of fund fees should generally give more weight to classes with more assets (and more shareholders). The typical fund investor is likely to own one of the larger classes, and to be charged an expense ratio at a large class's rate.
Weighting expense ratios by class size, we find that the expense ratio of the average class rose from 0.73% in 1979 to 0.99% in 1995, fell in 1996, 1997 and 1998 to 0.91%, and then rose to 0.94% in 1999.81 Although we find that the weighted expense ratio has increased since 1979, it is important to understand why this has occurred. In the sections that follow, we discuss changes in the fund industry that might explain this increase.
3. Expense Ratio Trends by Distribution Category
As previously described, a series of changes in mutual fund distribution patterns has blurred the lines between formerly distinct marketing categories -- load vs. no-load. Today, the no-load category includes directly distributed classes with and without 12b-1 fees, as well as certain classes of sales force distributed funds in which marketing expenses are reduced or eliminated because the class is sold only to selected groups such as institutional investors or retirement plans.82 The load category now includes classes with 12b-1 fees higher than 25 basis points, classes with 12b-1 fees and CDSLs, and classes with traditional front-end loads. Although the load category consists mostly of classes distributed by commissioned sales people or financial advisers, it includes some directly distributed funds.
In recognition of these changes, we divide classes into two categories for the purpose of analyzing trends in distribution expenses.
Tables 3 and 4 show how the number and total assets of load and no-load classes have changed over time. The trend in the study period is a gradual decline in the proportion of load classes and a faster decline in their proportion of assets. In 1999, for the first time, load classes had fewer assets, 49%, than no-load classes.
Number of Classes by Distribution Category
Class Assets by Distribution Category ($ Millions)
Table 5 shows the trend in average expense ratio by distribution category over the study period. (Expense ratios are weighted by asset size in all cases.) The expense ratio of the average no-load class rose from 75 basis points in 1979 to 80 basis points in 1992, before declining to 76 basis points in 1995, 75 basis points in 1996, 72 basis points in 1997, 68 basis points in 1998 and then increasing to 72 basis points in 1999.
In 1979 -- prior to the onset of 12b-1 fees -- the average load class had a lower expense ratio (72 basis points) than the average no-load class (75 basis points). From 1979 to 1992, load class expense ratios rose 24 basis points, on average, primarily because of the inclusion of 12b-1 fees in the expense ratio. Load class expense ratios increased another 21 basis points by 1995 (to 1.17%) before falling to 1.14% in 1997, 1.12% in 1998, and increasing to 1.17% in 1999.
Weighted Average Expense Ratios by Distribution Category
4. Total Ownership Costs
The results summarized in Table 5 do not take into account the decline in front-end sales loads that accompanied the increase in 12b-1 fees. The median front-end sales load (before quantity discounts) fell from 8.5% in 1979 to 4.75% in 1999.84 Some industry participants argue that evaluations of mutual fund expense trends should take into account all costs that a shareholder would expect to incur in purchasing and holding class shares ("total ownership costs"). Total ownership costs include fund operating expenses, 12b-1 fees, and sales loads.85
As part of this study, we performed a simplified analysis of total shareholder costs. The results are shown in Table 6. A key issue for any study that employs a total ownership cost approach is how to treat the sales load paid to purchase fund share classes. The analysis requires two data items that are not publicly available: the actual loads paid by investors (dollar amount or percentage of amount invested) 86 and actual shareholder holding periods.87
Because we do not have access to data that reflect actual sales loads paid or actual holding periods of fund investments, we make certain simplifying assumptions, which make the analysis less precise. We assume that shareholders hold their shares for either 5 or 10 years.88 We also assume that all investors pay the maximum front-end sales load. Using these assumptions, we then amortize the maximum sales load by dividing the sales load by the holding period. Finally, the amortized sales load is added to the expense ratio to arrive at the total asset weighted shareholder cost.
Table 6 indicates that the magnitude of total shareholder costs depends heavily on the amortization period chosen. Amortizing the average maximum sales load over a 5-year holding period shows that total shareholder costs for load classes have declined 18% between 1979 and 1999 -- from 2.28% to 1.88%. If the longer holding period of 10 years is picked, however, total shareholder costs remained basically unchanged between 1979 and 1999.
Total Ownership Costs for Load Classes
5. Expense Ratio Trends by Type of Investment
At the beginning of the study period, the mutual fund industry generally invested in U.S. securities and did not offer specialized funds. During the 1980s and 1990s many fund sponsors broadened their product lines in an effort to attract new assets and retain assets already under management.89 This strategy led to the introduction of two new major fund categories: international funds and specialty funds.90
In 1979, bond fund classes accounted for 38% of classes and 33% of assets, while equity fund classes accounted for 62% of classes and 67% of assets (see Tables 7 and 8). By 1992, bond classes had overtaken stock classes to become the largest fund category and international classes (10% of classes; 6% of assets) and specialty classes (6% of classes; 3% of assets) had become a significant part of the fund landscape.
Number of Classes
Seven years later, a bull market in equities enabled stock fund classes to become the largest category in terms of assets although bond fund classes still accounted for the largest number of classes. In 1999, stock fund classes accounted for 61% of assets compared to 21% for bond fund classes. Bond fund classes accounted for 44% of classes in 1999 and stock fund classes accounted for 34%. International fund classes grew steadily during the study period until they accounted for 16% of classes and 13% of assets, while the number of specialty fund classes stayed level at 5%, but their assets grew to 5% of total assets.
It is generally believed that equity funds are more expensive to manage than bond funds and that international and specialty funds are more expensive to manage than equity funds.91 Equity funds are thought to be more expensive to manage because of the increased research costs associated with picking stocks. Similarly, international funds are thought to incur additional costs over and above domestic equity funds because of the increased difficulty of researching international companies. Some of the increased cost results from the need to review and understand foreign accounting statements and to obtain company information not required to be disclosed under foreign securities laws. Custody costs generally are higher, as well.
The results shown in Table 9 are consistent with the opinions described above. Table 9 indicates that bond fund classes have lower expense ratios than equity fund classes, and that international and specialty fund classes have higher expense ratios than bond and equity fund classes. This fact, coupled with the increase in assets of equity, international, and specialty fund classes, helps explain some of the increase in mutual fund expenses.
Weighted Average Expense Ratio
By Type of Fund
6. Expense Ratio Trends by Class Age
Another common explanation for rising expense ratios is that large numbers of new funds have pushed up the averages. Commentators say that new funds often have higher expense ratios because they have not yet reached the critical size needed to pass on economies to their shareholders.92
Table 10 tends to confirm the notion that new fund classes have higher expense ratios. The average expense ratio (weighted by asset size) of classes that have been in existence 5 years or less is 1.23%, compared to 1.10% for classes in existence between 6-10 years, and 0.80% for classes in existence for more than 10 years.93
Years in Existence
7. Expense Ratio Trends by Class Size
The previous table indicates that expense ratios seem to be inversely correlated with age. That is, as classes get older they have lower expense ratios. Some industry commentators have suggested that the recent creation of newer smaller classes tends to increase the weighted expense ratio. Table 11 attempts to determine the relationship between class asset-size and expense ratios.
Table 11 divides all classes in 1999 into five groupings by asset size. As can be seen in the table, classes in the largest size category -- assets greater than $1 billion -- hold more than two-thirds of all fund assets. The data show that there is, in fact, an inverse relationship between size category and expense -- as the size category increases, expense ratios fall.
In Section C we found that the level of a class's expense ratio seems to depend on the following factors: asset size, age, investment category, and method of distribution. Because these factors appear to be important in explaining the magnitude of expense ratios at the class level, we sought to obtain more precise information about their impact.
To achieve this end, we built an econometric model of the relationship between the expense ratios of mutual fund classes and the factors described in Section C, as well as a few others. Our model hypothesizes that expense ratios of mutual fund classes can be explained by the following 11 factors: (1) fund asset size; (2) fund family asset size; (3) number of funds in its fund family; (4) portfolio turnover; (5) number of portfolio holdings; (6) fund age; (7) investment category; (8) method by which it finances distribution; (9) whether or not it is an index fund; (10); whether or not it is an institutional fund or class; and (11) whether it is part of a multi-class fund. 94 We used the model to analyze expense data for the 8,901 classes in our database in 1999.
2. Results of Econometric Model of Expense Ratios
We used our econometric model (see Appendix One, Regression Table) to analyze the expense ratio and operating expense ratio of classes in our database in 1999.95 As indicated previously, a fund's expense ratio is defined as its total expenses, including rule 12b-1 fees, divided by its average net assets. A fund's operating expense ratio is defined as its total expenses minus rule 12b-1 fees divided by its average net assets. In our analysis of total expenses (column 1) we observe that the maximum 12b-1 factor tends to explain the variance in total expenses due to actual 12b-1 fees and that the other factors explain only that part of the variance in total expenses that is due to differences in operating expenses. So the coefficients for the independent variables (except for the maximum 12b-1 fee) represent the influence of these variables on the operating expense ratio, not the total expense ratio.
We found that the following factors are important in explaining variations among fund operating expense ratios.96 Or, to put it another way, we found statistically significant relationships 97 between the operating expense ratios of funds 98 and the following factors. 99
The remainder of this section discusses the above results in more detail, using examples based on the data for 1999.
a. Fund Size
Other things held equal, a fund with assets of $10 million had an operating expense ratio that was 22 basis points lower than a similar fund with assets of $1 million. (Table 12). A fund with assets of $1 billion had an operating expense ratio that was 66 basis points lower than a similar fund with assets of $1 million.100
Relationship Between Fund Size and Operating Expense Ratio
b. Fund Family Asset-Size
In 1999, other things held equal, a fund's operating expense ratio fell 68 basis points if the total assets of its fund family rose from $1 million to $10 million (Table 13). A fund's operating expense ratio fell 75 basis points if fund family assets rose from $1 million to $10 billion.101
Relationship Between Fund Family Asset Size and Operating Expense Ratio
Increase in Fund Family
Change in Operating Expense Ratio
c. Investment Category
A very important factor in predicting a fund's operating expense ratio is its investment category. In 1999, bond funds were the lowest cost investment category. Other things held equal, in 1999 an equity fund had an operating expense ratio that was 44 basis points higher than a bond fund; a hybrid fund had an operating expense ratio that was 22 basis points higher than a bond fund; and a specialty fund had an expense ratio that was 62 basis points higher than a bond fund. These results are applicable to funds that invest primarily in securities issued by United States issuers. With respect to funds that invest primarily in securities issued by non-United States issuers, an international equity fund had an expense ratio that was 82 basis points higher than a domestic bond fund and an international bond fund had an expense ratio that was 31 basis points higher than a domestic bond fund.
d. Index, Institutional, and Multi-Class Funds
In 1999, other things held equal, the operating expense ratio of an index fund was 45 basis points lower than an equivalent fund that was not an index fund. The operating expense ratio of an institutional fund or class was 22 basis points lower than an equivalent fund or class that was not limited to institutional investors. Finally, a multi-class fund had an operating expense ratio that was 14 basis points higher than an equivalent single-class fund.
e. Number of Funds in a Fund Family
In 1999, other things held equal, a fund with ten funds in its family had an operating expense ratio that was 14 basis points lower than a fund with only 1 fund in its fund family (Table 14). A fund with 100 funds in its family had an operating expense ratio that was 28 basis points lower than a fund with 1 fund in its fund family.
