سیاست تقسیم سود، جریان پول نقد و سرمایه گذاری در ژاپن

This study provides evidence in support of the cash flow information (CFI) hypothesis focusing on the Japanese firms. Dividend changes indeed convey information about the firm's cash flows. Although the free cash flow hypothesis is to some degree supported by the evidence in firms' investment behavior, dividend policy is not used by Japanese firms to control the overinvestment problem. In addition, the dividend clientele effect does not appear significant around dividend announcements in Japan. Given the specific institutional features of the Japanese market, we find that investment spending is very sensitive to liquidity constraints for nonkeiretsu firms, but not so for keiretsu firms.

Dividend announcements have been the focus of extensive research in the U.S. markets. Hypotheses based on the information asymmetry and the overinvestment of free cash flow provide common interpretation of stock price reactions to dividend announcements. Announcements of dividend changes are usually associated with significant excess returns consistent in various ways with these nonmutually exclusive hypotheses. The purpose of this study is to test these hypotheses in Japan. Moreover, this research also provides general evidence about the relation between firms' cash flows and dividend changes. That is, how dividends are financed and whether the careful management of dividends is regarded as important by corporate Japan. Finally, it makes a contribution to the investment literature by studying the general investment behavior of Japanese firms. One popular hypothesis, the cash flow information hypothesis, suggests that management uses dividends to convey inside information about the firm's cash flow not available to other market participants. An announcement of a dividend increase (decrease) is a declaration by the management of their knowledge of favorable (unfavorable) future prospects for the firm. Larger changes in the dividend imply greater changes in the firm's cash flow. Therefore, the magnitude of the market reaction is positively related to the size of the dividend change (e.g., Asquith and Mullins, 1983 and Denis et al., 1994). The alternative hypothesis, the free cash flow hypothesis, implies that dividends are paid out to stockholders in order to prevent managers from building unnecessary empires in their own narrow interests. Entrenched managers have the tendency to invest free cash flow in size-increasing but nonprofitable projects. Stockholders would prefer to see an increase in dividend that would reduce the free cash flow available to the managers. As a result, stock prices react favorably to announcements of dividend increases and unfavorably to dividend decreases by overinvestors (Lang and Litzenberger, 1989). Most of the empirical studies that intend to distinguish between these competing hypotheses focus on the examination of the cumulative excess returns around dividend announcements. Clearly, either hypothesis is consistent with positive stock price reaction to an announcement of a dividend increase and a negative reaction to a dividend cut. Surprisingly, however, when the various studies, which are all carried out using U.S. data, separate the firms by their growth potential, some find results consistent with the information hypothesis Denis et al., 1994 and Yoon and Starks, 1995, while others conclude that the results are consistent with the free cash flows hypothesis (Lang and Litzenberger, 1989).2 Moreover, in an attempt to distinguish between the hypotheses, these studies take one step further and examine the revisions of analysts' earning forecasts after the announcements of dividend changes. Unfortunately, these tests do not help in finding the elusive consensus either. Denis et al. (1994) and Yoon and Starks (1995) find significant revisions of analysts' earning forecasts following dividend announcements, whereas Lang and Litzenberger (1989) obtain evidence of insignificant revisions. There is clearly a need for a fresh look at these mixed results. Examining dividend announcements in Japan can shed light on this debate. This approach is especially promising because the institutional features of the Japanese market differ substantially from those of the U.S. market. One unique institutional feature in Japan is the industrial organization of the firms. Most companies are affiliated with business groups or keiretsu, and engage in extensive reciprocal shareholding. These companies also have their senior managers on each other's boards of directors. In this context, the majority of the shareholders are likely to have access to the firm's inside information. Moreover, dividend announcements prior to the ex-dividend day are not required in Japan and are voluntary. In a recent paper, Kato et al. (1997) examine voluntary dividend announcements in Japan and document the characteristics of firms that make such announcements. They find that larger firms and keiretsu firms make dividend announcements more frequently than smaller firms and nonkeiretsu firms. More importantly, it appears that firms with greater information asymmetry are less likely to disclose dividend information. Therefore, the motive to make dividend announcements to reveal inside information (the cash flow information hypothesis) is less clear in Japan, where the majority of the shareholders are in reality quasi insiders. Similarly, the desire of managers pursuing their own agendas to invest suboptimally (the free cash flow hypothesis) should also be examined in a different light. When the monitoring of management is vigilant due to intercorporate shareholding and close ties with financial institutions, managers of Japanese firms may adopt a different pattern of investment behavior. Overall, it is not obvious whether any of these hypotheses will be supported by evidence obtained from a study conducted in Japan. Indeed, in a recent article, Dewenter and Warther (1998) suggest that Japanese firms face less information asymmetry and fewer agency conflicts than U.S. firms. This research first follows a standard approach to examine the market reaction to dividend announcements. It tests the competing hypotheses in a simple setting of multiple regression analyses by examining the combined effects of several explanatory variables, which are associated with these hypotheses, on the magnitude of abnormal returns around dividend announcements. Subsequently, this study implements two direct tests of the hypotheses. First, it examines in a univariate analysis the cash flow and investment behavior surrounding dividend announcements. Next, it investigates in a multivariate approach the relationship between the dividend policy and the investment spending while controlling for other components of cash flow from operations and financing. The results of the regression analysis of excess returns indicate that the magnitude of the market reaction to the announcements is positively related to the size of actual dividend change and to Tobin's q ratio, but negatively related to firm size. Contrary to results in the U.S., we do not find a significant impact of dividend yields on the magnitude of the market reaction during dividend announcements. These initial results are consistent with the information hypothesis, but lend no support to the dividend clientele hypothesis suggested by Bajaj and Vijh (1990) or to the free cash flow hypothesis. The results of the univariate analysis indicate that dividend changes are related to cash flow from operation, financing, and investment activities. Firms announcing dividend increases experienced on average higher earnings, reduced debt ratios, and increased investment levels. In contrast, dividend-decreasing firms underwent deterioration in earnings, raised their leverage ratios, and reduced their investment activities. Specifically, dividend changes not only are associated with earnings prospects in the near future but also reflect past and current earnings. These results are consistent with Lintner's (1956) model on dividend policy and can be compared with the findings in Benartzi et al. (1997), where dividend changes reflect the past rather than future earnings changes. Furthermore, the improvement in earnings is accompanied by an increase in investment activities that suggests an increased confidence in future prospects. The results of this analysis are also consistent with the cash flow information hypothesis, with one qualification. Since this analysis focuses on only one dimension of cash flow, it neglects the simultaneous changes in other cash flow components. To account for this shortcoming, we employ a multivariate approach in which we control the available cash flow and the investment opportunities while examining the dividend policy and investment behavior. This approach originates from the growing body of literature that studies the relation between investment and cash flow constraints (e.g., Fazzari et al., 1988, Oliner and Rudebusch, 1992 and Vogt, 1994). The common investment model in the literature is modified in this paper so that we can investigate the relationship between dividend changes and investment spending by firms. We find that firms with more cash flow engage in more investment. This result is consistent with the basic prediction of the free cash flow hypothesis. However, after controlling for the available cash flow and investment opportunities, an increase (decrease) in dividend payment is still followed by an increase (decrease) of the investment level. This evidence is not in line with the prediction of the free cash flow hypothesis. As a by-product of this research, we also gain insights into Japanese firms' investment behavior. Given the specific institutional background of the Japanese market, this study finds that investment is very sensitive to liquidity constraints for nonkeiretsu firms, and less so for keiretsu firms. This result should be compared with the findings by Hoshi et al. (1991), who studied the investment and liquidity of Japanese industrial groups. To summarize the results of the hypotheses testing, this study provides evidence in support of the cash flow information hypothesis. It seems that dividend changes indeed convey information about the firm's cash flows. Although the free cash flow hypothesis is to some degree supported by the evidence in firms' investment behavior, dividend policy is not used by Japanese firms to control the overinvestment problem. In addition, the dividend clientele effect does not appear significant around dividend announcements in Japan. This paper is organized as follows. Data are described in Section 2. 3, 4 and 5 explain and discuss the results of three separate but sequential analyses. Section 3 examines the market reaction to dividend announcements. Cash flow behavior around dividend announcements is presented in Section 4. Section 5 focuses on the relationship between dividend changes and investment spending. Section 6 offers a summary and conclusion.

