ABSTRACT This study investigates how extreme return components ( MAXs ) of smart beta exchange‐traded funds (ETFs) influence heterogeneous investor choice. We find that extreme overnight returns (overnight MAXs ) have substantial ETFs' flow predictability above and beyond that provided by standard risk, performance, and market‐timing measures. This suggests that individual investors tend to choose funds based on historical overnight MAXs . The persistence of overnight MAXs supports the lottery preference theory, which posits that retail investors disproportionately focus on the likelihood of high payoff states in a fund's past return distribution. Furthermore, we demonstrate that overnight and intraday MAXs negatively and positively predict future smart beta ETFs' performance, respectively. This divergence indicates that heterogeneous investors hold varying beliefs and employ distinct trading activities with respect to those lottery‐like fund return distributions. Finally, we show that investor attention is beneficial to explain our findings.
Liu et al. (Tue,) studied this question.