Relationship Between Fund Family Number and Operating Expense Ratio
f. Portfolio Turnover Rate
Portfolio turnover rate measures the average length of time that a security remains in a fund's portfolio. A fund that has a 100% portfolio turnover rate holds its securities for one year, on average. A fund with a portfolio turnover rate of 200% turns over its portfolio twice a year. In 1999, other things held equal, a fund with a portfolio turnover rate of 100% had an operating expense ratio that was 30 basis points higher than a similar fund with a portfolio turnover ratio of 1%. A fund with a portfolio turnover ratio of 200% had an expense ratio that was 4 basis points higher than a similar fund with a portfolio turnover ratio of 100%.
g. Number of Portfolio Holdings
Other things held equal, a fund that held 100 securities in its investment portfolio had an operating expense ratio that was 8 basis points higher than a similar fund that held 10 securities in its portfolio. A fund with 1,000 portfolio securities had an operating expense ratio that was 16 basis points higher than a fund with 10 portfolio securities.
h. Fund Age
Other things held equal, the operating expense ratio of a 10 year-old fund was 11 basis points higher than that of a 1 year-old fund in 1999; and the operating expense ratio of a 20-year-old fund was 4 basis points higher than that of a 10-year-old fund. Although the results indicate a positive relationship between age and expenses, the results appear to be driven at least in part by four older funds that have higher expenses than their peers. When the four funds are removed from the database, the positive relationship between a fund's age and operating expense ratio became considerably weaker.
i. Payment for Distribution Expenses: 12b-1 fee
The coefficient for the variable representing the maximum allowable 12b-1 fee is 0.93. This coefficient is statistically different from both 0 and 1.0. This indicates that, everything else equal, funds with 12b-1 fees had total expenses that were higher than those of other funds, but by an amount that was slightly less than the maximum 12b-1 fee.102 This may have occurred because funds do not always charge a 12b-1 fee, even if such a fee is approved, or charge less than the maximum fee. In addition, some funds with 12b-1 fees may use these fees to pay for expenses that other funds may consider part of operating expenses. In these latter cases, the imposition of a 12b-1 fee might reduce operating expenses slightly.
j. Payment for Distribution Expenses: Sales Load
In 1999, other things held equal, the operating expense ratio of a fund with a front-end sales load was 6 basis points lower than the operating expense ratio of an equivalent fund.
The results from our model confirm that the factors identified in Section C are important in explaining a fund's operating expense ratio. We next turn our attention to mutual fund management expenses and focus on the relationship between a fund's portfolio asset size and its management expense ratio.
Evidence developed above indicates that as mutual funds' assets grow larger, their operating expense ratios decline. In order to determine whether a similar pattern exists with respect to mutual fund management expenses,103 we hand-collected management expense data for the largest 1,000 classes in existence in 1999 and used a similar econometric model to analyze the data.104 The model is the same as previously described with one exception. This time, the dependent variable is the fund's management expense ratio. We are interested in a fund's management expense ratio because it includes the cost of providing the fund with portfolio management services -- e.g., conducting research, maintaining a trading desk, managing the investment portfolio in accordance with stated investment objectives and policies. Most observers believe that portfolio management is the fund cost with the greatest economies.105 Although we cannot analyze directly the cost of providing portfolio management services to a mutual fund in order to determine whether economies exist (because the data are unavailable), we can do the next best thing. We can analyze portfolio management costs indirectly by using the management fee charged to a fund by its adviser as a proxy for the adviser's cost of providing portfolio management services. Unfortunately, the proxy is far from perfect because management fees often pay for other services as well.106
One piece of evidence for the existence of economies in portfolio management is that many mutual fund management contracts contain fee breakpoints. Fee breakpoints are an arrangement under which the management fee rate on incremental assets is reduced as total fund assets surpass specified dollar levels.107
Breakpoints were first introduced during the 1960s after shareholders of investment companies sued over the fairness of advisers' fees.108 Although the management fee was not found to be "legally excessive" in any of the cases that came to trial, many other cases were settled before trial and the adoption of management fee breakpoints was often a condition of those settlements. 109
In our analysis we are interested in seeing whether fund management expense ratios decline as fund assets increase and breakpoints in management contracts are triggered.
2. Results of Regression Model of Management Expense Ratios
Our analysis produced interesting results. The management expense ratio of the 1,000 largest funds in 1999 did not show a statistically significant decline as fund assets grow, but rather, showed a statistically significant decline as fund family assets grew (see Appendix One). Other things held equal, a fund's management expense ratio fell 11 basis points in 1999 as fund family assets rose from $1 million to $10 million. A fund's management expense ratio fell 42 basis points as fund family assets rose from $1 million to $10 billion.110
Relationship Between Fund Family Asset Size and Management Expense Ratio
These results seem to indicate that, among large funds, economies in management expenses are present at the fund family level rather than at the fund level.111
In order to obtain additional information about the extent to which economies are present in management fees, we examined the management contracts of the 100 largest mutual funds in 1997, 1998, and 1999 for evidence of management fee breakpoints.112 Because management contracts are generally based on the total assets in a fund portfolio, we added together all the classes of multi-class funds to select the 100 largest funds.
An analysis of the management contracts of these funds produced some interesting results. Our analysis shows that not all management contracts incorporate fee breakpoints as fund assets increase. Instead, we observe contracts with five types of arrangements: 1) fee breakpoints based on fund assets (fund breakpoints); 2) fee breakpoints based on portfolio assets plus a performance fee (fund breakpoints-plus); 3) fee breakpoints based on fund family assets (fund family breakpoints); 4) a single, all-inclusive fee (single fee); and 5) at-cost arrangements. In addition, we observe that for funds with fund breakpoint or fund breakpoint-plus contracts, a substantial proportion of their assets are not subject to any further breakpoint reductions (Table 16). The remainder of this section discusses the different types of management contracts.
Fund breakpoint contracts have management fees that decline at selected asset intervals based on the asset size of the fund. Forty-seven funds in our analysis, with assets of $855.2 billion, have fund breakpoint contracts. The median number of breakpoints for the 47 funds is six. For these funds, the median asset-size level at which the first breakpoint takes effect is $500 million and the median asset-size at which the last breakpoint takes effect is $10 billion. The median management fee at the first breakpoint is 65 basis points and the median management fee at the last breakpoint is 41 basis points. Thirty-four funds have assets that exceed their last breakpoint. For these 34 funds, the combined assets that are not subject to any further breakpoints total $318 billion.
Management Fee Breakpoints
Fund family breakpoint contracts include breakpoints based on the asset size at the fund family level together with a single rate fee or a performance fee at the fund level. Twenty-one funds in our analysis, with assets of $506.3 billion, have a fund family fee. The median number of breakpoints at the fund family level is 37, with the first breakpoint at $3 billion in fund family assets and the last breakpoint at $1.2 trillion of fund family assets. The median fee rate for the first breakpoint is at 52 basis points and the median fee rate for the last breakpoint is 22 basis points. No funds have assets that exceed the last breakpoint.
Mutual funds that have fund breakpoints-plus contracts have an asset-based fee with breakpoints at the fund level and a separate fee that varies with the fund's investment performance. Eight funds in our analysis, with assets of $113.9 billion, have fund breakpoint-plus contracts. The median number of breakpoints is 4, with the first breakpoint at a fund asset-size of $150 million and the last breakpoint at a fund asset-size of $10 billion. For the median fund in this category, the first breakpoint is at fee rate of 27.5 basis points and the last breakpoint is at a fee rate of 11.3 basis points. Five funds have a combined $41.1 billion of assets that exceed the asset level of the last breakpoint.
Single fee contracts do not employ breakpoints. Nineteen funds in our analysis, with assets of $376 billion, have single fee management contracts. The median fee rate for single fee management contracts is 65 basis points, with a high of 100 basis points and a low of 24 basis points.
Five funds in our analysis have "at-cost" arrangements. For these funds, the management fee is not a function of asset size of the fund, asset size of the fund family, or the fund's investment performance. These funds have combined assets of $204.7 billion.
Americans entrust a significant portion of their retirement savings to mutual funds. As of December 31, 1999, mutual funds held $2.4 trillion (19%) of the $12.7 trillion in US retirement assets.113 Retirement assets represent more than one-third of total fund assets.
Retirement assets invested in mutual funds come primarily from 401(k) plans and other defined contribution arrangements, individual retirement accounts (IRAs), and variable annuities outside of retirement accounts. Over 40 percent of defined contribution plan and IRA assets are invested in mutual funds.
Because concern has been expressed about the level of 401(k) plan expenses, we sought to gain some insight into the level of expenses charged to 401(k) plans that invest their assets in mutual funds.114 Toward that end, we selected a sample of 50 funds with the most 401(k) assets (retirement-oriented funds) and compared their expenses to those of all funds. The retirement-oriented funds manage $340 billion in 401(k) assets and $993 billion of assets from all sources. For almost all funds in the sample, 401(k) assets represent a large portion of total assets. The average retirement-oriented fund derives 34% of assets from 401(k) plans, with the high being 95%, and the low 11%. Twelve retirement-oriented funds derive more than half of their assets from 401(k) plans.
Retirement-oriented funds do not have higher expenses than the average fund. In fact, the equally-weighted average expense ratio for retirement-oriented funds (96 basis points or 0.96%) is 28% below the average expense ratio for all mutual funds (1.35%). The asset-weighted average expense ratio for retirement-oriented funds is 24% below the average expense ratio for all funds (69 basis points compared to 91 basis points). It is likely that the primary reason why retirement-oriented funds have lower expense ratios is their size. The average retirement-oriented fund has $19.9 billion in assets, compared to $423 million for all funds.
Our goals in conducting this study were to provide summary data about the current level of mutual fund fees, describe how fee levels have changed over time, and identify some of the major factors that influence the current amount of fees charged. Some of the more significant findings are summarized below.
The current regulatory framework for mutual fund fees relies on a combination of disclosure, investor education, and procedural safeguards. To further improve the effectiveness of the current framework, we have the following recommendations.
1. Dollar Amount of Fund Fees
In its June 2000 report on mutual fund fees, the General Accounting Office recommended that the Commission require mutual funds and/or broker-dealers to send fund shareholders account statements that include the dollar amount of the fund's fees that each investor has indirectly paid. The GAO report surmises that adding personalized expense information to fund account statements may prompt fund shareholders to pay more attention to fees and to compare their fund's fees and services with those of similar funds, thus encouraging more fee-based competition among funds. The report acknowledges that requiring funds and/or broker-dealers to provide this information would impose additional costs on the industry because funds would have to change their account management systems to collect and calculate information that is not currently maintained. The GAO also recommends that the Commission consider alternatives that may provide similar information at lower cost, and identifies two such alternatives.
