5.1. Announcement Effect of Cash Dividend Changes on Share Prices
To assess the effect of cash dividend changes on the share prices, we collect a sample consisting of firms that changed their cash dividends during the period 2000 to 2004. Each observation in the sample satisfies the criteria mentioned in Part 4.
The resulting sample contains 422 announcements of cash dividend decreases and 460 announcements of cash dividend increases.
Table 6 presents the results of this analysis. For the cash dividend increase sample, there are 16 negative-abnormal-return trading days and only 3 positive- abnormal-return trading days in the 20 days preceding the declaration of cash dividend increases. For the cash dividend decrease sample, there are 13 negative- abnormal-return trading days and only 4 positive-abnormal-return trading days in the 20 days preceding the declaration of cash dividend decreases. Thus, the stock prices tend to perform poorly before the announcement of cash dividend changes. The 20-day cumulative abnormal return before the announcement of cash dividend changes is a negative 0.65% for cash dividend decreasing firms and is a significantly negative 0.75% for dividend increasing firms. This fact can also be gotten by observing the left half of Figure 1.
For the increasing cash dividend sample, the abnormal return is a significantly positive 0.23% at the day of a positive dividend change announcement. The abnormal returns of the 8 days succeeding the announcement of cash dividend increases are all positive, and four of them are even significant. Most of the
abnormal returns of the 20 days after the cash dividend increase announcement are positive. As a result, the cumulative abnormal return starts rebounding at the day of declaration.
For the decreasing cash dividend sample, the abnormal return is negative at the announcing date, but it is not significant. There are 13 positive-abnormal-return trading days and 6 negative-abnormal-return trading days in the 20 days after the announcement of the cash dividend decreases. All of these phenomena can be verified in Figure 1. Therefore, whether the cash dividend increases or decreases, the share prices are inclined to perform better after the announcement of dividend changes. Cash dividend changes have a positive influence on the share prices.
Figure 2 provides the graph of cumulative abnormal returns on days surrounding the announcement of cash dividend changes. Table 7 offers the cumulative abnormal returns for different event windows. As Figure 2 and Table 7 show that stock prices perform poorly before the announcement, but perform well after the announcement. Thus, Figure 2 echoes the viewpoints we have briefly concluded in the last paragraph. Figure 2 clearly shows the main difference in the price impact between the cash dividend increases and decreases. In the left half of Figure 2, two curves almost move synchronously. Nevertheless, the cumulative abnormal return curve of the increasing dividend sample rises much more intensely than that of the decreasing cash dividend sample after the announcement of cash dividend changes. Thus, after the announcement of dividend changes, the positive price impact of the increasing cash dividend sample is more significant than that of the decreasing cash dividend sample. The CAR of (0, 1), (0, 2), (0, 3), (0, 5), (0, 10), and (0, 20) event windows in Table 7 confirm this assertion very well.
In short, the empirical results only partly support the dividend signaling hypothesis. The analysis results of the increasing cash dividend sample are
consistent with the empirical implication of the signaling hypothesis, and the cash dividend increases are accompanied by the stock prices moving in the same direction.
However, the empirical results of the decreasing cash dividend sample do not comply with the signaling hypothesis, and the cash dividend decreases are accompanied by the stock prices moving in the opposite direction. Therefore, it is feasible to conclude that investors in China respond positively to the cash dividend announcement whether it is increasing or decreasing. In other words, cash dividends are welcome in China nowadays according to the results of our analysis.
5.2. Price Impact of Cash Dividend Changes on Different Markets
After assessing the effect of cash dividend changes on the market price, we try to compare the price impact of cash dividend changes on different markets. China has only two stock exchanges, one is the Shanghai Stock Exchange and the other is the Shenzhen Stock Exchange. The sample firms are all collected from the listed firms of these two exchanges. For the 460 increasing cash dividend firms, 287 firms are drawn from the Shanghai Stock Exchange and 173 firms are drawn from the Shenzhen Stock Exchange. For the 422 decreasing cash dividend firms, 253 firms are drawn from the Shanghai Stock Exchange and 169 firms are from the Shenzhen Stock Exchange. The results of this analysis are summarized in Figure 3, Table 8, and Table 9.
For the increasing cash dividend sample, most abnormal returns preceding the cash dividend change declaration are negative both for the Shanghai subsample and the Shenzhen subsample, but the negative abnormal returns of the Shenzhen subsample are more intense than that of the Shanghai subsample. On the day of dividend increase announcement, significantly positive abnormal returns occur for both subsamples. After the cash dividend increase announcement, the cumulative abnormal returns of both markets exhibit upward-moving trends. On the other hand,
the cumulative abnormal returns of the Shenzhen subsample increase more fiercely than that of the Shanghai subsample in a shorter time period, but the cumulative abnormal returns of the Shanghai subsample increase more steadily than that of the Shenzhen subsample in a longer time period. Generally speaking, although the overall abnormal return features of the Shanghai subsample and the Shenzhen subsample are slightly different, they are roughly the same with the pattern of full cash dividend increase sample as explained in Section 5.1.
