There’s a large body of evidence within the academic literature demonstrating that individual investors have a tendency to overpay for an asset that contains a large payoff in its past payoff distribution. Various explanations for this behavior, which has been shown to negatively impact returns, include overweighting the probability of tail payoffs in decision-making, overestimating the likelihood of high payoff states, and a preference for positively skewed fat-tailed (or lottery-like) distributions, even though such assets have a high likelihood of poor returns and a small probability of extreme positive returns.
The pricing implications of this behavior have been studied in the stock, option and IPO markets, as well as in experimental settings. For example, penny stocks, the stocks of companies in bankruptcy, IPOs and extreme small growth stocks with low profitability and high investment all have had very poor returns.
Analyzing the Literature
Ferhat Akbas of the University of Kansas and Egemen Genc of Erasmus University in The Netherlands — authors of the February 2016 paper “Do Mutual Fund Investors Overweight the Probability of Extreme Payoffs in the Returns Distribution? — contribute to the literature by investigating the impact that the preference for tail returns in a fund’s return distribution has on future fund flows. They provided the following example:
For the period from November 2004 through November 2005, the Van Kampen Value Fund and the Dreyfus Premier Fund had similar average style-adjusted returns, but differed in the right tail of their return distributions. The Van Kampen Fund posted a large positive return relative to its peers in May 2005 while experiencing fairly mediocre, sometimes even negative, style-adjusted returns throughout the rest of the year. On the other hand, the Dreyfus Premier Fund performed better than its peers during most of the months in this same period, underperforming in only one month.
The authors’ goal was to determine whether mutual fund investors tend to invest more in funds like the Van Kampen fund than they do in ones like the Dreyfus Premier fund because they overweight a single high payoff state (like that from May 2005) or overestimate the likelihood of this type of payoff state, despite the fact that the Van Kampen fund’s average past performance over the period from November 2004 through November 2005 and its other characteristics (as controlled for in their analyses) are similar to those of funds like the Dreyfus Premier fund.
To test their hypothesis, the authors used active share as a measure of how active a fund was (in other words, how much it deviates from its benchmark index). Their belief was that less risk-averse investors would be attracted to funds with higher active shares because they would be more likely to have fat-tailed distributions — higher maximum returns (MAX). As a result, the funds with higher active share should exhibit stronger MAX-flow relationships relative to funds with low active share.
Using Lipper investment categories, Akbas and Genc’s fund data is from the Center for Research in Security Prices (CRSP) Survivor-Bias Free U.S. Mutual Fund Database and includes domestic equity funds from January 1999 to December 2014. The authors’ full sample included more than 8,000 funds. Controlling for style and past performance, the following is a summary of their findings:
- High MAX funds trade more often and, as expected, are more volatile.
- Funds with high MAX over the previous year are more likely to show the same features in the future. If investors indeed derive utility from having a preference for extreme payoffs with small probabilities, it’s more likely they’ll gain extreme payoffs if they follow high MAX funds.
- There is a positive and statistically significant relationship between the MAX and future fund flows. The positive MAX-flow relationship suggests that mutual fund investors are willing to direct more flows to funds that have recently exhibited a large positive payoff in monthly return distributions.
- There is a strong MAX-flow relationship within more active funds, but no relationship within less active funds (for example, closet indexers).
- As expected, the effect of MAX on future flows is significantly more pronounced in retail funds (as compared to institutional funds).
- The findings are robust to various definitions of MAX, as well as to controlling for various other fund characteristics (such as past performance, two different measures of performance convexity, total volatility, total skewness, idiosyncratic volatility, idiosyncratic skewness, size (AUM), age, turnover, expenses and load fees).
- There is no evidence of a positive relationship between MAX and future performance. Thus, MAX isn’t a good metric for investors attempting to select funds with superior future returns.
The authors concluded: “Overall, the evidence suggests that fund investors are willing to direct more flows toward funds that have a large positive payoff in monthly return distributions.” They added: “The positive relation between MAX and future fund flows is best explained by these models of extreme payoff preference.”
Their contribution to the literature is that they show overweighting tail returns is a more common phenomenon in financial markets than previously thought, as it also affects investors’ decisions even in the mutual fund market.
The Bottom Line
The bottom line is that the research shows that overweighting maximum tail returns and a preference for lottery-like distributions is harmful to returns. Forewarned is forearmed. A simple way to avoid these apparently all-too-human traits is to invest in index funds, or other broadly diversified and passively managed funds (some of which even screen out stocks with these negative characteristics).