The GAO report identifies two alternatives that may merit further study. One alternative would be to multiply the fund's per share asset value by the fund's expense ratio, multiply the result by the average number of shares an investor owned during the period, and show the result in the investor's account statement. This alternative would provide each shareholder with an approximation of the dollar amount of fund expenses that he or she indirectly paid. A second alternative would be to provide information about the dollar amount of fees that were paid during the period for preset investment amounts, such as $1,000. Investors could use the results to estimate the amount they paid on their own accounts. The report notes that the Commission would need to weigh the costs of each approach against the benefits of the additional information to investors.
As the Commission considers how to best disclose to investors the fees and expenses that they incur with investment in a fund, including whether it would be appropriate for fund account statements to include personalized information about expenses or other fund-related data, it will need to consider the advantages and disadvantages of each alternative. For example, providing fund shareholders with personalized information, expressed as a dollar amount, about the fees and expenses that they paid indirectly during the year might increase shareholder awareness of fund fees and expenses. On the other hand, fees and expenses would need to be presented on a standardized basis - i.e., as a percentage of fund assets, for a defined time period, calculated in a manner that is uniform for all funds. Finally, as indicated in the GAO report, the compliance cost associated with a new personalized expense disclosure requirement, which ultimately would be borne by fund shareholders, may be considerable. Computer programs that perform shareholder accounting functions would have to be revised and other costs would be incurred. Administrative difficulties would present an additional obstacle. Shareholder accounting often is performed not by the fund, but by a broker-dealer who, in many cases, has no affiliation with the fund. Moreover, many investors hold their shares in omnibus accounts with broker-dealers. These broker-dealers do not have the information that would be needed to calculate the dollar amount of fees attributable to individual fund shareholders and would have to develop interfaces with the record owners of these accounts.
We believe that an approach that is based on the second alternative suggested by the GAO is likely to have the most favorable trade-off between costs and benefits. That alternative would provide information about the dollar amount of fees paid for preset investment amounts. Specifically, we recommend that information about the dollar amount of fees and expenses be presented in a fund's shareholder reports, so that investors can evaluate the information alongside other key information about the fund's operating results, including management's discussion of the fund's performance. In effect, shareholders would be able to evaluate the costs they pay against the services they receive. We also recommend that some or all of the information about the dollar amount of fees should be calculated in a manner that makes it easy for investors to compare the fees charged by their fund with the fees charged by other funds. Although our recommendation could be implemented in a variety of ways, we believe that the general approach embodied in our recommendation will encourage investors to incorporate information about the dollar amount of fund fees into their decision-making process.
Our approach would be to require fund shareholder reports to include a table that shows the cost in dollars associated with an investment of a standardized amount (e.g., $10,000) that earned the fund's actual return for the period and incurred the fund's actual expenses for the period. The Commission could require, in addition, that the table include the cost in dollars, based on the fund's actual expenses, of a standardized investment amount (e.g., $10,000) that earned a standardized return (e.g., 5%). This approach would provide additional information about fund fees, provide it in terms of dollar amounts, and provide it in a standardized manner that would facilitate comparison among funds. (The only variable in this calculation would be the level of expenses).115
Disclosure about fees and investor education about fees go hand-in-hand. As the primary information source for most fund investors, the mutual fund industry - funds, brokers, and other financial professionals - must play a major role in increasing investor awareness and understanding of fund fees. The fund industry should expand its efforts to educate investors about SEC-mandated disclosures and other information they can use to identify the fees that they pay, compare funds to each other and to other investment alternatives with respect to the level of fees, and consider the effect that fees will have in reducing the amount of wealth they may be accumulated as a result of an investment.116 The Commission has an important role to play, as well, and should continue its ongoing program (described in Section II) to improve the financial literacy of investors with respect to mutual funds and their costs. As the fee information described above (or other similar information required by the Commission) begins to appear in fund disclosure documents, the Commission should develop educational materials that help investors understand how they can use the new information. Also, as mutual fund fee structures become more complex, the Commission may be able to help investors make better-informed decisions. For example, although multiple share classes offer investors additional choices, investors may be confused by the various fund classes and find it difficult to determine which class represents the best value for their particular circumstances. Because the selection of the appropriate class of shares to invest in can be a complicated decision that generally depends on the unique circumstances of an investor, further investor education concerning these issues would be beneficial.
2. After-Tax Return
We recommend that the Commission adopt proposed amendments to our rules and to Form N-1A, the registration form for mutual funds, that would require disclosure of standardized mutual fund after-tax returns. Although fund expenses play a key role in determining ultimate shareholder wealth, taxes play an even larger role for many investors in mutual funds. A major accounting firm found, for example, that taxes reduced the investment performance of the median domestic stock fund by 2.6% per year.117 For comparison, we find in our fee study that the median expense ratio for all stock funds in 1999 was 1.3% per year and the weighted average expense ratio (See Section III, Table 9) was 0.90% per year. Due to the significant impact that taxes have on investors, we believe that investors would benefit greatly by receiving better disclosure concerning the effect of tax expense on returns.
1. Role of Independent Directors
We believe that the current regulatory framework would be enhanced by independent directors who more actively monitor fund fees and expenses.
In its October 1999 proposal of new rules and rule amendments, the Commission sought to strengthen the hand of independent directors in dealing with fund management and to provide fund shareholders with greater information to make their own assessment of the directors' independence. We recommend that the Commission consider these proposals as soon as practicable after the Commission staff finishes its review of comments from the public and the industry.
Of particular importance is the proposal that would, in effect, require that independent directors (directors not associated with the fund's management) comprise at least a majority of the members of fund boards. In our view, a fund board that has at least a majority of independent directors is likely to do a better job of representing the interests of fund shareholders than a board that has a lesser percentage of independent directors. An independent director majority would be able to elect officers of the fund, call meetings, solicit proxies, and take other actions without the consent of the adviser.118 The ability of a board to act without the approval of the inside directors should better enable it to exert a strong and independent influence over fund management. This is particularly true when the board considers the investment advisory fee rate, a situation in which the fund's interests conflict with those of the adviser. Although most funds already have boards with an independent majority, the proposals would ensure that shareholders of all funds that rely on certain Commission exemptive rules (virtually all funds) have the benefits of a board with an independent majority.
Fund directors also can strengthen their hand by educating themselves about issues concerning mutual fund fees and expenses.119 In particular, we recommend that fund directors focus further on the costs of providing investment management services and, in particular, on whether the funds that they oversee experience any economies of scale. In our study, we found that, for large funds, management expense ratios declined as fund family assets grew. We also found that the management expense ratios of large funds declined as individual fund assets grew, but the decline was not statistically significant. These results suggest that, in certain instances, economies of scale may be experienced primarily at the fund family level and only to a lesser extent or not at all at the fund level. Conclusions as to why economies of scale would be experienced in this way, however, cannot be drawn without knowing what the costs of supplying particular services were to the investment advisory firms.120
At the fund level, however, fund directors can obtain information about the cost of providing investment management services to the funds that they oversee. Fund directors can use this information to evaluate whether the funds that they oversee are experiencing any economies of scale and to assist them in ensuring that fund shareholders share in the benefits of any reduced costs. Whether increases in assets of a fund or fund family produce economies of scale is a factor that may influence fund directors' views on, among other things, the amount of fees that the fund should pay for advisory and other services and whether a rule 12b-1 plan for the fund is appropriate.
If the fund or fund family is experiencing economies of scale, fund directors have an obligation to ensure that fund shareholders share in the benefits of the reduced costs by, for example, requiring that the adviser's fees be lowered, breakpoints be included in the adviser's fees, or that the adviser provide additional services under the advisory contract. If the fund or fund family is not experiencing economies of scale, then the directors may seek to determine from the adviser how the adviser might operate more efficiently in order to produce economies of scale as fund assets grow. We believe that fund directors who ask pertinent questions about investment management costs can more effectively represent the interests of the shareholders they represent.
We believe that fund directors would benefit from learning about the types of information that they can review when making their decisions, including information that would enable them to determine whether their funds are experiencing any economies of scale. We believe that fund directors also would benefit from knowing about other sources of data and information that would enable them to compare the costs of investment management of the funds that they oversee with those of other funds. Fund directors who are equipped with this information can more effectively represent the interests of the fund's shareholders when setting and re-approving advisory and other fees.
Not all costs associated with investment in a mutual fund are paid for via the fund's expense ratio. The cost of effecting the fund's portfolio transactions, for example, is reflected in the amount paid when the fund buys or sells portfolio securities.121 For many funds, the amount of portfolio transaction costs incurred during a typical year is substantial.122 Clearly, fund directors should focus on portfolio transaction costs.123 As they review fund transaction costs, fund directors should pay particular attention to soft dollar practices -- arrangements under which the fund's investment manager obtains, from or through a broker dealer, products or services other than execution of securities transactions. The manager obtains these services in exchange for allocating client brokerage transactions to the broker-dealer.124
In addition to reviewing soft dollar practices, fund directors should carefully consider directed brokerage arrangements. Under a directed brokerage arrangement, the fund asks the investment adviser to direct securities transactions to a particular broker that has agreed to provide services, pay for services provided by others, or make cash rebates to the fund. Funds typically enter into directed brokerage arrangements to offset fund expenses, such as audit, legal, and custodial fees. Although directed brokerage does not involve the conflicts posed by soft dollars, it does raise issues related to how a fund's assets are being expended and other issues, including disclosure.125
2. Rule 12b-1
We recommend that the Commission consider whether it would be appropriate to review the requirements of rule 12b-1 that govern how funds adopt and continue their rule 12b-1 plans. We believe that modifications may be needed to reflect changes in the manner in which funds are marketed and distributed and the experience gained from observing how rule 12b-1 has operated since it was adopted in 1980.126 The rule essentially requires fund directors to view a fund's 12b-1 plan as a temporary measure even in situations where the fund's existing distribution arrangements would collapse if the rule 12b-1 plan were terminated. Under the rule, fund directors must adopt a 12b-1 plan for not more than one year, may terminate the plan even before the end of that year, and must consider at least annually whether the plan should be continued.127
In addition, many directors believe that when they consider whether to approve or continue a 12b-1 plan, they are required to evaluate the plan as if it were a temporary arrangement.128 The adopting release for rule 12b-1 included a list of factors that fund boards might take into account when they consider whether to approve or continue a rule 12b-1 plan.129 Many of the factors presupposed that funds would typically adopt rule 12b-1 plans for relatively short periods in order to solve a particular distribution problem or to respond to specific circumstances, such as net redemptions.130 Although the factors are suggested and not required, some industry participants indicate that the factors are given great weight by fund boards. Some argue that the recitation of the factors impedes board oversight of rule 12b-1 plans because the temptation to rely on the factors, whether they are relevant to a particular situation or not, is too great to ignore.131 Although the factors may have appropriately reflected industry conditions as they existed in the late 1970s, others argue that many have subsequently become obsolete because, today, many funds adopt a rule 12b-1 plan as a substitute for or supplement to sales charges or as an ongoing method of paying for marketing and distribution arrangements.132
The mutual fund industry utilizes a number of marketing and distribution practices that did not exist when Rule 12b-1 was adopted. For example, as described in Section III, many funds offer their shares in multiple classes -- an organizational structure that permits investors to choose whether to pay for fund distribution and marketing costs up-front (via front-end sales charge), over time from their fund investment (via 12b-1 fee), when they redeem (via deferred sales charge), or in some combination of the above.133 Rule 12b-1 plans are integral to these arrangements - they are the means by which the brokers that sell fund shares under these arrangements are paid. Some industry observers argue that fund principal underwriters and boards of directors may have good reason to view this type of 12b-1 plan as an indefinite commitment because a multi-class distribution arrangement could not continue to exist if the associated rule 12b-1 plan were terminated or not renewed.
