ความสัมพันธ์ระหว่างการประกาศจ่ายเงินปันผลกับการเปลี่ยนแปลง ของราคาหุ้นในตลาดหลักทรัพย์แห่งประเทศไทย / อนุวัฒน์ จงยินดี = Relationship between dividend announcement and the market price of stocks in the securities exchange of Thailand / Anuwat Jongyindee
The study of Eugene F. Fama, Lawrence Fisher, Michael C. Jensen and Richard Roll (FFJR) regarding the nature of abrupt changes in the market price of stocks caused by the announcement of stock splits in the New York Stock Exchange concluded that the New York Stock Exchange is an efficient market at the semi-strong level. This is an extremely interested study that inspired the author to adapt and use it as a guide for this thesis, which is a study of the relationship between dividend announcements and the market price of stocks in the Securities Exchange of Thailand under different market conditions in 1977 (B.E. 2520), which was considered bullish and 1979 (B.E. 2527), which was considered bearish. The study classified the common stocks available on the market into three categories, i.e. stocks with increasing dividend payments, decreasing dividend payments and sustained dividend payments in comparison with the preceeding year. Apart from this, the three classified stock catagories were combined into other catagories in order to see if the price of stocks would be affected by the dividend announcements in a shorter time period of 12 weeks prior to the dividend announcements, the day of the dividend announcements and 12 weeks after the dividend announcements of each particular classified stock category in this years mentioned. The study employed equations of the Market Model to calculate the average value of residual term and the cumulative value of residual term of each particular classified stock category in the years mentioned and analysed relationships by graphing the average value of residual term on the Y axis against the different timing of the announcements of the shorter plus/minus 12 weeks period in the years mentioned on the X axis. By the same means a graphic analysis was made of the cumulative average value of residual term against the different timing of the announcements of the shorter plus/minus 12 weeks period in the years mentioned. Further, the hypothesis was tested by the statistical T- Distribution method employing the average value of residual term of each particular classified stock category in the years mentioned with the assumption that the dividend announcements have no effect on the volatilities of the market price of stocks examined at the 99% statistical confidence level. In addition, a statistical Paired-Test analysis was conducted to test if under different prevailing market conditions that were imposed upon by the dividend announcements would the market price of stocks also behave differently or not. The results of the graphic analysis showed that under the market conditions of 1977 (B.E. 2520), which was considered bullish, the market price of stocks would be effected by the dividend announcements only on the day the announcements were made. Even if in the time period of 1-2 weeks after the day of the dividend announcements, the period that buyers were still receiving the dividend announcements which usually influenced the market price of stocks to increase, the graphic analysis showed that the dividend announcements had no effect on the volatility of the market price of stocks because the average value of residual term declined and bore negative value. Under the market conditions of 1979 (B.E. 2522), which was considered bearish, the market price of stocks would be affected adversely by the dividend announcements on the day announcements were made because the average value of residual term was nearly equal to the value in the time period prior to the day of the dividend announcements. And in the time period of about 1-2 weeks after the day of the dividend announcements the market price of stocks would be affected positively by the dividend announcements because the market price of stocks became higher than the market price on the day the announcements were made. This confirms that the behavior of the market price of stocks is insensitive to the dividend announcements. The reason is that if the market price of stocks were sensitive to the dividend announcements the market price of stocks must change abruptly as from the day the dividend announcements were made. Besides the above graphic analysis results, the results of statistical analysis using the same data also showed that the dividend announcements have no significant effect on the volatility of the market price of stocks at the 99% statistical confidence level. This might be due to the fact that other information which had some effect on the volatility of the market price of stocks came into the market before the dividend announcements and lessoned the importance of the dividend announcements. For example, information in the quarterly reports, a rather important factor used by the companies to manage their dividend policies. Furthermore, the Paired-Test statistical analysis showed that under either bullish or bearish conditions being imposed upon by the dividend announcements, the market price of stocks also behave indifferently. With the following assumptions: in the shorter plus/minus 12 weeks period, the dividend announcements was the only information coming into the market, the others was fixed, the studied stocks were the dividend paying stocks in 1977 (B.E. 2520) and 1979 (B.E. 2522). The Least Square Method and the Market Model to calculate the value of alpha/beta and residual term, respectively, were applied from the theory of the Capital Asset Pricing Model (CAPM), which has so many assumptions that in practice it’s not in compliance with all such assumption. Moreover, there were some variables such as the stocks with little buying and selling and nearly changeless prices which might not be good representative data to calculate the value of alpha/beta and respective residual term. In this study the value of residual term was calculated from the weekly data on the day the dividend announcements was made and in the shorter plus/minus 12 weeks period in the time period as from the day the Securities Exchange of Thailand was established – 30 April 1975 (B.E. 2518) to December 1980 (B.E. 2523). It is such a short time period in comparison with the study of FFJR which calculated the value of residual term from the monthly data of 940 stocks on the effective data of the stock split and in the shorter plus/minus 30 month period in the time period of 33 years (from January 1927 to December 1969). Thus, given not only the conditions of the theory but also so many conditions in practice, it is difficult to further conclude whether the Securities Exchange of Thailand is an efficient market at the semi-strong level. But if the conditions or the time period of the study were changed, the results might be different from the results of this thesis.
