W. Hardy Callcott

January 30, 2004

By Email to rule-comments@sec.gov

Re: File No. S7-29-03, Request for Comments on Measures to Improve Disclosure of Mutual Fund Transaction Costs

To the Commissioners:

I am pleased to submit this comment letter on the important issue of transaction cost disclosure for mutual funds. I was previously the Commission's Assistant General Counsel for Market Regulation, and later I was General Counsel for a major national broker-dealer which was a pioneer in the mutual fund supermarket area and distributed one of the nation's largest families of proprietary mutual funds. I submit this comment letter solely on my own behalf, and not on behalf of any client, my law firm, or any partners or associates at my law firm.

I concur with the Commission's assessment that mutual fund investors deserve useful information about a fund's transaction costs. While some members of the Commission staff have had a myopic focus on the commissions paid by mutual funds, in fact the other elements of transaction costs (spreads, market impact and opportunity costs) typically are much larger and have a correspondingly greater impact on investors' potential returns. A requirement to disclose only commission costs would in my view be affirmatively misleading, because it would cover such a small portion of total transaction costs. Moreover, as the concept release itself recognizes, a measure of commission costs alone could be gamed by simply moving away from agency trades toward principal trades, or by choosing execution venues that may give up execution quality in return for lower commissions - results that would cause net harm to a fund's investors. Nor are commissions a useful proxy for overall transaction costs; depending on the markets and the securities involved, the percentage of total transaction costs represented by commissions may be highly variable.

More generally, the focus of the concept release, on domestic equity trades, is too narrow. For a transaction cost disclosure to be meaningful, it must cover the universe of mutual funds available to US investors. Transaction cost disclosure must apply to money market funds, to bond funds (including both investment grade and high-yield debt), to blended funds (with both bond and equity investments), to international funds, and to global funds (with both US and international investments). It must apply to passively managed index funds as well as actively managed funds and funds managed with tax efficiency objectives. The disclosure must not only be measure the cost of conventional limit and market orders, but also of volume-weighted-average-price (VWAP) orders, market-on-close (MOC) orders, basket trading, stop-loss orders and other modern methods of portfolio management, including orders that are hedged in the options or futures markets and orders that arbitrage between equities and derivatives markets.

For many securities (notably many international equities and both US and foreign debt securities), there simply is no continuous two-sided firm-quote data available about the relevant securities.1 Most if not all of the proposed methods of quantifying spread costs, market impact costs, and opportunity costs are useless if there is no continuous quotation data (or if the available quote data consists merely of non-firm "invitations to deal" that are often far from actual transaction prices). For many order types, such as VWAP or market-on-close trades (or stop orders), the concepts of trade decision time and trade execution time at best are difficult to apply. Even in the US equities markets, there is rarely reliable, firm depth-of-book quote data available beyond a thin National Best Bid or Offer (NBBO) which is virtually irrelevant to institutional-sized orders.

In short, there is no "silver bullet" that will allow an easy quantification of transaction costs in a way that would be comparable among all the different types mutual funds. If the Commission were going to go down that path, it would have to develop and constantly modify specific rules for every possible asset class (Argentine high-yield corporate debt, Slovakian sovereign debt, Turkish equities, Chinese/Hong Kong dual-listed securities, etc.) and, within each market, for each order type. This is a classic example of a game that is not worth the candle. Even where the data is most available (for example, the market for US large-cap equities), different experts will assess differently the costs of an order. The market impact of an order may look very different two minutes after execution, five minutes or two days (not to mention trades executed and reported before or after market hours) - and there is no scholarly or practical consensus about the appropriate time frame in which to make the analysis. A transaction cost measure that may be appropriate for an index fund or a tax-managed fund may not be appropriate for an actively managed fund. An order sent to an electronic order-execution venue is not necessarily comparable to one sent to a floor-based specialist or market-maker environment. What constitutes a potentially market-moving order varies from security to security and even from hour to hour and day to day in a single security. In sum, transaction cost measurement is an art, not a science - and pretending that it is a quantifiable science would mislead investors, not enlighten them. A whole industry exists in the US to assist institutional investors in measuring transaction costs, and no two players in this industry come up with the same answers.2

Just because transaction costs are difficult to measure does not mean they do not have a real and important impact on investors - they do. However, in this instance, the Commission should look means of disclosure other than quantification. First, mutual funds should be required to include in their prospectuses textual disclosure about their transaction costs, the implications for transaction costs of their portfolio strategies, how they themselves attempt to measure and control transaction costs (including information about the experts they consult and how often they consult them), and how their boards monitor and evaluate those costs. As part of its examination program, the Commission should actively review mutual funds and sanction advisors which simply utilize boiler-plate disclosures and omit transaction cost information relevant to investors.

Further, among funds with similar goals and investments, I believe portfolio turnover is the most useful proxy for transaction costs. Although there is not an absolute one-to-one relationship between portfolio turnover and transaction cost, portfolio turnover does give investors information they can use to compare similar funds specializing in similar securities and markets. The Commission should consider requiring more prominent disclosure of portfolio turnover ratios, for example, in advertising containing historical performance data.

The Commission should endeavor to "first do no harm". A combination of enhanced textual disclosure and more prominent portfolio turnover disclosure is far preferable to the two alternatives suggested in the concept release - disclosure only of commissions, or an attempt to establish rules quantifying all types of transaction costs for all types of assets and orders for all funds dealing in all markets. Both commission-only disclosure and total-quantification disclosure would, because of their inherent severe limitations, in the end serve to confuse and mislead investors. I appreciate the opportunity to comment on this important issue and would be happy to discuss it further with members of the Commission or its staff.

Sincerely yours,

W. Hardy Callcott

1 The Commission's recent enforcement action against FT Interactive Data, Investment Advisers Act Rel. No. 2201 (Dec. 11, 2003), illustrates the shortcomings of pricing data in the market for infrequently traded US bonds - a market that is well more transparent than many foreign markets.
2 Nor do I believe it would be useful to require mutual funds to quantify and disclose transaction costs using their own proprietary methodologies. Such disclosure would not be comparable among different funds and different fund families, and investors who made the inevitable attempts to compare the data would be misled. The pressure to report low transaction costs would discourage candid evaluation of transaction costs. And allowing proprietary methodologies would not change the fact that transaction costs are not susceptible to precise measurement, and therefore that attempted quantifications have the inherent potential to be misleading.