Other funds offer their shares primarily through fund supermarkets -- programs sponsored by financial institutions through which their customers may purchase and redeem a variety of funds, with or without paying transaction fees. (Fund supermarkets are popular because they enable investors to consolidate their holdings of funds from different fund groups in a single brokerage account and to receive a consolidated statement listing all fund holdings.)134 Many funds that offer shares through fund supermarkets adopt rule 12b-1 plans to finance the payment of fees that are charged by the sponsors of fund supermarkets. Some may argue that because these 12b-1 plans are essential to the funds' participation in fund supermarket programs, these 12b-1 plans may be legitimately be viewed as indefinite commitments. In addition, because most funds pay fees to fund supermarkets for a mixture of distribution and non-distribution services, it can be difficult to determine when and how rule 12b-1 applies to these fees. Although the Division has provided additional guidance about what constitutes a distribution expense,135 questions still remain about how to determine whether a particular activity is primarily intended to result in the sale of fund shares, and therefore must be covered by a rule 12b-1 plan.
A third significant change in distribution practices is that some fund distributors are now able to finance their efforts by borrowing from banks, finance companies, or the capital markets because they can use anticipated 12b-1 revenues as collateral, or as the promised source of payment.136 If a fund adopts a 12b-1 plan, the right of its distributor to receive future 12b-1 fees from the fund is an asset of the distributor. Some distributors borrow from banks, finance companies, or other financial intermediaries, using this asset as collateral. Other distributors issue debt securities (asset-backed securities) for which the payment of principal and interest is backed by the distributors' contractual right to receive a stream of future 12b-1 fees.137 Although the independent directors of a fund have the legal right to terminate a fund's rule 12b-1 plan, the independent directors may be less likely to do so if the fund's future 12b-1 fees have been pledged to secure a bank loan or to pay principal and interest due on asset-backed securities.138
Because of these issues, the Commission should consider whether to give additional or different guidance to fund directors with respect to their review of rule 12b-1 plans, including whether the factors suggested by the 1980 adopting release139 are still valid. The Commission also should consider whether the procedural requirements of Rule 12b-1 need to be modified to reflect changes in fund distribution practices that have developed since the rule was adopted twenty years ago or may be developed in the future.
Over the past 60 years, the Commission has sought to protect the interests of fund investors with respect to fund fees and expenses through a combination of procedural safeguards to prevent conflicts of interest from resulting in excessive fees, full disclosure to make fund fees and expenses more transparent and easier to compare, and educational efforts designed to make investors more aware of the importance of fees and better able to use the fee disclosures that are available. We continue to believe that this approach is sound and is consistent with the regulatory framework established by Congress. We believe, however, that improvements can be made. The recommendations described above would provide investors with better information about fund fees, energize fund directors to take a more active role in monitoring fees, and enhance the Commission's ongoing efforts to improve investors' financial literacy with respect to mutual funds and their costs.
Sample is all classes of funds covered by Morningstar as of March 1999. Assets is Ln of fund assets. Famsize is 1/assets of fund family. Famnum is Ln of funds in the family. Turnover is Ln of class's turnover. Holdings is Ln of number of portfolio holdings. Age is Ln of fund age. Domestic equity is a indicator variable (1=domestic equity, 0=all others). Hybrid is an indicator variable (1=domestic hybrid fund, 0=all others). International bond is an indicator variable (1=international bond fund, 0=all others). International equity is an indicator variable (1=international equity fund, 0=all others). Specialty is an indicator variable (1=specialty fund, 0=all others). The omitted investment objective is domestic bond funds. Index is an indicator variable (1=index fund, 0=all others). Institutional is an indicator variable (1=institutional fund or class, 0=all others). Load is an indicator variable (1=front-end load, 0=all others). Multi-class is an indicator variable (1=multi-class, 0=single class funds). 12b-1 is the maximum 12b-1 fee authorized.
Note: In the body of our report, we analyzed expense ratio trends for two distribution categories -- load funds and no-load funds. In this Appendix, we subdivide the no-load fund category into two subcategories -- pure no-load and extended no-load -- and restate the data accordingly.
Number of Classes by Distribution Category
Class Assets by Distribution Category
Tables 1 and 2 show that 84% of the classes in the extended no load category are "pure" no-load classes (classes with no 12b-1 fee) and they account for 94% of the assets. In 1999, 547 (16%) of extended no-load classes charged a 12b-1 fee. These funds accounted for 6% of category assets. These figures represent a slight increase compared to 1995, when 337 (14%) of extended no-load classes had a 12b-1 fee and these funds accounted for 5% of category assets.
Weighted Average Expense Ratios by Distribution Category
Table 3 shows the trend in average expense ratio by distribution category over the study period. (Expense ratios are weighted by asset size in all cases.) The expense ratio of the average pure no-load class rose from 75 basis points in 1979 to 80 basis points in 1992, before declining to 74 basis points in 1995, 70 basis points in 1997, 66 basis points in 1998, before rising to 69 basis points in 1999. The inclusion in the extended no-load category of classes with 12b-1 fees of 1-25 basis points seems to have added 3 basis points to the average expense ratio in 1999.
1 This Report presents the results of an analysis of fee data for all stock mutual funds and bond mutual funds that were in our database at the end of 1979, 1992, 1995, 1996, 1997, 1998, and 1999; and for which data were available. Money market funds are excluded from the analysis because they have a different cost structure. Also excluded are the underlying mutual funds of insurance company separate accounts, closed-end investment companies, unit investment trusts, and face amount certificate companies. For an explanation of the data items used in the study, see infra Section III.B.3.
2 The Random House College Dictionary defines a fee as "a charge or payment for services," Random House College Dictionary 484 (Revised 1st Ed. 1982), and defines an expense as any "cost or charge." Id. at 465. We use the terms interchangeably in this report.
3 Retirement assets invested in mutual funds have increased from $300 billion in 1991 to almost $2.5 trillion in 1999. See Investment Company Institute, Mutual Fund Fact Book 50 (2000) (hereinafter "ICI Fact Book"). See also Karen Damato, Facing the Future of Funds, Wall St. J., Jan. 10, 2000, at R1 (discussing generally the increasing importance of the mutual fund industry during the 1990s).
4 See ICI Fact Book, supra note 3, at 69.
5 The number of funds represents the number of stock, bond and money market fund portfolios as of the end of the year. Id. at 71.
6 See Investment Company Institute, Fundamentals: Investment Company Research in Brief, Aug. 2000 at 1 (number of fund shareholders) (hereinafter "Fundamentals"); ICI Fact Book, supra note 3, at 67 (value of fund assets); Federal Reserve Board, Financial and Business Statistics, 85 Fed. Reserve Bull. A1, A15 (May 1999) (value of commercial bank assets).
7 See fundamentals, supra note 6, at 1.
8 See ICI Fact Book, supra note 3, at 50-51.
9 See, e.g., John C. Bogle, Do Mutual Funds Charge You Too Much?, Mutual Funds, Oct. 1998, at 80; Amy C. Arnott, The Rising Tide, Morningstar Mutual Funds, Oct. 11, 1996, at S1-S2.
10 ICI Fact Book, supra note 3, at 30.
11 The GAO report, Mutual Fund Fees: Additional Disclosure Could Encourage Price Competition, GAO/GGD-00-126 (General Accounting Office, June 2000) (hereinafter "GAO Report"), was delivered to the Chairman of the House Subcommittee on Finance and Hazardous Materials and the Ranking Member of the House Commerce Committee in June 2000.
12 However, Section 36(b) of the Act, 15 U.S.C. § 80a-35(b) (2000), authorizes the Commission to sue fund advisers that breach their fiduciary duty with respect to their receipt of compensation from a fund.
13 See, e.g., Dan Moreau, SEC Watches Over Mutual Fund Industry, Investor's Bus. Daily, June 15, 1999, at B1; Carole Gould, `Truth in Advertising' for Mutual Funds, N.Y. Times, Apr. 17, 1988, § 3, at 11; Jane Bryant Quinn, New Mutual Fund Table is Valuable Tool for Investors, St. Petersburg Times, May 12, 1988, at 19A; Bill Sing, Rules Offer Some Help on Shopping for Funds, L.A. Times, Apr. 30, 1988, § 4, at 3; Jan M. Rosen, Comparing Costs of Mutual Funds, N.Y. Times, Jul. 30, 1988, at 34.
14 See GAO Report, supra note 11, at 97-98.
15 These data are the type of fee information that GAO recommended that investors be given. See GAO Report, supra note 11, at 97 (second alternative).
16 See infra p. 74.
17 See Disclosure of Mutual Fund After-Tax Returns, Investment Company Act Release No. 33-7809, 65 Fed. Reg. 15,500 (Mar. 15, 2000).
18 Most notably, in 1970 Congress enacted Section 36(b) of the Investment Company Act to impose on advisers a fiduciary duty with respect to the amount of compensation that they receive; amended Section 15(c) to strengthen the ability of directors to scrutinize advisory contracts, and enacted Section 2(a) (19) to strengthen the standards for determining who may serve as an "independent" fund director. See Investment Company Act Amendments of 1970, Pub. Law No. 91-547, 84 Stat. 1413 (1970). See also S. Rep. 91-184 (1970), reprinted in 1970 U.S.C.C.A.N. 4897 (legislative history of the 1970 amendments); Division of Investment Management, Protecting Investors: A Half Century of Investment Company Regulation 257 n.14 (May 1992) (hereinafter "Protecting Investors").
19 The organizing entity might be an entity other than an adviser, such as a fund's administrator or its principal underwriter, which sells the fund's shares pursuant to an underwriting contract with the fund.
20 As enacted in 1940, the Investment Company Act had few limits on mutual fund fees, including sales loads and advisory fees. The Investment Company Act included a general prohibition on unconscionable or grossly excessive sales loads (that was modified in 1970 to prohibit excessive sales loads), to be defined by a securities association. See Investment Company Act of 1940, Pub. L. No. 76-768, § 22(b), 54 Stat. 789, 823 (1940) (codified as amended at 15 U.S.C. § 80a-22 (2000)); Investment Company Amendments Act of 1970, Pub. L. No. 91-547, § 12, 84 Stat. 1413, 1422 (1970) (codified as amended at 15 U.S.C. § 80a-22 (2000)). For example, in Saxe v. Brady, 184 A.2d 602 (Del. Ch. 1962), a leading case under the original Section 36, the court noted that because fund shareholders were properly informed of all material facts, plaintiffs had the burden of proving that the fee was so out of proportion to the value of services rendered as to make it unconscionable. Moreover, because the requisite disclosures to shareholders had been made, the court held that "corporate waste" and not fairness was the appropriate standard by which fees should be judged. The court made this finding even though it noted that:
[The adviser's] profits are certainly approaching the point where they are outstripping any reasonable relationship to expenses and effort even in a legal sense. And this is so even after making due allowance for incentive and benefit presumably conferred. This is not to say that no payment is justified after a fund reaches a particular size. It is only to say that the business community might reasonably expect that at some point those representing the fund would see that the management fee was adjusted to reflect the diminution of the cost factor.