A popular avenue of research to address the importance of dividend policy on firm value in the U.S. markets is to examine the market reaction to the dividend announcements. Many studies document a positive stock price reaction to dividend initiation or dividend increases and a negative market reaction to dividend cuts or omissions. The reason for these stock price reactions is, however, less clear. Two hypotheses are frequently proposed to explain the market reaction to the announcement of dividend changes: the cash flow information and the free cash flow/overinvestment. Under these hypotheses, dividend changes are considered either a signaling device for future cash flow or a monitoring device for controlling the overinvestment of free cash flow. Each of these hypotheses is supported by various empirical studies in the U.S. markets. To date, there is still no consensus on this issue. This study of dividend announcements in Japan sheds additional light on the continuing debate. In addition to providing new evidence from a different institutional background, we also apply various improved approaches to test directly the competing hypotheses. First, we examine the cash flow and investment behavior around the dividend announcements. Second, we investigate the relationship between dividend policy, available cash flows, and investment behavior. The findings of this study are generally supportive of the cash flow information hypothesis. Although dividend announcements do not appear to be associated with active signaling, the announcements of dividend changes do convey information about the announcing firm's cash flow from operations. Furthermore, dividend changes are not only associated with earnings prospects in the near future but also reflect past and current earnings performance. Interestingly, the results show that dividend announcing firms and the general population of firms exhibit different patterns of investment behavior. It seems that the investment activities of dividend announcing firms are in line with the description of the pecking order theory. In contrast, the investment behavior of firms in the comprehensive control sample is consistent with prediction of the free cash flow hypothesis. Specifically, firms with more cash flow do invest more, and firms with lower q value, which are more likely to be overinvestors, are more subject to cash flow constraints in their investment spending. If we focus, however, on our primary sample of firms that make voluntary dividend announcements, the free cash flow hypothesis does not predict correctly the market reaction to dividend announcements in Japan. Dividend announcements by firms with higher q value incur greater market reaction than announcements by firms of lower q value, which is inconsistent with the free cash flow hypothesis. Furthermore, after controlling for cash flow, investment opportunities, and external financing, an increase (decrease) in dividend usually accompanies an increase (decrease) in investment. Apparently, dividend policy is not adopted actively in Japan as a device to control the overinvestment problems, as the free cash flow hypothesis implies. 33 This last result should come as no surprise considering the institutional background of the Japanese market. The backbone of the free cash flow hypothesis is the agency problem between management and shareholders. In the Japanese industrial structure, the agency problems between management and shareholders is believed to be less severe because of the close monitoring by the major corporate shareholders and the affiliated financial institutions. All of our results for the sample of dividend announcing firms support this belief. Nevertheless, there are some hints for the general population of Japanese firms that suggest that the free cash flow problem might exists in Japan even though dividend policy is clearly not used as a device to control this problem there.