Charles Trzcinka

House Committee Testimony

Mutual Fund Fees and Competition in the Mutual Fund Industry

Testimony of Charles A. Trzcinka Before The House Committee on Commerce, Subcommittee on Finance and Hazardous Materials, September 29, 1998

Summary

My purpose in this testimony is to summarize the evidence on competition in the mutual fund industry and to advocate deregulation of sales fees and standardization of risk and return measures.

When mutual funds are compared across broad classes of investments, the mutual fund industry is spectacularly successful. If competition is defined within the mutual fund industry by comparing funds against each other, the story is very different. From the view of individual investors, a mutual fund is a consumer product where investors pay fees and get a return plus services. The theme of my work is simple. Investors have a hard time determining how much they are paying and an even more difficult time determining what they are getting. Some fees are hidden and many fees are charged in a complicated fashion. At best, the total fee can be estimated from the disclosure of most funds but if an investor decides to estimate fees, it is very difficult to compare portfolios of risky securities. There are limitations in applying all measures of risk and there is a lack of uniformity in their application. Finally, the law interferes with the market for financial consultants.

As a result of these difficulties, there is little pressure from investors to minimize fees. Disclosed fees have not changed much in this decade but the task of determining how much a mutual fund costs has become more complicated. Hidden fees appear to have grown. It is clear from the evidence that the current mixture of fees have little relationship with the quality of the fund when "quality" is defined as a better return for the risk taken. With this definition of quality, most economists view the competition in this market as imperfect and believe that the competition is not improving.

I believe the academic and non-academic evidence supports the following facts that are the basis for my conclusions.

• Total expenses paid by investors have not fallen over the past decade and probably have risen.
• There is no relationship between level of expense ratios and risk-adjusted performance except that large expense ratios substantially reduce performance.
• There is no evidence that managed mutual funds have performed better than funds that simply try to match an index or a combination of indices.
• There is little evidence of persistence of good performance, there is stronger evidence of persistence of poor performance.
• Good performance is rewarded by investors, poor performance is ignored except when the poor performance is extreme.
• Information available to investors on mutual fund portfolio management is poor.

In most successful industries, from calculators to VCRs, competition has dramatically reduced the prices that consumers pay. The success of the mutual fund industry has not produced the same price competition.

Introduction

I am Finance Professor at the State University of New York at Buffalo, who has authored numerous articles and monographs on mutual funds and institutional investing, such as The Costs and Benefits of SEC Rule 12b-1 and the Forbes Stock Market Course. The focus of my research and my public activities, such as working with the financial press, is the nature of competition in the market for mutual funds. My purpose in this testimony is to summarize the evidence on competition in the mutual fund industry and to advocate deregulation of sales fees and standardization of quantitative risk measures.

When mutual funds are compared across broad classes of investments, the mutual fund industry is spectacularly successful. Measured by the growth in assets, mutual funds have outperformed bank and insurance products, and have acquired assets faster than any competing financial product in history. The evidence is strong that investors continue to place a great deal of confidence in the industry. If competition is defined by comparing these broad types of financial institutions, mutual funds could hardly be more successful.

If competition is defined within the mutual fund industry by comparing funds against each other, the story is very different. From the view of individual investors, a mutual fund is a consumer product where investors pay fees and get a return plus some services. The theme of my work is simple. Investors have a hard time determining how much they are paying and an even more difficult time determining what they are getting. Some fees are hidden and many fees are charged in a complicated fashion. At best, the total fee can be estimated from the disclosure of most funds but, if an investor decides to estimate fees, it is very difficult to compare portfolios of risky securities. There are limitations in applying all measures of risk and there is a lack of uniformity in their application. Finally, the law interferes with the market for financial consultants.

As a result of these difficulties, there is little pressure from investors to minimize fees. Disclosed fees have not changed much in this decade but the task of determining how much a mutual fund costs has become more complicated. Hidden fees appear to have grown and it appears that the mutual fund industry is not competing very hard over fees. The identical portfolio offered by different fund families can sell for very different fees. This does not necessarily mean that mutual fund fees are too high since investors may value the services offered by a high-fee mutual fund. It is certainly possible that if it were easier for an investor to determine the total price of a mutual fund, investors would demand more services and fees would rise. But given the restrictions on this market and given industry practices, it is difficult to believe that investors would settle for the current mixture of fees in a more freely competitive market that has competes over standardized quantitative measures of risk. It is clear from the evidence that the current mixture of fees have little relationship with the quality of the fund when "quality" is defined as a better return for the risk taken. With this definition of quality, most economists view the competition in this market as imperfect and believe that the competition does not appear to be improving.

