Analysis of Dividend Yield Prediction Power of BIST Stock Prices: Evidence form Causality Tests

Abstract

The predictive power of dividend yield on BIST stock prices is examined in this study, adopting the Campbell and Shiller (1988a) model. Thus, it will be tested whether the dividend yield is a successful factor in predicting stock prices. The study has important implications for theorists, investors, and policymakers. First of all, revealing the effect of dividends on firm value will be decisive in determining dividend policies for managers. Once the predictive power of dividend yield is demonstrated, investors and portfolio managers can use the present value approach to determine whether a particular stock is overvalued or undervalued. In addition, knowing the effects of dividend payouts on market dynamics will guide the policy decisions of regulatory authorities. Toda-Yamamoto (1995) and Hatemi-J (2012) bootstrap asymmetric causality tests were adopted as a research methodology. Dividend yield and monthly closing price series of stocks that show continuity in the BIST 100 index and pay dividends continuously throughout the December 2011 – November 2021 period are used in the research. To the Toda-Yamamoto causality test findings, causal relationships from dividend yield to stock prices were determined in AGHOL, BIMAS, EGEEN, INDES, SARKY, SELEC, and TOASO stocks. Hatemi-J's (2012) bootstrap asymmetric causality test indicates fewer causality relationships compared to the Toda-Yamamoto test. A causal relationship could not be determined between the positive or negative components of dividend yield and price variables in many stocks. In the background of the mixed findings of the predictive power of the dividend yield, the structural breaks in the variables examined, the predictive power of the dividend yield changing over time, the weakening of the predictive power depending on the market conditions, the nonlinear relationship between the dividend yield and the price, the stock-based tests are more disadvantageous than the index-based tests, the estimation power of the dividend yield increases more when the annual frequency observations are used rather than the monthly frequency observations, the changes in the dividend policy of the companies and the share repurchase decisions may be weakened the estimation power of the dividend yield.