Return on Your Research Investments
Mandatory information disclosure regulations may negatively affect product innovation in the mutual fund industry.
Research by Mary Margaret Myers
The Securities and Exchange Commission (SEC) requires mutual fund managers to report the asset holdings in their funds on at least a semiannual basis. These disclosures give competitors and investors access to otherwise proprietary information. Others can learn which assets the funds consider attractive and purchase those assets themselves. Such mandatory reporting limits the amount fund managers can reap from their research investments.
Several years ago, Vanguard Group stopped the practice of reporting on cash flowing into and out of each of its funds, because third parties were using the information in an attempt to "front run" Vanguard funds. Knowing the largest holdings of a given Vanguard fund and that the fund had experienced an infusion of cash, the investor might be able to identify securities in demand in the near future. Fidelity Investments subsequently took similar action.
Just how much competitive advantage do actively managed mutual funds potentially sacrifice in these disclosures? In recent research, Mary Margaret Myers, an assistant professor of accounting at the University of Chicago Graduate School of Business, along with co-authors James Poterba from MIT, Douglas Shackelford from the University of North Carolina, and John Shoven from Stanford University, tackled that question.
In her research paper, "Copycat Funds: Information Disclosure Regulation and the Returns to Active Management in the Mutual Fund Industry," Myers demonstrates that hypothetical copycat mutual funds - funds that buy the same assets as actively managed funds - earn returns statistically equivalent to or higher than the funds they mimic, when those returns are adjusted for expenses.
Copycat funds are at a disadvantage vis-à-vis actively managed funds, because copycat fund managers can trade on new information only as often as it's disclosed - as infrequently as twice yearly in the case of some actively managed funds. In other words, they miss the opportunity to invest in assets that managers favor between disclosure dates. By contrast, actively managed fund managers can act immediately on the results of their research.
But actively managed funds incur research expenses in generating that information, while copycats have virtually no research expenses.
"If research is valuable, copycat funds should earn lower returns than their actively managed counterparts," says Myers.
However, the crucial question is how net-of-expenses returns on actively managed funds compare with those of copycat funds. Myers investigated this question by creating hypothetical copycat funds and tracking their returns vis-à-vis the actively managed funds she mimicked from 1992 to 1999.
Getting on Track
Myers's investigation began with the assumption that the hypothetical copycat funds could track only the asset selections of the comparable actively managed, or "primitive" funds, twice yearly, at the SEC-mandated disclosure dates of 60 days following the end of the fund's fiscal year, and 60 days after the end of its fiscal second quarter. That assumption is considered conservative. More frequent voluntary disclosure would result in even closer tracking of funds.
To provide an example of this conservative assumption, a fund with a calendar-year end would have to report at the end of February and again at the end of August. The copycat fund's assets would be purchased at those disclosure dates, and the copycat's performance would be compared over the next six months against the underlying primitive fund.
Assuming the potential net return advantage of copycat funds would be greatest versus primitive funds with high expense ratios, Myers examined a sample of large equity funds with high expenses, but repeated the analysis on a broader sample of funds.
The first, high-expense fund sample was made up of 20 equity funds whose large size and investor fees made them most susceptible to copycats. The second sample was comprised of the 100 largest funds (by net asset value) from a broader range of equity funds.
Taking a Closer Look
Myers examined the first sample and their hypothetical copycats, finding the mean return on the high-expense ratio primitive funds greater than the mean return on the copycat funds. But only in the sixth month was the differential significant. Not surprisingly, the gap in returns between primitive and the hypothetical copycat funds widened with the passing months, reflecting the copycats' declining ability to track their counterparts.
However, when the fund returns were adjusted for estimated monthly expenses, the copycat funds outperformed the primitive funds she mimicked in four of the six months, with the spread remaining insignificantly different from zero except in the fourth month. The basic finding: Copycat returns and primitive fund returns are not statistically different.
In addition to monthly returns, buy-and-hold returns were examined, finding much the same thing: The difference between primitive fund returns and copycat fund returns at the end of the six-month holding period was not significantly different.
The suggestion from the high-expense ratio sample is that the hypo-thetical copycat funds can track actively managed funds for up to six months after portfolio disclosures. What's more, the differences in returns are statistically insignificant whether or not the actively managed funds' expenses are deducted. That would also suggest that copycats should enjoy substantially higher returns once the additional costs of active management are factored in. Those costs, such as taxes and brokerage fees, were not considered in Myers's analysis.
Analyzing the second, broader sample, Myers found that relaxing the restrictions did little to change the findings. Looking at monthly returns, the primitive funds out-returned their corresponding copycat funds in four of six months, but not by significant margins.
However, when annual expenses were deducted from the returns or when the sample was restricted to the observations for which asset allocation is measured most precisely, the hypothetical copycats outperformed the primitive funds in five of the six months, but the difference was statistically significant only in one of the six months.
The buy-and-hold results produced much the same findings. Myers found that the difference between the primitive funds' buy-and-hold returns before expenses and the copycat funds' buy-and-hold returns before expenses at the conclusion of the six-month holding period was not significant. However, the net-of-expense, buy-and-hold returns for the copycat funds was significantly higher than the primitive funds over a six-month period.
In both samples, differences in pre-expense returns between copycats and actively managed funds declined with the percentage of fund assets that could be identified by the University of Chicago's Center for Research in Security Prices (CRSP). In short, the more assets that are known, the less able a fund manager is to earn superior results.
The End Game
Using a model designed to assess the factors determining the hypothetical copycat fund's precision in tracking the primitive fund, Myers found as expected that the absolute difference in returns decreased with the percentage of known holdings, and increased with the amount of turnover in both samples.
Her findings support the idea that funds with more assets invested in identifiable securities are easier to track, and funds that turn over their portfolios are harder to track.
"It makes sense that copycats are better able to mimic the performance of funds that trade infrequently, because a copycat would exactly match the performance of a fund that never traded," says Myers.
The broader sample findings also indicated that higher expenses and longer management tenure make mutual funds harder to track, while market capitalization of the funds' holdings make them easier to track.
Myers concludes that actively managed funds outperform copycat funds when returns before expenses are examined. But the returns net of expenses - the returns available to mutual fund investors - favor the hypothetical copycat funds. The return differential favors the copycat funds more in a broader sample than it does in a high-expense ratio sample.
Myers suggested subsequent research might examine the extent to which actively managed fund managers can stymie copycat funds through the use of "window dressing."
Window dressing is a strategy in which managers sell particularly attractive issues just before disclosure, only to purchase them again immediately following disclosure. Myers notes that window dressing would make the most sense if the securities in question are those of large companies whose shares trade actively in liquid markets, and less sense in the case of smaller and less liquid securities, whose transaction costs would outweigh the benefits of window dressing.
Myers's findings also raise several questions about regulatory policy design. Among them is the question of whether mandatory disclosure is appropriate, and if so, how often.
If concerned that a fund manager will invest assets irresponsibly,
investors are likely to demand disclosure. But if they believe that
Given the potential complexity of information disclosure regulations, Myers concludes that future regulatory policy decisions should take into consideration not only the potential gain to investors from disclosures, but also the potential costs to firms associated with those disclosures. - J.S.
Mary Margaret Myers is Assistant Professor of Accounting at the University
of Chicago Graduate School of Business.