Id. at 610. See also William P. Rogers and James N. Benedict, Money Market Fund Management Fees: How Much is Too Much?, 57 N.Y.U. L. Rev. 1059, 1074-88, & nn.79-88 (generally discussing the Saxe case). The National Association of Securities Dealers ("NASD") has promulgated a rule prohibiting NASD members from selling mutual fund shares if the sales charges on the shares exceed specified caps. See NASD Rule 2830, NASD Manual, (CCH) ¶ 4621 (2000).
21 Burks v. Lasker, 441 U.S. 471, 484 (1979).
22 Section 10(a) of the Investment Company Act of 1940, Pub. L. No. 76-768, § 10(a), 54 Stat. 789, 806 (1940) (codified as amended at 15 U.S.C. § 80a-10 (2000)).
23 Section 15(a) of the Investment Company Act generally makes it unlawful for any person to serve as an investment adviser to a fund except pursuant to a written contract that has been approved by a majority of the fund's outstanding voting securities and a majority of the fund's independent directors. Typically, the adviser, as the initial shareholder of the fund, initially approves the contract. After the initial two-year contractual period, Section 15 requires that the contract be renewed annually by a majority of the fund's independent directors or its shareholders. Similarly, Section 15 requires that the fund's underwriting contract be approved annually by a majority of the fund's independent directors. See 15 U.S.C. § 80a-15 (2000).
24 Section 15(c) of the Investment Company Act, 15 U.S.C. § 80a-15(c) (2000).
25 See Protecting Investors, supra note 18, at 256-258 (discussion of board evaluation of mutual fund fees). See infra pp. 20-21, for a discussion of the factors that directors consider when reviewing investment advisory contracts.
26 Rule 12b-1(b) under the Investment Company Act, 17 C.F.R. § 270.12b-1(b) (2000). A Rule 12b-1 plan also must be approved by a majority of the outstanding voting securities of the fund. See 17 C.F.R. § 270.12b-1(b)(1) (2000).
27 Rule 12b-1 addresses the potential conflicts of interest between a fund and its investment adviser that are created when a fund bears its own distribution expenses. An investment adviser that receives an asset-based advisory fee has an incentive to increase the amount of the fund's assets because the fee received would become larger as assets grow. As a result, an investment adviser often will pay for marketing expenses itself in order to increase the asset size of the fund. When a fund pays its own distribution expenses through a 12b-1 plan, both the adviser and fund shareholders may benefit from the increased size of the fund, but the adviser is spared the cost of paying for the distribution expenses itself.
28 We note that the NASD has imposed an annual cap on asset-based sales charges of 0.75% of average annual net assets and an additional 0.25% for service fees. See NASD Rule 2830, NASD Manual, (CCH) ¶ 4621 (2000). The NASD took this action to assure that shareholders paying for distribution indirectly through Rule 12b-1 fees would pay no more than shareholders paying for distribution directly through front-end loads. See Form 19b-4, Notice of Proposed Rule Change by National Association of Securities Dealers, Inc. Relating to the Limitation of Asset-Based Sales Charges as Imposed by Investment Companies, Exchange Act Release No. 29,070, 48 S.E.C. Docket 976 (Apr. 12, 1991).
29 In the adopting release to rule 12b-1, the Commission identified certain factors that the directors should consider, if applicable, when reviewing and approving a rule 12b-1 plan. Among other factors, the Commission stated that directors should consider the nature of the problems or circumstances which purportedly make implementation or continuation of such a plan necessary or appropriate; consider the causes of such problems or circumstances; and consider the way in which the plan would address these problems or circumstances and how it would be expected to resolve or alleviate them, including the nature and approximate amount of the expenditures; the relationship of such expenditures to the overall cost structure of the fund; the nature of the anticipated benefits, and the time it would take for those benefits to be achieved. See Bearing of Distribution Expenses by Mutual Funds, Investment Company Act Release No. 11,414, 45 Fed. Reg. 73,898, 73,904 (Oct. 28, 1980). In addition, the Commission stated that directors should consider the possible benefits of the plan to other persons compared to those expected to inure to the fund, and, in the case of a decision on whether to continue a plan, whether the plan has in fact produced the anticipated benefits for the fund and its shareholders. Id.
30 Because an adviser's duty under Section 36(b) applies to all fees received by the adviser and its affiliates, a fund's board of directors should review the dollar amounts paid and the services performed under any service contract between the company and the adviser or its affiliates. See Protecting Investors, supra note 18, at 258 and nn.23-24.
31 See S. Rep. No. 91-184 (1969), reprinted in 1970 U.S.C.C.A.N. 4897.
32 Congress adopted Section 36(b) as part to the 1970 amendments to the Investment Company Act in response to concerns that advisory fees were not subject to usual competitive pressures because of the external management of mutual funds. The Commission had recommended amendments that, among other things, required that compensation received by affiliated persons of investment companies for services furnished to the company be reasonable and that this standard be enforceable in the courts. Rather than impose a reasonableness standard, however, Congress imposed the fiduciary duty of Section 36(b). See Protecting Investors, supra note 18, at 317-19 (discussion of legislative history of Section 36(b)).
33 See Section 36(b) of the Investment Company Act, 15 U.S.C. § 80a-35(b) (2000). See also Krinsk v. Fund Asset Mgmt., Inc., 715 F. Supp. 472, 485 (S.D.N.Y. 1988), aff'd, 875 F.2d 404 (2d Cir. 1989).
34 See Krinsk, 875 F.2d at 412; Schuyt v. Rowe Price Prime Reserve Fund, Inc., 835 F.2d 45 (2d Cir. 1987); Gartenberg v. Merrill Lynch Asset Management, Inc., 694 F.2d 923, 930 (2d Cir. 1982); Kalish v. Franklin Advisers, Inc., 742 F. Supp. 1222 (S.D.N.Y. 1990), aff'd 928 F.2d 590 (2nd Cir. 1991).
35 See Gartenberg, 694 F.2d at 928; Krinsk, 875 F.2d at 409.
36 See Krinsk v. Fund Asset Mgmt., Inc., 875 F.2d 404 (2d Cir. 1989).; Schuyt v. Rowe Price Prime Reserve Fund, Inc., 835 F.2d 45 (2d Cir. 1987); Gartenberg v. Merrill Lynch Asset Management, Inc., 694 F.2d 923, 930 (2d Cir. 1982); Kalish v. Franklin Advisers, Inc., 742 F. Supp. 1222 (S.D.N.Y. 1990). Although the courts note that fees charged by other funds is not the principal factor to be considered in evaluating a fee under Section 36(b), such comparative information is significant.
37 Section 8 of the Investment Company Act requires mutual funds to register with the Commission. 15 U.S.C. § 80a-8 (2000). If the fund is conducting a public offering of its shares, it also must file a registration statement to register the offering of those shares under the Securities Act of 1933 ("Securities Act"). Form N-1A is used by a mutual fund both to register the fund under the Investment Company Act and to register the offering and sale of shares under the Securities Act. The registration statement includes the fund's prospectus.
38 Consolidated Disclosure of Mutual Fund Expenses, Investment Company Act Release No. 16,244, 53 Fed. Reg. 3192 (Feb. 1, 1988) (adopting release); Investment Company Act Release No. 15,932, 52 Fed. Reg. 32018 (Aug. 18, 1987) (reproposing release); Investment Company Act Release No. 14,230, 49 Fed. Reg. 45171 (Nov. 9, 1984) (proposing release).
39 Registration Form Used by Open-End Management Investment Companies, Investment Company Act Release No. 23,064, 63 Fed. Reg. 13916 (Mar. 13, 1998) (hereinafter "Form N-1A Adopting Release").
40 The fee table is Item 3 of Form N-1A.
41 The Commission also made several improvements to the fee table itself. For example, in order to give investors clearer information about the long-term costs of an investment, the Commission modified the manner in which a fund may show the effect of expense reimbursements and fee waiver arrangements that temporarily reduce costs. See Form N-1A Adopting Release, supra note 39, at 13924-25.
42 See SEC Chairman Arthur Levitt, Sept. 28, 1998 Testimony before the Subcomm. on Finance and Hazardous Materials of the House Comm. on Commerce, (visited Nov. 8, 2000) www.sec.gov/news/testmony/tsty1398.htm (concerning transparency in the United States debt markets and mutual fund fees and expenses).
43 See also Investment Company Institute, Report of the Advisory Group on Best Practices for Fund Directors: Enhancing a Culture of Independence and Effectiveness (June 24, 1999).
44 Role of Independent Directors of Investment Companies, Investment Company Act Release No. 24,082, 64 Fed. Reg. 59826 (Oct. 14, 1999).
45 Interpretive Matters Concerning Independent Directors of Investment Companies, Investment Company Act Release No. 24,083, 64 Fed. Reg. 59877 (Oct. 14, 1999).
46 See Mutual Fund Investing: Look at More than A Fund's Past Performance (last modified Jan. 1, 2000) http://www.sec.gov/consumer/mperf.htm.
47 Mutual Fund Cost Calculator (last modified Sept. 6, 2000) http: www.sec.gov/mfcc/mfcc-int.htm. During the first quarter of 2000, the calculator averaged over 8,500 hits per month - making it one of the most frequented portions of the Commission's web site.
48 Investment Options (last modified Sept. 7, 2000) http://www.sec.gov/consumer/investop.htm.
49 Investment Company Institute, Frequently Asked Questions About Mutual Fund Fees (1998) http://www.ici.org/pdf/mf_fee_faqs.pdf.
50 Financial Facts Tool Kit (last modified Apr. 21, 1999) www.sec.gov/consumer/tookit.htm.
51 Invest Wisely, An Introduction to Mutual Funds, Advice from the U.S. Securities and Exchange Commission (last modified Aug. 1, 2000) www.sec.gov/consumer/ inwsmf.htm.
52 Search Key Topics (continuously updated) http://www.sec.gov/answers.shtml. See, e.g., Mutual Fund Fees and Expenses (last modified Oct. 19, 2000) http://www.sec.gov/answers/mffees.htm. Investors can also order a hard copy of this brochure by calling the SEC's toll-free publications line at 800-SEC-0330.
53 Disclosure of Mutual Fund After-Tax Returns, Investment Company Act Release No. 24,339, 65 Fed. Reg. 15500 (Mar. 15, 2000).
54 See Scott Cooley, Revisiting Fund Costs: Up or Down, Morningstar Mutual Funds, Feb. 21, 1999, at S1-S2. The fund groups are American Funds, Fidelity, and Vanguard. For information about the relative asset-weighted ownership cost of 30 large fund groups, see the data table at http://www.morningstar.com/ news/MS/ Commentary/ 990219com.msnhtml (visited Feb. 26, 1999).
