Do investors/traders taking cues from the trades of top performers produce the momentum effect? In his December 2006 paper entitled “Follow the Leader: Peer Effects in Mutual Fund Portfolio Decisions”, Lukasz Pomorski investigates whether actively managed equity mutual funds tend to follow the stock trading leads of outperforming peers as the picks become known via the media and quarterly filings. He defines outperforming (leader) funds in two ways: (1) funds with alphas in the top 5% over the past two years, and (2) funds on the Forbes Honor Roll (high media exposure). He calculates overall leader activity in a stock based only on trades by leader funds with a position in the stock. Using mutual fund holdings and performance data for 1980-2003 (96 quarters), he finds that:
- Leader funds (ranging from 13 to about 100 over the sample period) average twice the monthly returns of the remaining funds, with average alpha of about 1.5% per month. The Forbes Honor Roll funds average only 61% turnover, compared to over 100% for the typical fund.
- Other funds do mimic the trading of leaders, amplifying leader buys of a given stock by 15-30% in the following quarter.
- Mimicking behavior concentrates in stocks with the most informative leader trades: initiations and deletions, and trades in stocks with low market capitalization, little analysts coverage or high idiosyncratic volatility.
- Funds with below-median alphas react to leader trades about twice as strongly as funds with above-median alphas. The best funds do not mimic.
- A zero-cost value-weighted portfolio, long in stocks likely to be bought and short in stocks likely to be sold
by mimickers, generates average returns (and alpha) of about 0.35% per month during the quarter after the leaders trade, with a relatively low monthly standard deviation of 2.3%. Results are statistically and economically significant. - Neither stock price momentum nor stock fundamentals (news) explain this “follow the leader” trading behavior.
The author notes that quarterly mutual fund data probably understate the amount of mimicking that occurs across all timeframes and all market participants.
In summary, mimicking the most informative actions of outperforming investors/traders reliably generates abnormal short-term returns. Such behavior may explain some of the momentum effect.
