Do changes in dividend policy signal the future or the past?

Abstract

This research study investigates whether Malaysian Main Board companies pursue changes in dividends to convey information on the future profitability of the companies by studying on a total of 2,679 firm-year observations from year 1998 to 2007. The Ordinary Least Square regression results show that Main Board listed companies do not use dividends as a signalling tool to convey information on the future prospect of the companies. In fact, positive and significant relationship is found to be stronger between changes in dividends in year T=0 with concurrent changes in earnings in year T=0, which is consistent with the previous findings by Benartzi, Michaely and Thaler (1997) and Nissim and Ziv (2001). Further, regression analysis on sub-samples of firm-year observations categorized by the period of stable dividends before the dividend change events (2 years, 3 years and 4 years) shows no relationship exists between the stability of dividends before dividend change events with the extent of dividend signalling. No relationship is found between size of dividend change, size of dividend yield with the extent of dividend signalling. Multiple regression by incorporating industry dummies in the regression equation shows that no difference in changes in earnings between the 3 major sectors (Industrial, Trading/Service and Consumer) when there are changes in dividends, indicating that industry effect does not have any influence on the extent of dividend signalling of the companies.