The Effect of Ownership Structures on Dividend Policy: Evidence from Jordan

Undoubtedly, dividend policy is an important component of companies that is in need to be investigated as it is found in literature to be an interesting issue. Therefore, the study purpose is to identify the influence of ownership structures involving family ownership, government ownership, private ownership, foreign ownership as independent variables. In addition, this study includes significant control variables involving firm size, future growth opportunity, free cash flow, and leverage. In fact, these factors added significance to the study. In order to fulfill the research objective, the analysis of descriptive statistics and regression were implemented. The sample of this study included public and private Jordanian companies listed in Amman Stock Exchange and they were 191 companies from 2014 to 2018.The results conclude that there is a strong relationship between the ownership structures, control variables used, and dividend policy. However, some hypotheses are not supported, while two of them are supported.  Ownership structure is an important factor to be always target for researchers and organizations. Besides, future studies are recommended to see other ownership structures such as managerial and institutional ones as they could support such findings and/or add some valuable information. Keywords: Dividend Policy, Family Ownership, Government Ownership, Private Ownership, Foreign Ownership, Firm Size, Future Growth Opportunity, Free Cash Flow, Leverage. DOI: 10.7176/RJFA/11-12-14 Publication date: June 30th 2020

2014-2018. The current study is motivated due to the importance of this relationship and effects of ownership structure and dividend policy as well as their effect on companies. In addition, previous studies found inconsistent results among this relationship. Therefore, the present study aims to identify the causal relationship and implementing independent variables: government ownership, private ownership, family ownership, and foreign ownership, and control variables including free cash flow, future growth opportunity, firm size, and leverage to see their effect on dividend policy as these factors were not included in such study according the researcher's best knowledge. The research outcome is expected to enable shareholders to know which ownership class will impact dividend policy. This study could also be a source to researchers investigating such issue.

Literature review 2.1 Jordan Amman Stock Exchange (ASE)
The Jordan Amman Stock Exchange (ASE) has been taken place since 1997 based on the "Securities Law 1997" being a non-profit institution, 1 with having administrative as well as financial autonomy in order to conducting the trading activities in the Jordanian Capital Market 2 . On 11 March 1999, the ASE has begun its operations in Amman (ASE Annual Report, 1999). It has started these trading activities with 151 companies that had been listed the ASE and with having Jordanian Dinar 4137.7 million as a market capitalization. The ASE had started its operations with including 22 members. There were 62 members on 31 December 2013. 61 members out of 62 ones are seen to be active members in this market. The ASE is still the only stock exchange in Jordanian kingdom (Al-Tal, 2014). The ASE' role is as a facilitative and regulatory body in this market. In doing so, the ASE examines its roles in trading and listing, in securities, ensuring the compliance with the disclosure requirements, and its role in corporate governance in Jordanian Kingdom. The Jordanian listed companies have been divided by the ASE into three sectors, namely, the services, financial, and industrial sectors.
A firm's dividend policies are deemed to be fundamental decisions by companies' managers related to what earning amount will be distributed to shareholders as a dividend as well it is also associated with how much should be retained in order to fostering the firm growth. Additionally, it also includes the managers' decision in which to retain earnings or pay dividends (Ibrahim & Shuaibu, 2016). According to Adesola and Okwong (2009) dividend payment is deemed to be relevant since it has information values and this information content regarding the payment of dividend could send some signals to prospective investors and shareholders; consequently, this can influence the share prices on the stock market of a company. According to Jensen (1986) Jensen and Meckling (1976) the agency theory is considered the most principal theory that have been implemented in a large number of research that included the ownership structures impacting the companies' dividend policy. The agency theory is associated to the conflicts which appears in the company relationships, more specifically, the relationship of principal-agent between the principal of the company "shareholders" and the agent "executives".