I believe the academic and non-academic evidence supports the following facts which are the basis of my conclusions on competition:

• Total expenses paid by investors have not fallen over the past decade. Expense ratios have been stable over the past decade while other expenses (softdollars and distribution expenses) appear to have increased.
• There is no relationship between the level of expense ratios and risk-adjusted performance net of fees except that large expense ratios substantially reduce performance.
• There is no evidence that managed mutual funds have performed better than funds that simply try to match an index or a combination of indices.
• There is little evidence of persistence of good performance, there is stronger evidence of persistence of poor performance.
• Good performance is rewarded by investors, poor performance is ignored except when the poor performance is extreme.
• The advertised investment philosophy or "style" of a mutual fund is informative about lower risk styles ("income" and "balanced) but the advertised style is does not describe the portfolio practices of higher risk styles ("growth" and "growth and income").
• Mutual funds tend to buy and sell the same securities at the same time, exhibiting "herding" behavior.

There is no litmus test or standard yardstick to measure the level of competition. From the view of an economist, all industries are imperfectly competitive. Taken as a whole, these facts certainly suggest that the mutual fund market is imperfectly competitive. The quality of the product does not seem to be related to the price of the product and the quality-price relationship does not seem to be improving in spite of the large number of suppliers in the market. I believe there are several reasons why investors do not improve the level of competition.

First, it is very difficult to determine the quality of the "product" being sold. Securities are difficult to price because risk is difficult to judge. The nature of this product is similar to health care services where the consumer has a very poor idea of the quality of a doctor’s service. No change in the law or in industry practices will alter this fact. There is no measure or group of measures that will make mutual quality easy to observe. If such a measure were available, an economically rational investor would care only about performance and not about fees. But since performance is very difficult to predict and fees are much easier – or should be much easier- to predict, a rational investor will care about fees.

Second, the disclosure required of mutual funds is inadequate. Some expenses are hidden, such as "softdollars". There is little standardization in the "classes" of funds. Most importantly there is little quantitative risk disclosure. Quantitative measures of risk can greatly aid in judging the quality of a mutual fund but each investor must develop his or her own standard for every measure.

Third, the law interferes with the market for financial consultants by restricting the market for financial advice. Mutual fund sales fees have not be deregulated like brokerage commissions on common stock. A mutual fund investor, unlike an institutional investor, is not free to hire a consultant who has access to every fund and is responsible only to the investor. By law the financial consultant who recommends "load’ mutual funds must be paid at least partly by the fund. If this were done on the institutional side of the investment business, this practice would be considered unethical by many institutional investors.

I conclude that from the investor’s point of view, mutual fund expense ratios are stable, softdollars are hidden and distribution expenses are complicated. The investor has little incentive to carefully examine the fees and compare funds. From an economist’s view, fees that are hidden and complicated are not likely to fall - especially when there is a restricted market for advice.

Mutual Fund Fees

In every successful industry, from calculators to VCRs, competition has dramatically reduced the prices that consumers pay. The success of the mutual fund industry has not produced the same benefits. Disclosed fees paid by mutual fund investors have not fallen and there are reasons to believe that the total fees paid by mutual fund investors have risen. The fees paid by investors are broken into three parts. The first is the expense ratio which is fully disclosed in the prospectus of the fund. The second is softdollars which are rebates from brokers who charge the fund a commission that covers more than the cost of buying and selling securities. These fees are not disclosed. The third are the expenses paid for buying and selling the shares of the fund. This is the "load" of the fund and the "12b-1" expense of the fund. These are disclosed but "12b-1" fees require investors to estimate the future return of the fund.

