Investor over- and underreaction to earnings announcements: AN EXPERIMENTAL STUDY

Investors do not react appropriately to earnings reports: this is the conclusion of many researchers in the fi eld of fi nance. In this regard it has been amply demonstrated that the inappropriate reactions of investors are expressed in two main phenomena in cases where studies have been made out of errors in assessing risks, frictions in fi nancial markets, and procedural shortcomings. The fi rst anomaly is described as underreaction. It leads to the extension of stock-market prices over the short term (Bernard and Thomas, 1989; Theobald and Yallup, 2004; Nguyen, 2005). Conversely, the second anomaly is described as overreaction. It is more of a long-term matter, following certain specifi c phenomena. These include, for example, the occurrence of consecutive good or bad news (Barberis et al., 1998) or the "persistence" of the performance of the securities1 (DeBondt and Thaler, 1985; Chopra et al., 1992). This under- and over-reaction may be attributed to various factors. According to Bernard and Thomas (1989) and Ball and Bartov (1996), it refl ects ignorance or incorrect assessment of the auto-correlation between the quarterly components of a series of earnings over time. In contrast, according to Bloomfi eld et al. (2003), the inadequate or excessive market reactions persist even if we consider series of annual earnings only. They all arise from the over-reliance of investors on the preceding announcement2.