To examine whether and how media coverage influences the way professional investors trade, the authors use a measure of a mutual fund’s propensity to buy or sell stocks covered in the media. They identify a new manager/fund characteristic that predicts fund performance and provide novel evidence that limited attention detracts from that performance.
What’s Inside?
Recently, researchers have provided increasing evidence of the impact that mass media coverage of individual stocks can have on individual investors’ trading behavior. Media coverage can play a significant role in putting certain stocks on the radar screens of attention-constrained individual investors.
Professional investors have more access to institutional information-processing resources and timely newswire services than do retail investors. The authors examine whether mass media coverage of individual stocks has an impact on the trading and performance of professional investors—specifically, mutual fund managers.
How Is This Research Useful to Practitioners?
The authors hypothesize that the investment decisions of mutual fund managers, like those of individual investors, are subject to limited attention and are consequently influenced by attention-grabbing mass media coverage. Overall, their findings are consistent with this hypothesis.
Funds tend to buy more of those stocks that receive heavy media coverage (indicated by an above-median article count). Fund sells, however, are not significantly associated with higher levels of stock media coverage, possibly because of short sale constraints.
Most notably, there is a strong negative relationship between funds’ propensity to buy stocks covered by the media and their performance relative to other funds. Funds in the strongest-propensity decile underperform funds in the weakest decile by 1.1%–2.8% per year.
If fund investors are attracted to media-covered stocks—and consequently channel money into funds that hold such stocks—then fund managers’ tendency to buy media-covered stocks might simply represent a rational response to client preferences. But further analyses conducted by the authors indicate that media coverage of stocks bought by the funds has little impact on the inflow of money from clients.
Managers appear to be buying and selling based on media coverage regardless of its content. The authors use contemporaneous stock returns to infer coverage tone and find that positive-toned coverage and negative-toned coverage are both associated with more buys as well as with more sells.
Moreover, funds’ propensity to buy high-coverage stocks is persistent. Compared with funds classified as having the lowest propensity to buy, the highest-propensity funds continue to display a higher propensity three years after the classification period. Manager changes, however, are accompanied by significant reductions in propensity to buy, which is consistent with the notion that limited attention is a manager characteristic rather than a fund characteristic.
How Did the Authors Conduct This Research?
The authors obtain media coverage data from LexisNexis for all NYSE stocks and 500 randomly selected NASDAQ stocks for 1993 through 2002. They include articles published about the sample stocks in four newspapers in daily circulation, including the Wall Street Journal and the New York Times. The authors match the stocks to those in the CRSP stock database and to mutual fund holdings data.
The mutual fund sample is constructed by merging the CRSP Survivor-Bias-Free US Mutual Fund Database with the Thomson-Reuters Mutual Fund Holdings database (formerly known as the CDA/Spectrum database). The authors restrict their analysis to open-end domestic equity funds, excluding index funds. They also exclude funds that have less than half of their holdings in stocks belonging to their media-covered stock list.
They infer funds’ buys and sells from quarterly changes in fund holdings. They examine the relationship between stock media coverage and fund trading activity by estimating panel regressions of funds’ buys and sells—scaled by funds’ total net assets—on stocks’ lagged media coverage measures. The authors control for relevant firm characteristics, including firm size and book-to-market ratio.
The authors use cross-sectional regressions, in conjunction with shrinkage estimation, to define regression coefficient–based measures that represent funds’ propensity to buy/sell media-covered stocks. They then estimate panel regressions to examine the relationship between funds’ lagged propensities and five measures of their performance (i.e., four factor-model alphas and one “manipulation-proof” measure). The authors control for relevant fund characteristics, including fund size and fund turnover.
Abstractor’s Viewpoint
The cross-sectional regression coefficient–based measures that the authors use to represent propensities to buy and sell certain kinds of stocks are intuitively appealing. The measures seem to effectively capture fund manager tendencies to sell and, in particular, buy stocks with a high amount of media coverage.
In their closing paragraph, the authors mention the notion of investors taking the other side of trades that are driven by limited attention. I think this is an intriguing idea and look forward to seeing research examining such investors.