The evidence is that disclosed fees are stable. The table below shows the average fee for equity funds existing at three dates: June 1987, June 1992, and June 1998. I did not require that the fund last the whole period, from 1987 to 1998, nor did I require that the fund last even one month after the each date. I simply averaged all expense ratios that were reported by Morningstar for each of the dates. The numbers do not include the load but include "12b-1" fees.

The "asset-weighted" average gives more weight to the expense ratio the more assets that a fund has. It represents the expense ratio for the average dollar in the industry. The "equal-weighted" expense ratio simply adds all the expense ratios together and divides by the number of funds. It represents the expense ratio for the average fund and does not adjust for the fact that small funds typically charge more. This is an example of what many academic researchers, and non-academic researchers, including Morningstar, have found. (For a recent study see W. Dellva and G. Olson "The Relationship between Mutual Fund Fees and Expenses and Their Effects on Performance" published in the Financial Review in February 1998.) The equal-weighted expense ratio is rising over time and the "asset-weighted" average is comparatively stable, especially in the past five years. The industry has, on average, consistently charged mutual fund investors 1% of assets, that is, one cent per dollar invested. The range of expenses is large as shown by the equal-weighted average which has consistently risen over the past twelve years.

Assuming that these fees represent the total payments by investors, the stable expense ratios support one of two conclusions. First, the expense ratio fully reflects all costs of doing business and funds are vigilant in controlling costs. The stability reflects the success of the industry in controlling costs. Second, expense ratios do not reflect the costs of doing business and funds are not competing over fees. The stability reflects the lack of searching by consumers across the market.

Expense ratios do not represent total payments. They do not include softdollars and they only partially include "distribution expenses". I deal with these expenses individually below.

Softdollars

The total payments of investors to mutual funds are not observable since mutual funds do not disclose their "softdollars". In "The Inspection Report on the Soft Dollar Practices of Broker-Dealers, Investment Advisers and Mutual Funds", the Securities and Exchange Commission has defined softdollars as

"arrangements under which products or services other than execution of securities transactions are obtained by an adviser from or through a broker-dealer in exchange for the direction by the adviser of client brokerage transactions to the broker-dealer."

The Commission found that the through its inspection program that 280 investment advisers paid $274 million in softdollar payments for third-party research from January through October 1996. The $274 million was not included in the expense ratios of the mutual funds who charged the expenses. The investors of these mutual funds have no idea how much their mutual fund spent through this mechanism. However, the net returns of the mutual funds are lower because of this payment since the price of the securities bought was increased by this amount.

Other evidence indicates that commissions add about 20% to expense ratios. M. Livingston and E. O’Neal (Journal of Financial Research, Summer 1996) using a large sample of mutual funds found that median brokerage commissions are .21% of net assets and the commission levels are smaller for larger funds. The average brokerage commission measured as a percentage of net assets exceeds the typical execution-only commission for large institutional traders. They found that expense ratios are positively correlated with commissions per trade, indicating that mutual fund managers do not reduce fees when they use softdollars.

Further, the amount of softdollar payments appears to have grown over time. A. Hopfner, Plan Sponsor, 1995 (pages 60-61) estimates that softdollar payments were $800 million in 1994. The Commission estimates that softdollars were around $1 billion in 1996.

It is plausible that softdollars have some benefit. S. Horan and D. Bruce Johnsen, in a paper titled "The Soft Dollar Debate: Agency Theory and Evidence", a working paper distributed by George Mason Law School, argue that softdollar use can potentially help a money manager obtain exclusive access to valuable research. They provide indirect evidence that is consistent with this conjecture but do not have actual softdollar expenditures. Their evidence is from pension fund money managers and not from mutual funds.

Since softdollars are not reported there is no way that economists can test whether they are beneficial and there is no way that mutual fund investors can make the determination for themselves. The Commission did not recommend that mutual funds report their use of softdollars.

It is difficult to determine from publicly-available data, such as the above data in Table 2, whether distribution expenses have increased or decreased over the past five years. Each part of the distribution expense depends on the action of investors. The "number of funds" is really a misnomer. A single portfolio can have many classes of shares. Each of these share classes is treated as a different fund. Each class of shares involves a different combination of front-end fees, 12b-1 fee and deferred load. Investors can choose how to pay the broker that sold the fund by selecting from the menu of classes. It is difficult to determine which menu item is most expensive since it depends on the behavior of the investor. Investors cannot determine the exact cost of each class of shares since the 12b-1 fee depends on the future performance of the fund.

