DOC PREVIEW
Chicago Booth BUSF 35150 - Persistence in Mutual Fund Performance

This preview shows page 1-2-23-24 out of 24 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 24 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 24 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 24 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 24 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 24 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

Persistence in Mutual Fund Performance: Analysis of Holdings Returns Samuel A. Kruger June 8, 2007 Term paper for: Research Projects: Finance (35908) Spring Quarter, 2007 University of Chicago Graduate School of Business Professor Eugene F. Fama1 I. Introduction Mutual fund performance is a rich topic for academic research. The very existence of the mutual fund industry with its high expenses and transaction costs is a puzzle to the efficient markets hypothesis. If markets are efficient, why are so many resources invested in stock picking? The conventional wisdom is that at least some mutual fund managers have stock-picking skills. Investors frequently look to past returns and historical ratings such as Morningstar for guidance on which funds will perform well in the future. Successful fund managers are glorified in the media and turned to for investment advice for years to come. The academic literature is much more mixed. It is generally agreed that there is some persistence in mutual fund performance.1 However, the source of this persistence remains a topic of debate. For example, Carhart (1997) attributes almost all persistence in mutual fund performance to four-factor loadings, expenses, and transaction costs. On the other hand, Wermers, Yao, and Zhao (2007) finds that “good” managers pick better-performing stocks even after controlling for style characteristics. By examining the future performance of stocks held by “good” and “bad” mutual funds, I conclude that fund managers do not have stock picking skills. This paper adds to the literature by applying the performance-sorted portfolio methodology of Carhart (1997) to a dataset that includes returns on the underlying holdings of mutual funds as well as the returns of the funds themselves. Section II provides a brief review of the relevant literature. Section III describes my dataset. Section IV introduces my analytical approach. Section V discusses the results of my analysis. Section VI concludes. 1 See Carhart (1997).2 II. Literature Review Mark Carhart’s 1997 Journal of Finance article, “On Persistence in Mutual Fund Performance,” is the primary motivation for my analysis. Carhart examines the returns of mutual funds from 1962 to 1993 to look for evidence of performance persistence. Carhart’s primary analytical technique was to form performance decile portfolios of mutual funds on January 1 of each year based on returns over the past year. The portfolios are then held for one year and monitored for any abnormal performance. If performance is persistent, funds that performed well in the past should perform well in the future, and the top decile portfolios should outperform the other portfolios. Carhart finds that past winners do outperform past losers. However, most of this persistence is explained by a four-factor model including factor-mimicking portfolios for the market return, size, book-to-market, and one-year momentum. Momentum is the biggest explanation of the persistence. The remaining persistence is mainly explained by fund expenses and transaction costs, which are higher in the lower performance deciles. Of the 8% difference in annual returns that Carhart finds between the top and bottom deciles, 4.6% is explained by four-factor loadings, 0.7% is explained by expense differences, and 1.0% is explained by transaction cost differences. This leaves an unexplained return spread of 1.7%, almost all of which is concentrated in the difference between the ninth and tenth deciles. In other words, Carhart finds some evidence that the very worst funds continue to underperform but finds no evidence of persistent skill in any of the other deciles.3 Wermers, Yao, and Zhao (2007) analyzes the returns of mutual fund holdings and comes to the opposite conclusion that good managers have significant stock-picking skill. Wermers et al identifies “good” managers based on past alpha estimates and then forms an equal weight portfolio that is long stocks held disproportionately by “good” managers and short stocks held disproportionately by “bad” managers. They find that this trading strategy results in significant alpha, which they interpret as evidence of stock-picking skill. A significant problem with this analysis is that forming portfolios based on the same model used to analyze returns could bias the results. Significant alpha could come from biases in the model. Additionally, employing equal-weight portfolios of stocks will overweight small stocks. I avoid these problems by sorting based on returns instead of alpha estimates and forming equal-weight portfolios of fund returns instead of stock returns. My analysis builds on Carhart (1997) by applying Carhart’s analytical approach to data on the returns of mutual fund holdings. If fund managers have stock-picking skills, the stocks they choose to buy should have high future returns. Analyzing holdings returns without the noise of expenses and transactions should be a cleaner way to identify stock-picking skill. Using this approach, I find no compelling evidence of stock-picking skill. Additionally, examining the holdings returns helps to identify the source of the persistent underperformance of the worst funds in Carhart (1997). Is this underperformance due to picking bad stocks, or is it the result of transaction costs, hidden expenses, or other unobserved actions? Given that I find no unexplained persistence in the holdings returns, I conclude that Carhart’s persistent underperformance must be due4 to other actions of the mutual fund, likely transaction costs from being forced to sell shares as investors pull out money. Finally, analyzing the holdings data is useful because any anomaly can be translated directly into a trading strategy. Persistence in mutual fund returns is interesting but difficult to exploit. I can purchase the winning funds, but I cannot short the losers. On the other hand, if return persistence is present in fund holdings, I can exploit this with a zero-investment portfolio by buying stocks held by winning funds and shorting stocks held by losing funds. The fact that this is a tradable strategy also makes holdings returns a better test of market efficiency. III. Data My dataset covers domestic equity mutual funds from January 1980 to December 2006. My data includes information on both returns and holdings. As a result I am able to analyze both fund performance and the


View Full Document

Chicago Booth BUSF 35150 - Persistence in Mutual Fund Performance

Documents in this Course
CLONES

CLONES

8 pages

Load more
Download Persistence in Mutual Fund Performance
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Persistence in Mutual Fund Performance and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Persistence in Mutual Fund Performance 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?