Not logged in.

Contribution Details

Type Master's Thesis
Scope Discipline-based scholarship
Title Can Dividends Predict the Future? A closer look at the Swiss Equity Market
Organization Unit
Authors
  • Claudio Läuchli
Supervisors
  • Michel Habib
  • Jacqueline Haverals
Language
  • English
Institution University of Zurich
Faculty Faculty of Business, Economics and Informatics
Number of Pages 63
Date 2017
Zusammenfassung Over the last 70 years, the dividend policy was amongst the most researched topics in corporate finance. Even though the literature is extensive, if one compares the findings of the major research, it becomes clear that many questions have not yet been answered conclusively. The main question remains the same: Why do companies pay dividends? Black (1976) named this riddle famously ‘The Dividend Puzzle’. Before Miller & Modigliani’s revolutionary work, most researchers argued that dividend payments were directly linked to the value of a firm. The higher the dividends, the more a firm is worth. Williams (1938) for example concluded that a stock is only worth the present value of all its future dividend payments. Gordon (1959) shared these beliefs and constructed a firm valuation model based only on the size of the dividend payments and a rate of discount. In a widely recognized study, Lintner (1956) provided proof that most firms smooth dividends over time. He further found that the management does not set dividend payments every year anew, they rather use last year’s dividends as a benchmark and subsequently decide on whether dividend payments should be changed. Unhappy with this approach, Miller & Modigliani (1961) proved in a theoretical paper that as long as a firm’s investment policy is unchanged, changing the payout policy does not influence the value of a firm. For this theory to work, they had to make the critical assumption that capital markets are perfect. A lot of research was dedicated to examine Miller & Modigliani’s dividend irrelevance hypothesis, but no unambiguous results could be found. This is mainly due to the strict assumptions Miller & Modigliani had to make in order for their theory to work. Those assumptions have to be relaxed when testing empirically. Taxes, agency restriction, clientele effects and other investor biases all do occur in the real world, and they all do have an influence on the payout policy. Another possible answer to Black’s question may be given by the dividend signaling theory. This hypothesis is based on Miller & Modigliani’s (1961) and Watts’ (1973) discovery that dividends may contain information. More precisely, the dividend signaling theory states that managers use dividend changes to communicate their assessment of the future of the firm to the markets. Dividend increases should therefore be followed by increasing future profits and vice versa. Bhattacharya (1979), John & Williams (1985) and Miller & Rock (1985) provided theoretical proof for this theory. But the empirical results were everything but conclusive. Watts (1973), DeAngelo, DeAngelo & Skinner (1996), Benartzi, Michaely & Thaler (1997) or Grullon, Michaely, Benartzi & Thaler (2005) found only weak support for the hypothesis. Nissim & Ziv (2001) investigated the relationship between dividend changes and future profitability. After adapting the models used in previous studies, they find strong support for the dividend signaling hypothesis. The studies mentioned before are just a small selection of the research conducted on the payout policy and the dividend signaling hypothesis. The mostly mixed and contradicting results show the complexity of the subject and subsequently provided the main motivation for this thesis: to examine, if changes in dividend payments do have predictive power and can predict the future earnings of Swiss firms. Most previous studies have been conducted in US or EU markets. As each market has its own specialties, such as taxes and regulations, it will be worthwhile to analyze, how the Swiss stock market responds. The main regression follow Nissim & Ziv (2001). A cross-sectional fixed effects model was used to account for time-invariant effects, such as the legal environment, management or shareholder-structure. The key predictor variable in this study will be the yearly changes in dividend payments and the dependent variable will be the annual change in earnings. Additionally, several control variables will be added to strengthen the results. In this study, companies listed in the UBS 100 Index will be examined over the time period from 01.01.2011 to 31.12.2016. The UBS 100 Index represents the 100 largest stocks in the Swiss Performance Index (SPI) measured by market capitalization. This will provide a broad overview over the publicly traded shares in Switzerland. Data was excluded from the sample, if the company was not represented in the index during the whole time period. In total, this study covers a sample of 92 companies over six years. In the second part, financial institutions will be excluded from the study. This is done to increase the homogeneity amongst the firms, because these sectors are highly regulated. This extra oversight should theoretically reduce the agency costs and consequently reduce the significance of dividend payments. Only very weak support for the dividend signaling theory was found in this study for Swiss firms. If financial institutions are excluded, a positive significant correlation can only be found for the dividend increase dummy for the first year following the dividend change. In the second year after the change, no significant relationship can be detected. Similar results emerge, after multiplying the dummies with the dividend change ratio. Neither in the first, nor in the second year after the change can any significant estimates for the dividend changes be found. All in all, dividend changes cannot predict future earnings changes in Switzerland. The results suggest, that dividend changes follow past earnings changes, rather than predicting future earnings changes. The data further reveals, that Swiss firms are very reluctant to decrease dividends, as only 29 dividend decreases occur in the time frame of this study. Both those findings support Lintner’s (1956) view that executives raise dividends very cautiously and only when they can be paid permanently.
Export BibTeX