Investors cannot negotiate the contract individually with a broker who sells the fund. The Investment Company Act of 1940 (section 22d) requires a mutual fund to set the same fee for all brokers who sell the fund. From 1940 to 1981 the SEC, empowered by this act to regulate funds, allowed only one type of broker compensation - the "front-end load". If an investor wanted the services of a broker, he paid a percentage of the money invested at the time of the investment and nothing more. While brokers were forced to charge uniform fees, it was easy to compare the services of brokers with their fees since the fee was simple. In 1981 the SEC passed rule "12b-1" allowing funds to deduct a percentage of the fund's assets each year to pay the broker an annual fee. The "12b-1" fee continues for as long as an investor is in the fund. With this fee, investors must guess a broker's ultimate compensation since it depends on the future value of the assets and the length of time an investor spends in the fund - and the fund may increase the fee up to a limit. Over the past five years funds have given investors a menu of choices between paying a front-end load, a deferred load or a "12b-1" fee. If investors want all their money invested immediately and avoid the front-end fee, they must guess at broker compensation since it will depend largely on 12b-1 fees which are a percentage of net assets.

As a result of these regulations, there is a striking - and predictable - difference between the regulated selling of mutual fund shares and the competitive pricing of broker services in the retail market for common stock. An investor who buys common stock can easily compare the prices of broker services which range from the simple execution of trades performed by discount brokers to stock selection by a full service broker based on specialized research. The same investor who decides to own common stock by buying shares in a mutual fund is confronted by identical prices for broker services no matter which broker is selected or what service is provided. If the fund has a 12b-1 fee the investor must compare the expected compensation. Thus, it is much less worthwhile for the investor in a mutual fund to compare services offered by brokers with the fee than it is for the investor in common stock. Since one motive for buying mutual fund shares is to pay someone else to investigate the risks and return of securities, it is not surprising that investors who need the services of brokers often ignore fees - allowing funds to raise their charges.

The economic evidence on fees supports these conclusions. In 1990 I published (with Robert Zweig) a study of mutual fund distribution expenses titled, "An Economic Analysis of The Cost and Benefits of SEC Rule 12b-1". The study concluded

• Expense ratios are higher for 12b-1 funds than non 12b-1 funds.
• Expense ratios for new 12b-1 funds were higher than expense ratios for older 12b-1 funds
• For new funds there is no evidence of substitution between the front-end load and the 12b-1 fee.
• Net returns (returns minus the expense ratio and portfolio transactions costs) were lower for 12b-1 funds than non 12b-1 funds.
• The adoption of a 12b-1 plan has no effect on the percentage change in total assets of the fund.

Since this study was finished, there have been substantial regulatory changes but recent academic studies have confirmed many of the conclusions. R. Mcleod and D. Malhotra, in "A Reexamination of Mutual Fund Expense Ratios" published in the Journal of Financial Research, summer 1994 conclude "The findings are consistent with previous studies that find charges are a dead-weight cost borne by shareholders. However, we show that this cost increases over time". In another article, "The Effect of Rule 12b-1 on Bond Expense Ratios" published in the Journal of Economics and Finance, Spring 1996, the same authors conclude

Using a model specific to bond funds, as opposed to the generic models used in previous studies on equity funds, the analysis confirms that the 12b-1 fee is an additional cost borne by shareholders of the fund without any additional benefit. However, this cost as a percent o the net asset value of the fund has decreased from 1991 through 1994.

W. Dellva and G. Olson in a study titled "The Relationship between Mutual Fund Fees and Expenses and Their Effects on Performance" published in the Financial Review in February 1998 conclude

In this paper we find that, on average, 12b-1 fees, deferred sales charges and redemption fees increase expenses whereas funds with front-end loads generally have lower expenses.