55 See Lipper Inc., Lipper Directors' Analytical Data app. (1st ed. 2000) (Summary Table by Complex). The asset figures include stock, bond, and money market mutual funds and exclude underlying mutual funds of insurance company separate accounts. For stock funds, the market share of the three fund families in 1998 was 35%. See John Rekenthaler, Which Way is Up? The Debate About Fund Costs (visited Dec. 23, 1998), http://www.morningstar.com/ news/MS /IvoryTowers/ 981223Rek.msnhtml.
56 See Janet Novack, Custom-made Mutual Funds, (visited Sept. 11, 2000) http://www.forbesbest.com/0911/072.htm
57 See New online brokers let you build your own mutual fund at a bargain price, S.F. Chron. (visited Aug. 1, 2000) http://www.sfgate.com/cgi-bin/article.cgi? file=/chronicle/archive/2000/08/01BU107294.DTL. For a description of FOLIO[fn], one version of this type of product, see Financial Research Corp., Shake and Bake Mutual Funds: Technology Enables Creation of lndividualized Mutual Funds, Mutual Fund Cafe (visited Nov. 8, 2000) http://www.mfcafe.com/pantry/bps_062600.html.
58 In what may be a sign of things to come, the Vanguard Group recently announced that it would reduce the fees charged to fund shareholders with large account balances and long holding periods - generally speaking, a fund's preferred customer-base. Fees paid by large, long-term investors in one fund, the Vanguard Index 500, for example, would be reduced from 18 basis points to 12 basis points. One commentator speculates that this reduction is an attempt to compete with ETFs. Dan Culloton, Vanguard Lets Big Retail Investors Become Admirals, (visited July 26, 2000) http://www.morningstar.com/news/Wire/0,1230,2393,00.html. The fee rate charged to holders of the largest ETF, Standard & Poor's Depository Receipts Trust, Series 1 -- popularly known as Spiders -- is 12 basis points.
59 The management expense ratio is the dollar amount of a fund's management expenses divided by its average net assets. Management expenses include payments made by the fund to its investment adviser (or to affiliates of the adviser) for investment management, administrative or other services. See infra Section III.B.1 (What Costs are Included in a Fund's Expense Ratio?)
60 Some funds define the term management fee narrowly, to cover only the cost of selecting portfolio securities. These funds pay for administration, record keeping, and other services under separate contracts with other service providers. Other funds define the management fee broadly, to cover a variety of administrative and other services, in addition to expenses associated with selecting portfolio securities. A few funds have "unified" fees under which the management fee pays for all fund expenses (the management fee is equal to the expense ratio). Thus, if Fund A has a higher management fee than Fund B, it may mean that Fund A pays a higher fee to its adviser. Alternatively, it may mean that Fund A's management fee pays for services that are provided and charged for separately by Fund B's adviser, an affiliate of the adviser, or outside contractors.
61 Rule 12b-1 fees are most commonly used to pay for sales commissions, printing prospectuses and sales literature, advertising, and similar expenses. Some funds, however, adopt 12b-1 fees to cover expenses considered by other funds to be advisory or administrative expenses for which no plan may be required. To complicate the issue further, a fund might pay broker-dealer firms under a 12b-1 plan for services provided to fund shareholders who are the broker-dealer's customers while paying banks under an administrative agreement for providing the same services to fund shareholders who are bank customers. In addition, because it is unclear what expenses are properly considered distribution expenses, some funds, out of an abundance of caution, adopt "defensive" 12b-1 plans. Defensive plans exist solely to ensure that if a court found any fund operating expense to be also a distribution expense, the expense would be covered under a 12b-1 plan. The result: some funds have 12b-1 plans although no assets are used for distribution purposes. Similarly, other funds, that do use their assets to pay for distribution, extend their 12b-1 plans to cover operating expenses as well.
62 See Protecting Investors, supra note 18, at 320-26.
63 The sales load -- representing the difference between the price per share at which fund shares are offered to the public and the net amount per share invested in the fund -- is retained by a fund's principal underwriter and/or the selling broker-dealer and no part is paid to the fund. The sales load is used to finance the broker's commissions, other sales and promotional expenses, and the underwriter's profit (if any).
64 During the 1970s, the Commission received a number of requests to allow fund assets to be used to pay for distribution. Reasons cited to approve these requests included rising net redemptions, growing public resistance to high front-end sales loads, the increased popularity of no-load funds, and the availability of competing investment products without front-end loads. Another rationale was that use of fund assets for distribution expenditures would result in a net flow of cash into funds, and in turn, economies of scale and more effective portfolio management. In 1979, after extensive consideration, the Commission proposed rule 12b-1, stating that funds should be permitted to bear distribution expenses if they were disclosed and regulated. Bearing of Mutual Fund Expenses by Shareholders, Investment Company Act Release No. 10,862, 44 Fed. Reg. 54014 (Sept. 17, 1979). The Commission adopted rule 12b-1 in October 1980. Bearing of Distribution Expenses by Mutual Funds, Investment Company Act Release No. 11,414, 45 Fed. Reg. 73898 (Oct. 28, 1980).
65 Investment Company Institute, Statement of the Investment Company Institute Regarding the Operation of Rule 12b-1 Plans, 23 (Aug. 8, 1986).
66 See Protecting Investors, supra note 18, at 294.
67 Many fund families offer their funds in a multi-class structure. One common structure consists of a share class with a front-end load and a small 12b-1 fee, commonly referred to as "A Shares"; a share class with a CDSL and a larger 12b-1 fee that expires after, typically, 5-8 years, commonly known as "B Shares"; and a share class with a larger 12b-1 fee that never expires, but no front-end load or CDSL, commonly referred to as "C shares".
68 Order Approving Proposed Rule Change Relating to the Limitation of Asset-Based Sales Charges as Imposed by Investment Companies, Securities Exchange Act Release No. 30,897, 57 Fed. Reg. 30985 (July 7, 1992).
69 A basis point is equal to 1/100 of 1%.
70 See infra Section III.C.4 for a further discussion of total shareholder cost analysis.
71 Although we attempted to use all available data, we eliminated some funds from the study because of missing data. For example, in 1999 the Morningstar Principia Pro database included observations for 11,078 classes. We excluded 2,177 classes because they were missing data for one or more of the variables in our regression. There were 1,084 classes without values for the expense ratio, and another 1,093 classes without values for one or more of the independent variables. This left us with 8,901 classes for which we have complete data.
72 The Morningstar databases use fund classes, rather than funds as the basic data item. The ramifications of this approach are discussed below and infra, note 97.
73 See supra note 67 and accompanying text.
74 Master-feeder arrangements are another organizational structure that is designed to offer additional choice to fund investors. Like a "regular" mutual fund, a master fund invests in stocks, bonds, and other portfolio securities. Unlike a regular mutual fund, the master fund distributes its shares not directly, but through other funds (feeder funds). A feeder fund sells its shares to the public, but invests only in shares of the master fund. Feeder funds, like classes, may offer varying levels of service or alternative ways of paying for distribution costs. The Morningstar Principia Pro database includes feeder funds as separate observations. Principia Pro identified 556 feeder funds with total assets of more than $200 billion as of March 31, 2000.
75 Although investors purchase shares of a specific class and incur that classes' expenses, analysis of fund expenses at the class level can sometimes produce anomalous results. Consider the following example: Class S of Big Fund, Inc. (Big Fund: S) is a small (in terms of asset size) share class of a very large fund. Small Fund, Inc. is identical to Big Fund: S in all respects (same asset size, investment objective, etc.) except that it is a stand-alone fund. Big Fund: S is likely to have a lower expense ratio than Small Fund, Inc. because Big Fund: S is likely to benefit from scale economies that are produced by Big Fund's other, larger classes. A mutual fund expense analysis that is performed at the class level would incorrectly identify Big Fund: S as a small fund with low expenses, when it may more appropriately be identified as a large fund with low expenses.
76 In constructing our econometric model, we consider each class of a multi-class fund to have an asset size equal to the sum of the assets of all the classes that share a common investment portfolio. See infra, note 98 and accompanying text.
77 All mutual funds are required to provide reports to shareholders, including expense ratios, 60 days after the end of their fiscal years. To capture data on a calendar year basis, we used Morningstar data for the end of March.
78 Although we recognize that the sample may not adequately portray the experience of smaller funds, we believe that the sample reflects the results that are likely to be experienced by funds with the most assets and the most shareholders.
79 For this analysis, multiple-class funds were evaluated at the fund level because all classes of a multiple class fund are subject to a single management contract.
80 An equally weighted average assumes that all members of a population are equally important and gives equal weight to all data points. In populations where some members are more important than others, an average that gives more weight to the more important members (weighted average) may be more appropriate.
81 It would appear that the weighted expense ratio increased in 1999 as a result of the growth in assets of equity, international and specialty classes relative to bond classes. Assets of equity classes increased 2.9%, international classes increased 1.7% and specialty classes increased 2.0%, while assets of bond classes declined 6.6%. Because equity, international and specialty classes generally have higher expense ratios than bond classes, any increase in the proportion of assets in these investment categories would tend to increase the weighted average for all classes. See infra, Section III.C.5.
82 For a discussion of the extent to which lines between mutual fund distribution expense categories and marketing channels have become blurred, see Financial Research Corp., The Alphabet Soup of Share Classes: Or Whatever Happened to Simplicity, (visited Aug. 30, 2000) http://www.mfcafe.com/pantry98/bps_100598.html.
83 We refer to classes that may call themselves no-load under current NASD rules as "extended no-load classes." The data for pure no-load classes and extended no-load classes are broken out separately in Appendix Two.
84 Sales load data reported by Morningstar are the maximum sales loads charged.
85 See John D. Rea and Brian K. Reid, Trends in Ownership Cost of Equity Mutual Funds, Investment Company Institute Perspective, Nov. 1998, at 4 ("Rea and Reid"). This study found that, for stock mutual funds, sales-weighted average shareholder costs decreased from 2.25% of new investments in 1980 to 1.49% of new investments in 1998 -- a decrease of almost 34%. Stock fund operating costs rose by 12 basis points during the period, however. Subsequent Investment Company Institute studies have yielded similar results. See generally, John D. Rea and Brian K. Reid, Total Shareholder Cost of Bond and Money Market Mutual Funds, Perspective, Mar. 1999, at 5; John D. Rea et al., Operating Expense Ratios, Assets, and Economies of Scale in Equity Mutual Funds, Perspective, Dec. 1999.
86 Data about the maximum sales load that investors could pay are readily available. Data about the extent to which investors actually pay less than the maximum sales load (because they are eligible for discounts for large purchases, for purchases through retirement accounts, or for other reasons) are not available.
