Success/Failure of Past Trades and Trading Behavior of Investors

Using a unique and large dataset, the present paper contributes new insights to the growing literature on behavioral biases in portfolio decisions of individual investors. First, success/failure of recent past trades, where a profitable trade indicates success and an unprofitable trade failure, influences the current trading decisions of investors more than the size of the gains or losses from the trades. In the process, our research establishes the sign of the outcomes of past trades (positive or negative) rather than their dollar size as the main source of a behavioral bias induced by reinforcement learning. Second, the trading decisions we consider include not only volume but also frequency of trading by investors. We find that both of them reflect the behavioral bias in question. The effects of the bias are economically large as well as statistically significant. We also find that, on an average, the bias results in losses for the investors. We introduce several methodological innovations in this paper.