Keele Research Repository

The main focus of the research reported in this thesis is the dividend signalling theory. More specifically, I investigate the dividend signalling theory from the perspective of orthodox finance and from the perspectives of two unorthodox areas of finance (behavioural finance and the calendar anomalies literature) using a sample of firms from the FTSE 350 index observed between 1990 and 2015.
The thesis revolves around four research questions. First, do dividend changes contain any information about future earnings? Second, do dividend-increase (decrease) announcements have a positive (negative) effect on stock returns? Third, does investor sentiment play any role in the reaction of the stock market to dividend announcements? And fourth, do calendar anomalies play any role in the relationship between dividend announcements and stock returns? My thesis employs an empirical approach and makes original contributions to the behavioural finance literature, the corporate finance literature and the literature on calendar anomalies.
According to my analysis, there is no evidence that dividend-increase (decrease) announcements are followed by increases (decreases) in firm earnings. However, at the same time, I document that dividend-increase (decrease) announcements are accompanied by abnormal increases (decreases) in stock market returns. These conflicting findings represent a puzzle that I would hope to investigate in my future research. From a behavioural finance point of view, I also find some evidence that investor sentiment influences the response of stock prices to dividend announcements. More specifically, I document that dividend-decrease announcements have a smaller than usual negative effect on stock returns when temperature in London is high and, as a result, investor sentiment is likely positive. Similarly, I find that the negative impact of dividend-decrease announcements on returns is bigger than usual when the air pollution level in London is high and, as a result, investor sentiment is negative. With regards to calendar effects, consistent with the “sell in May and go away” anomaly, I document that the stock market reacts more positively to dividend-increase announcements during the November-April period than during the rest of the year, and the stock market reacts less negatively to dividend-decrease announcements during the November-April period than during the rest of the year. Regarding the turn-of-the-month anomaly, I find that that the stock market reacts less negatively (actually positively) to dividend-decrease announcements if they occur at the turn of the month than if they occur during the rest of the month. Counter-intuitively, the stock market seems to react less positively (more negatively) to dividend-increase (decrease) announcements if they occur in January than if they occur during the rest of the year. Previous investigations of the dividend-signalling theory have relied exclusively on an orthodox approach; the findings documented in this thesis suggest that future investigations about this theory could benefit from the insights produced by the behavioural finance literature and the literature on calendar anomalies.