Weighing the Evidence of Dividend Initiations

This paper analyzes post-event stock performance based on the strength and the weight of evidence of dividend initiations. The strength refers to how extreme the evidence is; the weight refers to how credible the evidence is. Griffin and Tversky’s (1992) experiments show that people are underconfident when presented with low strength and high weight information and overconfident with high strength and low weight information. This study shows that Griffin and Tversky’s results extend to dividend initiation as well. Portfolios with high strength and low weight overreact to dividend announcement, while portfolios with low strength and high weight underreact. The empirical results in this paper show that the strength-weight model can explain the underreaction-overreaction puzzle and suggests that psychological biases may impact stock performance.