eThesis - School of Business electronic theses

Abstract

PURPOSE OF THE STUDY The purpose of the thesis is to study whether established psychological concepts facilitate understanding various dividend related phenomena, such as dividend smoothing. More specifically, this thesis examines the role of reference points and signaling in a dividend framework among EU15 companies and aims to lend support to earlier findings in different geographical settings. The set of hypotheses examine both the dividend policy setting characteristics (managers' perspective) and the market reaction to dividend changes (investors'perspective). At the core of the study is the assumed non-linearity of DPS change inflicted market reactions. Also other interesting features associated with dividends, such as the role of dividend streaks and round numbers, are explored.

DATA AND METHODOLOGY The dataset utilized in the study comprises 31,520 DPS observations and 21,740 DPS change observations. The sample is based on companies domiciled in EU15 countries. The sample period extends from January 1993 to December 2012. All data is retrieved from Thomson ONE database. To test the hypotheses, various basic statistical methods are utilized with OLS regressions and descriptive statistics serving as the starting point. Piecewise linear regressions are also employed to study the expected non-linearity of dividend change inflicted market reaction.

RESULTS The empirical results of this study show the important role of psychological concepts in explaining dividend related phenomena. The vast majority of results are statistically and economically significant and consistent with the hypotheses set forth. The results imply that investors evaluate the dividend payments in comparison to DPS rates established in prior periods. Investors react to dividend changes in a manner which suggests that managers are believed to signal information through dividend policy. Moreover, prospect theory value function characteristics can be seen in investors' market reaction patterns and the salience of a reference point is strengthened by dividend streaks. In addition, dividend smoothing elements are evident in the distribution of raw DPS observations as well as the centering on round numbers for both raw DPS and DPS change observations. It seems that managers are reluctant to cut an established level of dividend in fear of an adverse market reaction. This effect is not, however, present in the scarce number of dividend cuts observed within EU15 companies.