Summary

Given the nature of the industry regulation and practices it is difficult to determine whether total fees paid by investors have increase or decreased over the past five years. Most studies find that expenses have little effect on returns. It is worth noting that none of these studies examined the services of a mutual fund. Investors may be willing to pay higher fees and accept a lower return if they value the services provided by a mutual fund. It is difficult to determine this because of the difficulty of determining the exact fees that investors are paying. From the investor’s point of view, mutual fund expense ratios are stable, softdollars are hidden and distribution expenses are complicated. From an economist’s view, fees that are hidden and complicated are not likely to fall.

Mutual Fund Performance

Regardless of how fees are charged and disclosed, they do alter the net return of the fund. Higher fees will mean lower returns if everything else is equal. Of course, nothing else is equal and this section briefly describes the recent evidence on mutual fund performance.

The first study of mutual fund performance was in 1969 by Micheal Jensen who concluded that there was no evidence that mutual funds earned greater than a risk -adjusted return. Since the Jensen study there have been over 300 performance studies, most of which have agreed with the Jensen conclusion. Here is a list of the recent studies and a brief synopsis of their conclusions;

• Recently, Hendricks, Patel and Zechkhauser, (Journal of Finance, 1993) concluded that short-term mutual fund performance persists with the poor funds performing poorly in the future and good funds performing well in the future.
• B. Malkiel (Journal of Finance, 1995) using all funds that had at least one year of existence between 1971-1991 finds that no strategy, including buying the Forbes recommended funds, outperforms a passive strategy.
• M. Gruber (Journal of Finance, 1996) finds that buying the top decile firms and selling the low decile firms gives 28 basis in performance per year over a risk-adjusted return for a group of funds that began in 1984 (dead funds included in sample and ending in 1994.
• Carhart(Journal of Finance,1997) examines all mutual funds that had at least one month of existence between 1962-1993 (no survivorship bias) finds that common factors explains the persistence in the performance of top mutual funds. He finds that expense ratios, turnover and load fees (12b-1 charges) are significantly and negatively related to performance

All these studies find that risk-adjustment and survivorship bias is a critical issue in measuring performance. Not adjusting for survivorship bias adds as much as 1.4% per year to estimated performance over a risk-adjusted return for funds that survive ten years. (Estimates for one year bias is about 20 basis points).

Taken as a whole, it is reasonable to conclude that

• There is no evidence that managed mutual funds have performed better than funds that simply try to match an index or a combination of indices.
• There is no relationship between expense ratios and performance except that large expense ratios substantially reduce performance.
• There is little evidence of persistence of good performance, there is stronger evidence of persistence of poor performance.

Competition in Mutual Fund Portfolio Management

Mutual funds compete by offering investors different investment "styles". Mutual fund offer investors a wide variety of different types of portfolios such as "income", "balanced", and "growth". The terms are used to describe the portfolio management process which involves picking securities and monitoring their performance. Recent studies have found that this competition is imperfect. Some styles are descriptive and some are not. The range of mutual fund descriptions much larger than the actual practice of mutual funds which tend to buy and sell the same securities at the same time.

S. Brown and W. Goetzman in "Mutual Fund Styles" published in the Journal of Financial Economics, (March 1997) found evidence that the descriptions for the "styles" with lower risk, "income" funds and "balanced" funds, are accurate descriptions of the actual investment practices of the mutual fund. The portfolio advertised is close to the portfolio that investors get. However, the descriptions of higher risk styles, such as "growth and income", "growth", and "aggressive growth", do not represent the portfolio practices of the mutual fund. For example, the mutual fund "Mutual Fund Shares", that is run by portfolio manager Micheal Price, is a "value" fund that follows the "value" strategy first advocated by Benjamin Graham. This fund is classified as a "growth and income" fund in the prospectus.

There is evidence that mutual funds tend to buy and sell the same securities at the same time. This is called "herding" behavior. M. Grinblatt, S Titman and R. Wermers, in "Momentum Investment Strategies, Portfolio Performance and Herding: A Study of Mutual fund Behavior" published in the American Economic Review in December 1995, found that 77 percent of mutual funds were "momentum" investors, buying stocks that were past winners, but most did not sell past losers. On average funds that invested in momentum realized significantly better performance than other funds (neither group outperformed a risk-adjusted return). The authors found evidence that funds tended to buy and sell the same stocks at the same time. E. Falkenstein in "Preferences for Stock Characteristics as Revealed by Mutual Fund Portfolio Holdings" published in the Journal of finance March 1996, found that mutual funds have a significant preference for stocks with high visibility and low transactions costs. Mutual funds are adverse to stocks that do not move with the market.

