A stock price process model reflecting dynamics of traders' behaviors

Abstract

In this paper, we propose a new stock price process model that reflects the dynamics of traders behaviors. Our model has two implications: First, in the both seller group and the buyer group, the stock price moves in favor of the minority group, not the majority group, and the smaller the minority group is, the larger the change in the price. Second, in both the seller group and the buyer group, traders follow (herd to) the behavior of the minority, and the smaller the minority group is, the larger the herding. Then, exploiting our proprietary data set, we show that our model explains the market well. We also use our model to show that we can predict stock prices via a machine-learning technique that we develop.