The Earnings Information Content of Dividend Announcements
Kiiskinen, Karo (2008)
Kiiskinen, Karo
2008
Kuvaus
Opinnäytetyö kokotekstinä PDF-muodossa.
Tiivistelmä
A firm’s dividend policy has been the object of extensive studying. One of the dividend policy explanations considers dividends as a possible signalling device. The basis for the signalling studies is the information asymmetry between firms’ managers and ow¬ners, and dividend changes are then seen as an effort to reduce it. Reviewing the subject, the purpose of this thesis is to examine the relationship between current dividend changes and future earnings changes. A positive relationship would indicate that ma¬nagers are signalling future prospects of their firms through dividends.
The sample under research included 76 Finnish companies traded in the Helsinki Stock Exchange during the research period 1990-2001. The required accounting data was gathered from the database of ETLA, and was supplemented with the year-end share prices and number of shares of the firms studied. The panel data set was then formed to investigate dividend signalling across the market and through time. Regression analysis was used as a statistical methodology. Alternative models and procedures used in this thesis are defined in the Benartzi, Michaely & Thaler’s (1997) and Nissim and Ziv’s (2001) study.
The empirical results were very inconsistent. None of the models or statistical proce¬dures provided results strongly supporting the signalling hypothesis. The most common result was that the relationship between dividend changes and future earnings changes is weak and insignificant. Additionally, both positive and negative relationships were found. In conclusion, on average the earnings information that possibly has been sig¬nalled by dividend changes was relatively small.
The sample under research included 76 Finnish companies traded in the Helsinki Stock Exchange during the research period 1990-2001. The required accounting data was gathered from the database of ETLA, and was supplemented with the year-end share prices and number of shares of the firms studied. The panel data set was then formed to investigate dividend signalling across the market and through time. Regression analysis was used as a statistical methodology. Alternative models and procedures used in this thesis are defined in the Benartzi, Michaely & Thaler’s (1997) and Nissim and Ziv’s (2001) study.
The empirical results were very inconsistent. None of the models or statistical proce¬dures provided results strongly supporting the signalling hypothesis. The most common result was that the relationship between dividend changes and future earnings changes is weak and insignificant. Additionally, both positive and negative relationships were found. In conclusion, on average the earnings information that possibly has been sig¬nalled by dividend changes was relatively small.