There is evidence that mutual funds alter their portfolios based on the fund’s most recent performance. If returns in the first part of the year are good, the fund will lower the riskiness of the portfolio in the second part of the year. If the returns of the first part of the year are poor, the fund will increase the riskiness of the fund in the second part hoping to earn a good return. There are two independent studies that show this. J. Chevalier and G. Ellison in "Risk Taking by Mutual Funds as a Response to Investors" published in the Journal of Political Economy, December 1977, and K. Brown, C. Harlow and L. Starks, "Of Tournaments and Temptations: An Analysis of Managerial Incentives in the Mutual Fund Industry", Journal of Finance, March 1996. The last study shows that this effect became stronger as industry growth and investor awareness of fund performance increased over time.

There is some discipline by investors of managers who perform poorly. A. Khorana in "Top Management Turnover: An Empirical Investigation of Mutual Fund Managers", (Journal of Financial Economics, March 1996) found that manager replacement is positively related to the growth rate in the fund’s assets and its portfolio returns. Managers who leave a fund have higher portfolio turnover and higher expenses than managers who are not replaced. However, E. Sirri and P. Tufano in "Costly Search and Mutual Fund Flows" found that consumer base their fund purchase decisions on prior performance information asymmetrically. They invest disproportionately more in funds that performed very well than they pulled money out of funds that performed very poorly. Search costs seem to be an important determinant of fund flows. Fund flows are directly related to a fund’s marketing effort and media attention.

Summary

From the view of individual investors, a mutual fund is a consumer product where investors pay fees and get a return plus some services. I believe the academic and non-academic evidence supports the following facts which are the basis of my conclusions on competition:

On Mutual Fund Fees

• Total expenses paid by investors have not fallen over the past decade. Expense ratios have been stable over the past decade while other expenses (softdollars and distribution expenses) appear to have increased.
• There is no relationship between the level of expense ratios and risk-adjusted performance net of fees except that large expense ratios substantially reduce performance.
• Expense ratios are higher for 12b-1 funds than non 12b-1 funds.

On Fund Risk-Adjusted Performance

• There is no evidence that managed mutual funds have performed better than funds that simply try to match an index or a combination of indices.
• There is little evidence of persistence of good performance, there is stronger evidence of persistence of poor performance.

On Portfolio Management

• Good performance is rewarded by investors, poor performance is ignored except when the poor performance is extreme.
• The advertised investment philosophy or "style" of a mutual fund is informative about lower risk styles ("income" and "balanced) but the advertised style is does not describe the portfolio practices of higher risk styles ("growth" and "growth and income").
• Mutual funds tend to buy and sell the same securities at the same time, exhibiting "herding" behavior.
• Mutual Funds alter the risk of their portfolios depending on the return early in the year. Managers likely to end up as "losers" will increase fund risk and managers likely to be "winners" will decrease fund risk. This effect has grown stronger as industry growth and investor awareness has increased over time.
• Managers of poor performing funds are more likely to be replaced than managers of good performing funds where performance is measured by growth in assets and portfolio return. Departing managers have higher expense ratios and higher turnover than remaining managers.

attachment: example of return and risk table in a comment letter to sec

July 27, 1995

Mr. Jonathan Katz
Securities and Exchange Commission
450 Fifth Street
Washington, DC. 20549

Re: Concept Release; Request For Comments on Improving Descriptions of Risk by Mutual Funds and Other Investments Companies; File No. S7-10-95.

Dear Mr. Katz,

This letter is submitted in response to the Securities and Exchange Commission ("SEC") concept release and request for comment on improving descriptions of risk by mutual funds and other investment companies. (Release No.s 33-7153; 34-35546; IC-20974 ("Release")).

As a finance professor who has written about, advised and worked for various participants in the mutual fund industry, including the SEC, I fully support the SEC's overall objective of improving mutual fund risk disclosures and the SEC's more specific focus on disclosures concerning comparative fund risk levels. The Concept Release requests comments and suggestions on eight general areas covering a broad range of risk disclosure issues. I will comment on the goals of risk disclosure and the use of quantitative risk measures.