For the cash dividend decrease case, the abnormal return pattern of the Shanghai subsample has a distinct difference from that of the Shenzhen subsample - that is, most of the abnormal returns of the Shenzhen subsample are negative before the announcement of the cash dividend decrease, but the abnormal returns of the Shanghai subsample show a comparatively positive performance before the announcement of cash dividend decrease. On the day of cash dividend decrease declaration, the abnormal return of the Shanghai subsample is negative, but that of the Shanghai subsample is zero. After the announcement of cash dividend decreases, the cumulative abnormal returns of both markets display upward-moving trends.
On the right-hand side of Figure 3, we can easily see that all of the four cumulative abnormal return curves move upward. We thus conclude that, on the whole, the announcement effect of cash dividend changes is positive for both markets, and there is no great difference between the announcement effects of the two markets.
5.3. Announcement Effect of Different Sample Period
The empirical results from above are completely derived from the sample of the period 2000-04. We wonder whether the announcement effect of a dividend change on share prices alters with time. Hence, we analyze the announcement effect with the sample year 1999 and cash dividend changes. The announcement effect of cash dividend changes in 1999 is quite important, because it provides the announcement
effect of a quite different cash dividend distribution situation. We have pointed out that, in 1999, the ratio of cash-dividend-paying firms was low (36.78%) and the rules which severely influence the cash dividend payment decision were not promulgated then. Nevertheless, in 2001 the rules compelling listed firms to distribute cash dividends were announced and executed, and the ratio of cash-dividend-paying firms increased dramatically.
Figure 4 and Table 10 present the results of this analysis. Figure 4 and Table 10 clearly show the facts that the announcement effect of a cash dividend decrease is significantly positive, and the announcement effect of a cash dividend increase is insignificantly negative. The absolute value of cumulative abnormal returns for (0, 20) the event window of the 1999 sample is more than that of the 2000-04 sample.
These empirical results are not only totally opposed to the dividend signaling hypothesis, but are also different from the analysis findings we got from 2000-04 dividend change sample.
The announcement effects of dividend changes before and after year 2000 present an immense variation. We think the promulgation and execution of two administrative rulesmay reasonably explain the variation. Before these two laws were declared, the ratio of cash-dividend-paying firms was lower, and capital gains were the main source of income for investors. Therefore, investors may look down upon cash dividends and react negatively to cash dividend increases and positively to cash dividend decreases. However, when the ratio of cash-dividend-paying firms increased sharply after these two rules were announced, except for capital gains, cash dividends became another major source of income. Thus, investors’ attitudes towards cash dividends may become positive, and react positively to all cash-dividend-paying firms.
5.4. Affecting Factors Analysis of the Announcement Effect
To further investigate factors that may affect the market reaction around the announcement of cash dividend changes, we estimate the following cross sectional regression using the OLS regression methodology:
+ whereCAR is the cumulative abnormal returns for different event windows around the announcement of the dividend changes; DC is the percentage change in the cash dividend payment; DY is the dividend yield at the time of the announcement of the cash dividend changes; DP is the dividend payout ratio (cash dividend per share / earnings per share); ASSETS is the logarithm of the book value of the total assets at the time of the announcement of the cash dividend changes; P/B is the price-to-book ratio at the end of the year; DEBT is the debt ratio ( book value of total liabilities / book value of total assets); TATR is the total assets turnover rate; ROA is the return on assets; NF is proportion of non-floating shares. Table 11 summarizes the investigation results.
Table 11 indicates that only four financial variables have a significant impact on the announcement effect of cash dividend changes. First, dividend yield has a significantly positive relation with CAR, but the significance disappears gradually with the time interval extension of event windows. In other words, the high dividend yield stocks are inclined to perform better than the low dividend yield ones. Next, for the event window of a longer time interval, P/B has a significantly negative impact on cumulative abnormal returns. In other words, value-oriented (low P/B) firms tend to have higher cumulative abnormal returns in a longer time interval, and the intrinsic value of a firm may be reflected on the longer cumulative abnormal returns. Third, market investors react positively to the profit index ROA for the (0, 10) event window, but the significance of this positive connection disappears with the extension or
curtailment of the event window. Fourth, the ratio of non-floating shares has a negative impact on the cumulative abnormal returns of (0, 3), (0, 5) event windows.
We have concluded that non-floating shareholders usually occupy the majority of the shareholder structure and dominate the board of directors. Because the holding costs of non-floating shareholders are lower (dividend yields are higher), and the dividend distribution is the only mechanism they can realize incomes, the controlling shareholders (usually non-floating shareholders) may make an over-lavish cash dividend policy that hurts the interest of floating shareholders. Therefore, the negative relation between the ratio of non-floating shares and cumulative abnormal returns does make great sense.