87 Available data about investor holding periods are limited, and anecdotal evidence is contradictory. Looking first at the overall picture, during recent years, the annual redemption rate (redemptions as a percentage of average assets) for all stock funds has been 17-18%, implying an average holding period of just under 6 years. The annual redemption rate for all bond funds has been roughly 20%, implying an average holding period of 5 years (See ICI Fact Book, supra note 3, at 69, 87). A recent article in the trade press cited 5 years as the average mutual fund holding period. Gavin Daly, Edward Jones Starts Selling Funds in U.K., Ignites.com, (visited Dec. 13, 1999) http://www.ignites.com. Another article claimed 3 years as the average holding period for stock funds, citing a long-term study of investor behavior by Dalbar, Inc., a mutual fund research firm. Stock Fund Investors Who Stay Put Double Returns: Dalbar, Dow Jones News Svc., Dec. 8, 1999. Financial Research Corporation, another mutual fund research firm, concludes that, based on an analysis of figures published by the Investment Company Institute, the average holding period for mutual funds has declined from about 10 years in the early 1990s to a current holding period of two-and-a-half years. Financial Research Corp., Is Three the Magic Number?, Mutual Fund Café, (visited Oct. 9, 2000) http://www.mfcafe.com/blue/bps.html. Some observers believe that as access to information has increased and trading has become easier, the average holding period has declined. See, e.g., Darlene DeRemer, et al., High Turnover May be Hurting Fund Company Profits, Mutual Fund Cafe, (last modified Nov. 1998) http://www.mfcafe.com/pantry/is_1198.html. Others argue that a minority of active traders are skewing the statistics and that a large majority of fund shareholders are buy- and-hold, long term investors. See, e.g., Gavin Daly, Fears about Short-Term Trading Called Overblown, (visited Aug. 23, 2000) http://www.ignites.com (citing results from a study conducted by Strategic Insight, a mutual fund consulting firm). Of course, aggregate figures about average holding periods may conceal wide variations among different groups of investors and funds. For example, according to one recent article, the typical holding period for an investor who utilizes the Charles Schwab mutual fund supermarket is "...in the two-to-three year range." Bridget O'Brian and Pui-Wing Tam, More and More Dollars Flow to Hotshot Funds, Wall St. J., June 7, 1999, at R1 (quoting Guy Mozkowski, an asset-management analyst at Salomon, Smith Barney, Inc.) In contrast, other recent articles indicate that for one large load fund family the average holding period is 12 years; and that clients of one medium-size brokerage firm hold fund shares for more than 18 years, on average. Oster, Capital Appreciation, Smart Money, Mar. 1999, at 130-35.
88 Rea and Reid used holding period estimates contained in a study performed by The Wyatt Company for the NASD in 1990. The Wyatt Company selected a random sample of stock and bond fund accounts that were opened in 1974 at funds with front-end loads and determined the percentage of the original share purchases that was redeemed in each of the subsequent 15 years. See Rea and Reid, supra note 85, at 7.
89 E. Sirri and P. Tufano, Competition and Change in the Mutual Fund Industry, Financial Services: Perspectives and Challenges, 190-91 (1993).
90 International funds invest in stocks and bonds of non-U.S. companies and governments. Specialty funds (sometimes referred to as sector funds) concentrate their investments in specific industries or industry sectors.
91 See, e.g., Andrew Leckey, Market Sag Puts a Harsher Light on Fund Fees, Chicago Tribune, Mar. 11, 2000, available in 2000 WL 3644678. Are Your Managers Overpaid?, Los Angeles Times, at S6, Oct. 10, 1999, available in 1999 WL 26182762. Scott Cooley, Revisiting Fund Costs: Up or Down?, Morningstar Mutual Funds, Feb. 21, 1999, at S1.
92 See Lipper Analytical Services, Inc., The Third White Paper: Are Mutual Fund Fees Reasonable? at 12-13 (Sept. 1997).
93 Results of the econometric model presented in the next section differ from the results described in this section. The results of the model show that as funds get older, their expense ratios increase.
94 A number of researchers have used similar mathematical models in their studies of issues related to mutual fund expenses. See, e.g., Stephen P. Ferris and Don M. Chance, "The Effect of 12b-1 Plans on Mutual Fund Expense Ratios: A Note," 42 J. Fin.1081 (1987); Don M. Chance and Stephen P. Ferris, "Mutual Fund Distribution", 5 J.Fin. Services Res. 39 (1991); Charles Trzcinka and Robert Zweig, An Economic Analysis of the Cost and Benefits of S.E.C. Rule 12b-1 at 22 (N.Y.U. Leonard School of Business Monograph Series in Finance and Economics No. 1990-1, 1991).
95 The basic model is as follows:
E=a + b1*Ln(Assets) + b2*(1/Famsize) + b3*Ln(Famnum) + b4*Ln(Turnover) + b5*Ln(Holdings) + b6*Ln(Age) + b7*Equity + b8*Hybrid + b9*I Bond + b10*I Equity + b11*Specialty + b12*Index + b13*Institution + b14*Load + b15*Class + b16*12b-1 + e
|E||=class's expense ratio|
|Ln(Assets)||=natural log of fund's net assets in millions|
|1/Famsize||=reciprocal of family net assets in millions|
|Ln(Famnum)||=natural log of number of funds in family|
|Ln(Turnover)||=natural log of class's turnover|
|Ln(Holdings)||=natural log of number of issues in class's portfolio|
|Ln(Age)||=natural log of fund's age in years|
|Equity||=an indicator variable that equals 1 if the fund is a domestic equityfund, 0 otherwise|
|Hybrid||=an indicator variable that equals 1 if the fund is a domestic hybrid fund, 0 otherwise|
|I Bond||=an indicator variable that equals 1 if the fund is an international bond fund, 0 otherwise|
|I Equity||=an indicator variable that equals 1 if the fund is an international equity fund, 0 otherwise|
|Special||=an indicator variable that equals 1 if the fund is a specialty fund, 0 otherwise|
|Index||=an indicator variable that equals 1 if the fund is an index fund, 0 otherwise|
|Institution||=an indicator variable that equals 1 if the class is an institutional class or fund, 0 otherwise|
|Load||=an indicator variable that equals 1 if the class has a front-end load, 0 otherwise|
|Class||=an indicator variable that equals 1 if the class is part of a multi-class fund, 0 otherwise|
|12b-1||=maximum 12b-1 fee|
96 We define a factor as important if its t test statistic is greater than the critical value, approximately 1.96. At this value, we are statistically confident 95% of the time that the attribute is associated with an effect on the expense ratio. The t test statistic for each expense factor is shown in Appendix One.
97 Our approach of using classes, rather than funds, as a data item presents two problems in our regression analysis. First, it potentially gives more weight to the results of multi-class funds than to the results of single-class funds. Second, not all observations are independent of each other. One of the fundamental assumptions of regression analysis is that the observations are independent. While each class typically has its own expense ratio, many fund expenses, including the management fee, are incurred at the portfolio level and then allocated among a fund's classes typically based on the relative net assets of each class. Other expenses, including 12b-1 fees and some administrative fees, are incurred directly at the class level. Because a fund's classes bear many expenses in common, the operating expense ratios of a fund's classes usually are very similar and frequently are identical. In addition, most of the independent variables in the model are identical across classes in the same fund. This lack of independence among observations may cause the regression analysis to understate the standard errors and overstate the t-statistics. To determine whether our approach led to erroneous conclusions, we also regressed a proxy for operating expenses (the expense ratio less the maximum 12b-1 fee) on the independent variables exclusive of the maximum 12b-1 fee. In this second model we used only one observation for each fund. For multi-class funds we used as the expense ratio variable the asset-weighted average operating expense ratio of all classes in the fund. The institutional and load variables were the proportion of assets in classes with these characteristics. The results of this model are not qualitatively different from the results presented in this section. The coefficients of the second model are very similar to those of the basic model and all remain statistically significant.
98 Although each fund class is represented as a separate data item, with its own expense ratio, the asset size of each class is calculated as the sum of the assets of all classes that that we could identify as sharing a common investment portfolio. In other words, asset size is calculated at the fund level. The age of a fund is considered to be the age of the fund's oldest class.
99 Our standard errors also may be biased downward because expense ratios among the funds in a fund family are likely not independent.
100 The reader should note that, for certain factors (fund assets, number of funds in the fund family, number of portfolio holdings, and turnover) the associated variable in our model is the natural logarithm of the factor. For a second group of factors (those associated with a fund's investment category, whether not it is an index, institutional, or multi-class fund) the factor in the model is known as an indicator variable. That is, the value of the factor in the model can be only 1 or 0.
101 A number of funds that are part of very small fund families have, everything else equal, relatively high operating expense ratios. We did not observe a relationship between fund family assets and operating expense ratios for funds that are members of larger fund families (except, as noted in note 110, with respect to four large fund families). One way of capturing this relationship is to include as an independent variable the reciprocal of fund family assets. The t-statistic for the coefficient of the reciprocal of family assets is considerably larger than that obtained when the natural logarithm of fund family assets is used, further supporting the reciprocal as the better measure of the relationship.
102 If the coefficient were equal to 1.0, then everything else held constant, funds with 12b-1 fees would have expenses that are higher than the expenses of other funds by an amount that equaled the maximum 12b-1 fee.
103 Management expenses consist of fees paid for investment advice and other services provided under a fund's management contract. Not all funds account for management expenses in the same way, however. Some funds define the management fee narrowly, to cover only the cost of selecting portfolio securities, while other funds define it more broadly, to cover a variety of administrative and other services. See supra Section III.B.1 (What Costs are included in a Fund's Expense Ratio?).
104 The 1,000 classes included in the regression analysis represent approximately 82% of fund assets in 1999. The smallest class in the sample had assets of $704 million in 1999. Although we recognize that the sample may not adequately portray the experience of smaller funds, we believe that the sample reflects the results that are likely to be experienced by the funds with the most assets and the most shareholders.
105 See, e.g., Protecting Investors, supra note 18, at 256 n.12 ("Advisory fees typically are calculated as a percentage of assets under management, although the cost of providing investment advisory services -- consisting largely of salaries and overhead -- is relatively fixed (i.e., a portfolio manager can manage $500 million nearly as easily as $100 million.) An advisory fee that does not scale down as company assets increase consequently may yield enormous profits to the adviser, to the detriment of shareholders.").
106 See supra note 60 and accompanying text.
107 Although breakpoints are not legally required to be included in the advisory contract, "the fee structures of many funds have been specifically designed to pass along economies of scale by means of breakpoints." Statement of Matthew P. Fink, President, Investment Company Institute, before the Subcomm. on Finance and Hazardous Materials of the House Comm. on Commerce, Sept. 29, 1998, at 21-22, available in 1998 WL 18088868.
108 See 2 Tamar Frankel, The Regulation of Money Managers 260 (1978).
109 See id.
110 Although the magnitude of change in a fund's management expense ratio that is associated with changes in fund family asset size appears to be large, this result may be attributable to four large fund families. When we reran the regression model with the four fund families omitted, we found no statistically significant relationship between a fund's management expense ratio and the asset size of its fund family.
111 Other fund attributes found to be important in explaining a fund's management expense ratio in 1999 were investment category, portfolio turnover, fund age, and whether or not a fund is an index fund or an institutional fund. Equity funds had higher management expense ratios than bond funds, and international and specialty funds had higher management expense ratios than equity funds. Funds with more portfolio turnover had higher management expense ratios. Older funds had lower management expense ratios than newer funds.
112 The 100 largest fund portfolios had combined assets of $1.4 trillion, in 1997, $1.6 trillion in 1998, and $2.0 trillion in 1999. The assets of these funds represented 47% of all stock and bond fund assets in 1997 and 45% of total assets in 1998, and 1999. We observed that during the three-year period some funds adjusted their breakpoints to account for more assets, and that in 1999 the funds in one large fund complex eliminated their fee breakpoint arrangements.