The Goals of Risk Disclosure

I believe that the primary goal of risk disclosure should be to help investors choose a mutual fund that meets the investors' tolerance for risk. Consistent information should be available for each mutual fund so that investors can match their investment objectives and investment horizon to the objectives and general risk of the mutual fund.

Investors differ widely in their goals, circumstances and ability to use investment information. I believe it is unlikely that the SEC can design risk disclosure, either narrative or non-narrative, that directly meets the needs of most investors. A more realistic goal is to require disclosure that both improves the market for information about risks of mutual funds and improves investors' awareness that risk is an important consideration in choosing a mutual fund.

I fully support the proposals to require mutual funds to provide a supplemental service in the form of a narrative brochure, or questionnaire that would help investors in determining their risk-tolerance, investment horizon and investment objectives. However, I do not believe that a general brochure is sufficient to address investor concerns about risk. Investors must have enough information to make comparisons between specific mutual funds on a consistent basis. Investors are in the same position as consumers buying products who need consistent, easily-understood measures of product's quality. The SEC should develop risk disclosures that allows investors to "shop" in the mutual fund market.

Quantitative Measures of Risk

Narrative risk disclosure is not sufficient for investors to achieve the goals described above. This is why quantitative measures were first developed over forty years ago. It is especially true today when many financial products are complex combinations of more basic securities. The failure of narrative disclosure for institutional investors, such as pension fund sponsors, has lead to the development of an extensive consulting industry using quantitative methods (e.g. the Frank Russell Corporation, SEI corporation, Richards and Tierney). Narrative disclosure is necessary to understand many of the activities of mutual funds but it is clearly not sufficient.

The difficulty with quantitative risk measures is that there are limitations of any single measure. If these limitations are not understood by the user, the risk measure can lead to poor investment decisions. I believe that the SEC should not support the use of any single measure but should develop a table of measures. This is analogous to the requirement by the Federal Drug Administration that all consumer food products carry a nutrition label describing various characteristics of the product. No single measure on a nutrition label is sufficient to describe the quality of the product and some argue that the table itself is inadequate. But nutrition labeling has lead to increased consumer awareness and has established a set of core facts that vendors of information about food products employ.

I have attached a "Return and Risk Table" which I believe generally address most of the concerns about quantitative risk measures that have been expressed by interested parties. I have also prepared a short summary of the strength and weakness of each measure in the table.

Thank you for giving me the opportunity to comment on this Release.

Sincerely,

Charles Trzcinka

Historical Return and Risk Table for Fidelity Magellan as of June 30, 1995

I. Summary of Monthly Returns
	Most Recent 12 Months	Most Recent Five Years	Most Recent Ten Years
Annualized Average Return
Fidelity Magellan	20.80%	15.20%	19.00%
SP500	26.50%	12.80%	16.00%
Growth of a $1,000 Initial Investment
Fidelity Magellan	$1,190	$1,925	$4,863
SP500	$1,261	$1,769	$3,932
II. Risk of Monthly Returns
	Most Recent 12 Months	Most Recent Five Years	Most Recent Ten Years
Volatility (standard deviation)
Fidelity Magellan	5.10%	4.20%	5.10%
SP500	2.50%	3.30%	4.40%
Number of Months with Negative Returns
Fidelity Magellan	3	18	36
SP500	2	20	42
Fidelity Magellan's Responsiveness to the marke (beta)	1.15	1.09	1.07

III. Interest Rate Risk of Fidelity Magellan

interest rate sensitivity: Low
average maturity of bonds: No bonds in the portfolio

Distribution Expenses

An important part of mutual fund fees are the expenses used to "distribute", that is, sell the shares of a fund. These include the "front-end" load, the "deferred load" (or "back-end" load) and the "12b-1 fee". It is difficult to determine whether these fees have increased or decreased since the "12b-1" fee is charged each year, the "front-end" load is charged when the shareholder initially purchases the fund and the "back-end" load is charged when the shareholder sells the fund. Table 2 below shows that the averages for these numbers have changed in different directions over time.

