Decomposition of the dividend forecast model: firm characteristics or market discrimination? Evidence from the Korean market

Abstract

Historically, KOSPI firms have shown consistently higher chances of dividend payment compared to KOSDAQ firms. The counter factual decomposition technique popularised by Oaxaca and Blinder is adapted to explain the dividend differentials between the two markets. Our results suggest that there exists market discrimination that cannot be explained by the traditional dividend forecast model of Fama and French (2001). After controlling for the group differences in firm characteristics such as size and profitability, KOSPI firms still have a higher probability of paying dividends, and, especially during the crisis period, they deal with risks better and are less affected by macroeconomic shocks. We observe the level difference after controlling for firm characteristics as well as the difference in the sensitivity of dividends with respect to the dividend predictors between two markets. For the crisis period, the explanatory power of firm characteristics drops, and the market effect plays a dominant role in explaining dividend behaviours.