Discussion of “Underreaction to Self-Selected News Events”

Abstract

During the past decade there have been several studies that suggest that stock prices tend to underreact to corporate events. These include announcements of earnings, dividend changes, capital structure changes (e.g., share issuances and repurchases) as well as stock splits, which are analyzed by Ikenberry and Ramnath (2001). 1 Two theoretical articles that have explicitly examined the tendency to underreact to these corporate events include Barberis, Shleifer, and Vishny (1998 hereafter BSV) and Daniel, Hirshleifer, and Subrahmanyam (1998 hereafter DHS). The BSV article is based in part on experimental evidence of what is known as a “conservatism bias,” which essentially means that individuals tend to underweight new information when updating their prior beliefs. This conservatism bias is consistent with the observed underreaction to all of the above-mentioned corporate events. However, as Hirshleifer (2001) discusses, conservatism bias is really just a description of the tendency of individuals to underreact to certain types of information and cannot really be viewed as an explanation for why they underreact. DHS develops a model of investor overconfidence that also examines how investors update their priors in response to new information. In DHS, investors are overconfident about the information they collect on their own, and as a result, overreact to this information. In addition, since they overweight the information they collect on their own, they tend to underweight and, as a result, underreact to information that is provided by managers.