Discussion of Return and Risk Table

The above table is intended to be a suggestion of how investor could be informed about portfolio risk with consistent and systematic measures. It uses actual data from Fidelity Magellan and the SP500. There is no measure that is critical to the table nor is the table an exhaustive list of measures that could be used.

The table has the important limitation of all risk disclosure, narrative and quantitative, namely, the portfolio manager can easily change the securities in the portfolio and easily change the risk of the portfolio. This is why I labeled the table "Historical" risk measures. These are all "backward-looking" measures but since a portfolio manager can easily alter the risk exposures of a portfolio I believe that there is no meaningful distinction between "backward" measures and the "forward" measures of some vendors. So-called "forward measures" simply apply statistical and mathematical techniques to "backward" data.

Each of the risk measure in the table has strengths and weaknesses. Taken together, I believe that they provide an accurate picture of how risky the portfolio was over the horizon specified (e.g. 12 months, 5 years, 10 years).

Benchmark Portfolio

I believe it is important to provide a basis of comparison for any return and risk measure. I chose the SP500 but there are several benchmarks that will be adequate. The SEC should allow mutual fund to choose from a list of approved benchmarks but all benchmarks should represent the market for all securities. The benchmarks should not be specific to any sector or investment style. Information on risk is already complicated for many investors and adding the additional complication of sectors or styles will greatly reduce the effectiveness of the table in achieving the goals outlined above. Moreover, there is no consensus among investment professionals, academics and regulators on the definitions of sectors and style. Finally, the "beta" of the portfolio should be computed relative to the benchmark so that the investor has a consistent basis to judge the measures.

Volatility (Standard Deviation)

The standard deviation of monthly returns represents the dispersion of the returns. It captures all sources of risk that cause prices to change such as market expectations, illiquidity, interest rate sensitivity, inflation, etc. The weakness is that it is difficult to interpret and it is affected by extremely low or high returns.

Number of Negative Returns

Simply counting the number of returns below zero gives the investor a straightforward measure of how often his portfolio can fall in value. It increases the awareness of investors that risk means they can lose money. Technically, this is a "non-parametric" measure that does not involve the assumptions of the standard deviation. The weakness of this measure is that it discards information such as the magnitude of the negative returns.

Responsiveness to the Market (Beta)

This measure allows the investor to judge how risky the portfolio is relative to the overall market. The strength is that this measure is stable overtime for mutual fund portfolios and has an simple interpretation (e.g. "a beta of 2.0 means when the market rises by 10% the portfolio will rise by 20%). The weakness of this measure is that there are other factors such as an industry factor, leverage, etc. (The BARRA corporation sells 68 factors for domestic equities.)

Interest Rate Risk

The Interest Rate Sensitivity category is intended to be the same as the "duration" of a portfolio (e.g. Macaulay's duration) appropriately adjusted for any derivatives. All the investor cares about is that this portfolio has low, medium and high sensitivity. I would define "low" as the sensitivity that the value of a portfolio of short term bonds with no derivatives has to interest rate changes. I define "medium" as the sensitivity of a medium term bond portfolio (with no derivatives) to interest rate changes and "high" as the sensitivity of a long term bond portfolio (with no derivatives). Most equity portfolios do not change in value with interest rates in any predictable fashion and the appropriate response may be "insensitive" rather than "low".

Average Maturity of bond portfolios is a common measure of interest rate risk. Investors should see both the average maturity and interest rate sensitivity since the short term bond portfolios of some fund groups are very sensitive to interest rates because of derivatives. Some short term bond portfolios are actually more sensitive to interest rates than a portfolio of 30 year bonds.

Omitted Risk Measures

There are several risk measures that are deleted from the table that are commonly found in the institutional investment market. First, I excluded "style" comparisons for the reasons cited above. Second, there is no measure of liquidity risk (such as percent of illiquid stocks). Clearly this risk is greater for some portfolios (e. g. high-yield bond portfolios) than others. I did not include these measure because the standard deviation will capture any illiquidity price risk and it is difficult to define "illiquidity" precisely. Third, I omitted any measure of risk caused by macroeconomic factors such as inflation and unemployment. Again, the standard deviation will be larger to the extent that these risks are large and research in financial economics demonstrates that these risks are small relative to the movement of the market.

---