113 Investment Company Institute, Mutual Funds and the Retirement Market, Fundamentals: Investment Company Research in Brief, May 2000, at 1-2.
114 We recognize that not all expenses associated with 401(k) plans are included in mutual fund expense ratios.
115 Another option would be to mandate that mutual funds include in their prospectuses or shareholder reports a new standardized "ending-value" table. The ending value table would utilize historical information about a fund's expenses to illustrate how seemingly small changes in expenses can have a large impact on the amount of money accumulated for a long-term goal. For example, if a retirement saver invested $5,000 per year starting at age 25, earned an average annual rate of return of 9% over 40 years, and incurred no expenses, his or her ending value would be $1,841,459. If the same investment were subject to annual expenses of 50 basis points, his or her ending value would be reduced by more than $257,000, or 14%.
The ending value table would compare the ending value after ten or twenty years of an investment (e.g., $10,000) that incurred the fund's historical expense ratio, to the ending value of an investment that incurred an expense ratio of zero, 1%, or any other number mandated by Commission rule. The expense numbers would be applied to a standardized return such as 5% (the return used in the fee table example) or a number between 9-12% that would reflect the historical return on equities over the last 20-80 years. The table would enable investors to readily compare funds with respect to the long-term impact of fund expenses on the ending value of an account.
For more information about the long-term effect of expenses on the ending value of an investment account see Mamta Murthi, J. Michael Orzag, and Peter R. Orzag, The Charge Ratio on Individual Accounts: Lessons from the U.K. Experience, (Birkbeck College, University of London, Discussion Papers in Economics, Mar. 1999).
116 The Investment Company Institute produces a series of educational brochures, in English and Spanish, to help individuals make well-informed investment decisions. These include "Frequently Asked Questions About Mutual Fund Fees." In reference to efforts of the ICI to educate investors, Chairman Levitt recently stated, "[T]here is no better way to bring opportunity to more people than to educate them on the fundamentals of sound investing. By providing the guidance and resources for these programs, the ICI moves more Americans closer to realizing their long-term financial goals." SEC Chairman Arthur Levitt, Address on the 60th Anniversary of the Investment Company Act and the Investment Company Institute, Oct. 5, 2000 (last visited Dec. 15, 2000) http://www.sec.gov/news/speeches/spch403.htm.
117 KPMG Peat Marwick, LLP, An Educational Analysis of Tax-Managed Mutual Funds and the Taxable Investor 14 (1999). The KPMG study analyzed the performance of 496 domestic stock funds for the ten years ended December 31, 1997. The average annual total return for the median fund in this group was 16.1% before taxes and 13.5% after taxes. (The median fund is the fund at the midpoint of the frequency distribution. An equal number of funds have a higher or lower return than the median fund.) Annual performance given up to taxes ranged from a low of zero to a high of 5.6%, with a median of 2.6%.
118 See Role of Independent Directors of Investment Companies, Investment Company Act Release No. 24,082, 64 Fed. Reg. 59,826 (Nov. 3, 1999).
119 Information may be available from a variety of legal, accounting, and academic organizations. The Directors Program Committee of the Investment Company Institute sponsors a number of educational and information programs for fund directors. We also believe that the recently formed Mutual Fund Directors Education Council (described in Section II.B.2.) will serve as a useful source of information for fund directors. As part of the Council's plan to develop programs to promote a culture of independence and accountability in the boardroom, we recommend that the Council focus on the directors' role in negotiating fees and expenses.
120 Any study of the costs of investment management would require fact-finding and analysis similar to that previously conducted by the Wharton School of Finance and Commerce ("Wharton School"). The Commission retained the Securities Research Unit of the Wharton School of Finance and Commerce of the University of Pennsylvania to make a fact-finding survey and report on certain aspects and practices of registered investment companies. See Investment Company Act Release No. 2,729, 1958 WL 5755 *1 (SEC) (Jun. 13, 1958). The Wharton School produced A Study of Mutual Funds, H.R. Rep. No. 2274, 87th Cong., 2d Sess. 491-95 (1962), which formed a basis for the 1970 amendments to the Investment Company Act.
121 As described supra in Section III.B., other fund costs that may be paid for outside of the fund's expense ratio include costs related to marketing and distribution, financial advice to fund investors, and maintenance of shareholder accounts. In many cases, some or all of these costs may be paid separately by the shareholder.
122 During the period 1989-1993, according to one study, the average stock fund paid annual brokerage commissions equal to 0.28% of net assets. This figure excludes the market impact costs of fund portfolio transactions, i.e., changes in the price of a security that result directly from a fund's trading activity. See Miles Livingston and Edward O'Neal, Mutual Fund Brokerage Commissions, 19 J.Fin.Res. 272 (1996).
123 Although a mutual fund's investment manager has an obligation to seek the best execution of securities transactions arranged for on behalf of the fund, the manager is not obligated to obtain the lowest possible commission cost. The manager's obligation is to seek to obtain the most favorable terms for a transaction reasonably available under the circumstances. See Securities Brokerage and Research Services, Exchange Act Release No. 23,170, 51 Fed. Reg. 16,004, 16,011 (Apr. 23, 1986). Section 15(c) of the Investment Company Act requires a fund's board of directors to request and review, and the fund's manager to supply, such information as may reasonably be necessary for the fund's board to evaluate the terms of the advisory contract between the adviser and the fund. Research and other services purchased by the adviser with the fund's brokerage bear on the reasonableness of the advisory fee because the research and other services would otherwise have to be purchased by the adviser itself, resulting in higher expenses and lower profitability for the adviser. Therefore, mutual fund advisers that have soft dollar arrangements must provide their funds' boards with information regarding their soft dollar practices. See SEC Office of Compliance, Inspections and Examinations, Inspection Report on the Soft Dollar Practices of Broker/Dealers, Investment Advisers and Mutual Funds 30 (Sept. 22, 1998) (hereinafter "Soft Dollar Report").
124 See Soft Dollar Report, supra note 123, at 5-6. Soft dollar arrangements developed as a means by which brokers discounted commission rates that were fixed at artificially high levels by exchange rules. Prior to 1975, institutional advisers took advantage of competition among brokers and their willingness to accept compensation lower than the fixed rates in order to recapture portions of the commissions paid on institutional orders. Fixed commission rates that far exceeded the costs of executing trades provided the fuel to support an increasingly complex pattern of practices to recapture portions of these commissions by advisers, including "give-ups" and other "reciprocal practices". Investment company managers directed give-ups to brokers that sold fund shares in order to motivate or reward such sales efforts. Fund managers also used give-ups as a reward for research ideas furnished by brokers to them in their capacity as investment advisers to funds. The Commission abolished the system of fixed commission effective May 1, 1975. Soon thereafter, Congress enacted Section 28(e) of the Securities and Exchange Act of 1934 in order to clarify that, under certain circumstances, an investment manager may pay more than the lowest available commission in recognition of research and other services provided by the broker-dealer. See id. at 6-7.
125 All advisers, including the investment advisers of mutual funds, have an obligation to act in the best interests of their clients and to place client interests before their own. They also have an affirmative duty of full and fair disclosure of all material facts to their clients. See 15 U.S.C. § 80b-6 (2000) (Section 206 of the Investment Advisers Act of 1940); S.E.C. v. Capital Gains Research Bureau, 375 U.S. 180 (1963).
Some of the funds that engage in directed brokerage disclose the practice in the prospectus, the statement of additional information, and/or the annual report to shareholders. Others use the footnotes to the financial statements to make the disclosure. In 1995, the Commission adopted accounting rules which require funds to report all expenses gross of off-sets or reimbursements pursuant to a directed brokerage arrangement. See 17 C.F.R. § 210.6-07(g) (2000) (Rule 6-07(g) of Reg. S-X).
This requirement is designed to allow investors to compare expenses among funds.
126 See generally Bearing of Distribution Expenses by Mutual Funds, Investment Company Act Release No. 11,414 (Oct. 28, 1980), 45 FR 73898 (hereinafter "Adopting Release"). When rule 12b-1 was adopted, the Commission stated the rule was intended to be flexible, and that the Commission would monitor and make adjustments as necessary. Id. at 22. Since 1980, the rule has not been substantively amended.
127 See Joel H. Goldberg and Gregory N. Bressler, Revisiting Rule 12b-1 under the Investment Company Act, 31 Rev. Sec. and Commodities Reg., 147, 147-152 (1998).
128 Id. at 151.
129 See Adopting Release, supra note 126. See also supra note 29 and accompanying text.
130 See Goldberg and Bressler, supra note 127, at 151.
131 See Goldberg and Bressler, supra note 127, at 151; Paul G. Haaga, Jr. and Michele Y. Yang, Distribution of Mutual Fund Shares: Rule 12b-1, (Practicing Law Institute '40 Act Institute, 1990).
132 See Goldberg and Bressler, supra note 127, at 151.
133 See supra Section III.B.2.
134 In a typical fund supermarket, the sponsor of the program - a broker-dealer or other institution - offers a variety of services to a participating fund and its shareholders. The services include establishing, maintaining, and processing changes in shareholder accounts, communicating with shareholders, preparing account statements and confirmations, and providing distribution services. For the services that it provides, the sponsor charges either a transaction fee to its customer or an asset-based fee, generally ranging from 0.25% to 0.40% annually of the average value of the shares of the fund held by the sponsor's customers. The asset-based fee is paid by the fund, its investment adviser, an affiliate of the adviser, or a combination of all three entities. See Letter from Douglas Scheidt, Associate Director of the Division of Investment Management to Craig S. Tyle, General Counsel of the Investment Company Institute (pub. avail. Oct. 30, 1998) at 2-4, available in 1998 WL 1543541 *2-4 (SEC 1998). (hereinafter "Investment Company Institute").
135 See The Shareholder Services Group, Inc. (pub. avail. Aug. 12, 1992) and Investment Company Institute, supra note 134.
136 See, e.g., Rochelle Kauffman Plesset and Diane E. Ambler, The Financing of Mutual Fund 'B Share' Arrangements, 52 Bus. Law. 1385 (1997); Tania Padgett, First Union Group Plans to Cater to Cash Needs of Mutual Fund Firms, American Banker, May 17, 1996, at 20; Michael Brush, Are Managers Counting on a Rubber Stamp?, N.Y. Times, Dec. 29, 1996, at F9.
137 Some distributors sell their rights to receive certain 12b-1 fees to a commercial bank or a finance company. Other distributors securitize their 12b-1 fees by transferring the rights to receive certain 12b-1 fees to a special purpose entity. The entity, in turn, issues one or more classes of securities. The holders of these securities receive payments of interest and principal from the cash flows generated by the 12b-1 fees. See Plesset and Ambler, supra note 136, at 1398-1402, 1405.
138 When investors and rating agencies evaluate the quality of asset-backed securities, a key criterion is the degree of assurance that the revenue stream of 12b-1 fees will remain uninterrupted over the life of the security. See Plesset and Ambler, supra note 136.
139 See Adopting Release, supra note 126. See also supra note 29 and accompanying text.