Opening the Black Box : From an Individual Bias to Portfolio Performance

We suggest an experimental design that can help opening the black box of investor behavior by documenting a channel of how biases affect portfolio performance. We study two of the most important investor biases (overreaction and overconfidence), show how they are related, and analyze their consequences for portfolio choice and resulting portfolio performance in a controlled experimental setting with 104 participants. The main innovation of our study is that we go beyond just documenting a correlation between overconfidence on the one hand and investor behavior as well as resulting portfolio performance on the other hand. We empirically identify the precise channel (overreaction) which is proposed by some models. We find that subjects overreact on average, i.e. forecasts are too optimistic after positive signals and too pessimistic after negative signals. Furthermore, there is greater overreaction when subjects are more overconfident. Moreover, overreaction is related to risk taking in a portfolio choice task thereby adversely affecting portfolio efficiency