Behavioral Finance argues that several anomalies could be explained by relaxing the central proposition of EMH (Efficient Market Hypothesis), that is, investors’ rationality. Indeed, recent empirical and experimental work provides additional evidence that human judgment errors may impact financial market price behavior instead of simply canceling each other out.

Our work aims to provide a better understanding of several cognitive biases, human judgment errors occurring during mental information processing and decision making. We propose a general mathematical framework of price formation, allowing for subsequent modeling of individual cognitive biases, without affecting the underlying assumptions. Price characteristics, such as expected equilibrium price, volatility but also trading volumes, can be potentially analyzed for each bias’ functional form.

This theoretical work on some specific cognitive biases allowed for further empirical research, using quarterly earnings estimates, announcements and subsequent price reaction data for listed US companies over the period 1983-1999. We selected events, for which previous earnings information pointed to a highly probable bias; which could be either an anchoring bias to previous earnings values or a representativeness bias given a series of past earnings surprises. We showed that these events exhibit, at the time of the current earnings announcement, important and highly significant abnormal returns, indicating a correction phenomenon of the (assumed) previous under- or over-reaction.