Comparing Dividend Puzzle Solutions by Polish, Canadian, Norwegian and American Executives

This study compares firms’ dividend policy antecedents in four countries. The author surveys managers of 230 companies listed on the Warsaw Stock Exchange that paid dividends over the 2001-12 period and compares the findings to survey reports from the USA, Canada, and Norway. The main antecedents of dividend policy in these countries are the level of current earnings, their stability over time, the level of expected earnings and the pattern of past dividends. The last factor is of little importance for Polish managers, as the history of dividend payouts in Polish-listed companies is brief. The establishment of factors that shape the dividend policy in public companies has crucial importance in global financial markets. Investors make cash transfers to markets that ensure a high return on invested capital. Orders from foreign investors generate almost 50% of the turnover on the Warsaw Stock Exchange. Therefore, in their decisions on the division of earnings, Polish companies should be guided by similar considerations as those displayed by companies in developed economies that seek to attract foreign investors. The study expands on the existing survey research on dividends and provides new evidence from managers of companies in Poland.


Introduction
Dividend policy plays a key role in the distribution of listed companies' net profits, but despite the existence of numerous studies, Brealey, Myers, Allen and Mohanty (2011) list dividends as one of the ten unresolved problems in contemporary corporate finance. The problem of decision-making with regard to dividend payments, described in the literature as the "dividend puzzle" (Black, 1976;Baker, Powell, & Veit, 2002), emphasizes the role of three forces that influence dividend policy: managers, shareholders, and potential investors. Baker, Singleton and Veit (2011) assume that a major reason for this ongoing debate is the heavy reliance on economic modeling approaches and the lack of an in-depth understanding of how investors and managers behave and perceive dividends. Dividend policy continues to attract attention due to its linkage with corporate financing and investing decisions and its impact on shareholder wealth (Baker & Jabbouri, 2016).
The Polish capital market is a unique case in the study of corporate dividends. Since the start of the reg-exchange, the proceeds from which served to replenish the state budget. However, the Treasury continues to be the majority shareholder in strategic companies in the fuels, energy and mining sectors. Faced with a budget deficit, the Treasury expects to receive dividends from the companies in which it has a stake. Hence, these companies are the main dividend payers in the Polish market (Sierpińska-Sawicz, 2014).
In studying dividend policy, researchers typically rely on two main approaches: managerial surveys and statistical analyses of published financial data. Survey research complements research based on secondary data and provides additional insights into why firms make certain dividend policy decisions (Baker & Weigand, 2015). Data on Polish companies listed on the regulated market are distorted by high inflation and interest rates, especially in the early years of privatization. Therefore, these data are difficult to apply in comparisons with long-term data from highly developed markets. In addition, compared with exchanges in mature economies, the Warsaw Stock Exchange   Lintner (1956) is the first to initiate manager surveys applied to dividend policy studies. His seminal research on US executives reports that the basic premises for dividend payments are the pattern of past dividends and the current level of earnings. Baker and Powell (2000) (Baker, Saadi, & Dutta, 2008). Lintner (1956) reports that an industry effect may influence firms' corporate dividend policy and investor perceptions of dividends for firms in different industries. For example, firms operating in saturated and mature industries have a higher propensity to pay dividends than do firms in high-growth industries. Based on research in the USA, Canada, the UK, Germany, France and Japan, Denis and Osobov (2008), conclude that the propensity to pay dividends is connected to firms' size, growth opportunities and profitability and is higher among larger, more profitable firms. Baker, Mukherjee, and Paskelian (2006)

Sample and survey
The research into the antecedents of dividend policy was conducted among a group of companies listed on the regulated market of the WSE that paid out dividends in the period from 2000-2012. The author assumed that companies that had not paid out dividends for over a decade would not be able to earnestly determine the factors that influence dividend policy. In the second stage of the study, the author sent out a paper version of the survey and contacted the companies by telephone to ask them to complete the survey. In addition, the data collection procedure included an offer to send the companies the aggregate survey results so the companies could compare their own dividend policy with the results for the rest of the industry. As a result, over a three-month span, an additional 48 surveys were received, which resulted in a total of 84 complete questionnaires and a total response rate of 36.6%. In line with the pertinent literature, this sample is deemed sufficient to verify the research hypotheses. For example, some studies feature the following response return rates: 32.9% for the NYSE (Baker, Powell, 2000), 29.8% for the NAS-DAQ (Baker et al., 2001), 27.3% for the Oslo Stock Exchange (OSE) (Baker et al., 2005) and 35.4% for the TSX (Baker et al., 2007).  Denis and Osobov (2008) argue, the propensity to pay dividends is higher among larger, more profitable companies whose retained earnings account for a large part of the value of their property. Similar reasoning appears in earlier comparative studies (Baker et al., 2007).

Statistical tests
To test the first hypothesis, the author calculated the arithmetic mean ( ) and standard deviation (s d ) of each of the 26 factors influencing the companies' dividend policy. The author used a t-test for the null hypothesis that the mean response for each of the 26 factors influencing dividend policy equals 0 (no importance). For each factor, the author verified the null hypothesis by testing for the absence of significant differences between the arithmetic mean and the value of 2.0 (average weight of the 5-point scale). This comparison indicates whether the factor is relatively less or more important than the validity referred to as "average".  Table 2 shows the results of the ranking of the factors affecting firms' dividend policy in Poland and the results of the rankings of the same factors by Canadian, Norwegian and American authors (Baker et al., 2000(Baker et al., , 2001(Baker et al., , 2006(Baker et al., , 2007.

Comparison of the hierarchy of the antecedents of dividend policy in the compared countries -survey results and discussion
In Poland, the most important factors determining dividend payouts are the level of current earnings (F3), their stability over time (F1) and the level of expected future earnings (F4). The data in Table 2  In Canada, the three most important factors are the level of expected future earnings (F4), the stability of earnings (F1) and dividend payouts in previous years (F2). In Poland, the last factor was considered to be of average importance (insignificant difference between the arithmetic mean and the value of 2.0) and was ranked fifteenth place, compared to eleventh for the OSE. The weight that managers attach to future earnings is rational because this factor is strongly con- Norwegian managers' approach to regulations and legal restrictions on the payment of dividends (e.g., equity depletion) also differs. This factor is ranked sixth in Norway, compared to its much lower ranking by Polish companies (20th out of 26), and its rank of fifteenth on the NYSE, nineteenth on the TSX and twelfth on the NASDAQ. The difference stems from regulatory differences between the countries. As mentioned previously, Norway's centralized government (Baker et al. 2006)

Summary and conclusions
To summarize the discussion of the hierarchy of factors shaping the dividend policy for firms in Poland, the USA, Canada and Norway, one first observes that this hierarchy does not differ significantly. The most important antecedents of dividend policy for compa-