|
Abstract:
Chapter One examines stock price reaction to earnings news on earnings announcement date and its implication for post-earnings announcement drift. We document that stock prices often move in the opposite direction of earnings surprises on announcement date. Thirty percent of the stocks with extreme negative earnings surprises experience significant positive returns, while thirty percent of the stocks with extreme positive earnings surprises experience significant negative returns. Prices of the former subsequently reverse, whereas prices of the latter do not drift. Further analysis shows that the divergent price reaction patterns cannot be explained by two widely documented behavioral biases, the representativeness heuristic and the disposition effect. Alternatively, we find that investors react differently to earnings-related qualitative news, which contributes to the divergent price reactions and subsequent price movements. Chapter Two investigates whether the timing of earnings announcements in SEC filing seasons affects market responses to earnings surprises. We document that market responses are more favorable when news announcements are made early in earnings season ("timing effect"). Price reactions on earnings announcement dates and price movements in the following 60 trading days are significantly stronger (weaker) for positive (negative) earnings surprises announced at the beginning of earnings season. Further investigation shows that the timing effect associated with positive earnings surprises is consistent with information transfer in that late good news announcements are accompanied by significant pre-announcement price increase. The timing effect associated with negative earnings surprises is mainly driven by firms' strategic delay of bad news announcements since the timing effect does not exist among advanced bad news announcements. Chapter Three studies the impact of clustered quarterly earnings announcements on stock returns at both the firm level and the aggregate level. We find quarterly earnings announcements are concentrated in a month starting two weeks after the fiscal quarter end. Pre-announcement month contributes the most to quarterly returns. At the firm level, earnings news is positively correlated with contemporaneous and future stock returns (evidence of post-earnings announcement drift). At the market level, aggregate earnings news are negatively correlated with contemporaneous returns. Moreover, aggregate earnings news cannot predict future market returns. The differential price discovery process at the firm and the market level indicates that the implications of individual earnings news is qualitatively different from that of aggregate earnings news.
|