Dividend Policy and Ownership Structures
In this regard, Zraiq and Fadzil (2018) investigated the dividend policy, ownership structures, and control relationship and they indicate that family share creates some challenges as it dominates overall the company. This finding was supported by some other study (e.g., Febriani, & Margaretha, 2019;Putri, & Ramli, 2017). In addition, Sakinc and Gungor (2015) and Ibrahim (2016) state that the foreign ownership structure negatively and significantly affects the relationship with dividend payout. They explain that an increase in the foreign ownership reduces dividend payout, in fact, these foreign shareholders do not prefer to distribute the profits of a company to the shareholders or they only distribute low amount in order to utilize the earnings for more investments. As for the government ownership, it was indicated that this government ownership structure negatively affects the dividend payout (Ben-Nasr, 2015). On the contrary, Pham, Jung, and Nguyen (2018) find that government ownership positively impacts the dividend payout. Besides, private ownership negatively impacts on dividend payout policy (Azzam, 2010).
Many studies studied the relationship between ownership structure and dividend policy (e.g., Thanatawee, 2013;Shariff, Salehi, & Bahadori, 2010;Ullah Fida, & Khan, 2012) for example, Thanatawee (2013) investigated the ownership structures and dividend policy relationship over the period 2002-2010 in Thailand. The analysis' findings concluded that a company pays dividends when this company has higher ownership concentrations or when the highest shareholdings are institution and companies distribute many dividends when the largest shareholders, more precisely if institutions hold more equity. This study also indicated that the possibility of the degree of dividend payouts and distributing dividends will increase with higher institutions having individual ownerships. In addition, Shariff ett al., (2010) studied the shareholders ownership identity's influence on payout ratio among the Iranian companies for the period of 2002-2008. The study found that a significant positive correlation between the institutional ownerships and the ratios of payout. Besides, Ullah et al., (2012), found a research to see the ownership structure influence on firms' dividend policy among listed Pakistan stock exchange for 2003-2010. Their study found that institutional ownership is seen to be positively associated with dividend policy among listed companis found in the Pakistan stock exchange.
In addition, several studies investigated such relationship (e.g., Chen, Cheung, Stouraitis, & Wong, 2005;Gonzalez, Guzman, Pombo, & Trujillo, 2014;Wei Wu, Li, & Chen, 2011;Zraiq & Fadzil, 2018). Chen, et al (2005) concluded that a significant negative relationship has been found between both family ownership and dividend payouts and this was rated as up to 10% of the stockholdings of firms and a positive association for shareholding of families ranging between 10 % and 35 % for only small companies in Hong Kong. Furthermore, Gonzalez, et al (2014) carried out a study to see the family involvement effects on dividend policy and how such family involvement impacts the cost problems of agency between minority as well as large shareholders. Their study revealed that family influence varied considerably according to the family involvement type. Additionally, Wei et al (2011) indicated that families having lower cash dividend payouts and lower tendencies to distributing dividends, in China, when comparing to non-family firms. As Zraiq and Fadzil (2018), they investigated the dividend policy and ownership structure and control relationship in Jordanian Kingdom comparing to a benchmark sample of West European companies. The study found that the prominent problem of agency in both regions is seen to be wealth expropriation by controlling shareholders "predominantly the families" from outside shareholders.
Moreover, many recent studies have also investigated the relationship between ownership structure and dividend policy (e.g., Ajadi et al., 2019;Jabeen & Ahmad, 2019;Kautsar, 2019;Tran, & Le, 2019). The present study is still different from the above-mentioned studies as they did not include private ownership, government ownership, family ownership, and foreign ownership and control variables including firm size, free cash flow, future growth opportunity, and leverage as same as our present study. This kind of difference makes the current study distinguished from previous studies. In addition, it can provide to the body of knowledge some values as it includes many variables to identify their effects to dividend policy. Hence, the model of this study will be presented to explain more specifically the current study.

3 Firm Size, Leverage, Future Growth Opportunity, Free Cash Flow, and Dividend Policy.
Control variables are seen to be important while dealing with dividend policy (Lin et al, 2016;Li & Zhao, 2008). Firm size inclusion is presented to serving in the current research as a control variable since it is found that large companies are being expected to paying the dividends to their own shareholders (Redding, 1997). Additionally, these larger companies are also expected to be different in their nature and hence, they are less prone to the bankruptcy. Thus, large companies can boost the companies to paying dividends to the shareholders (Titman & Wessels, 1988). The size of the company can be said that companies with large assets will be easier to enter the capital market, to maintain reputation and that makes the company can distribute larger dividends (Hadinugroho & Handayani, 2009). Pribadi and Sampurno (2012) suggest that company size has a negative and significant impact. But it differs from the study of Hadinugroho and Handayani (2009) which suggests the opposite is positive and significant.
As for the leverage, the greater the company's leverage, it tends to pay lower dividends with the aim to reduce dependence on external funding so that the greater the proportions of debt used for the capital structures of a company, the greater the amount of liabilities that will affect the size of the dividend will be shared (Akhyar Gunawan & Candrasari, 2014). Their study's results concluded that leverage negatively and significantly affected the dividend policy (Akhyar et al., 2014). However, it is different from what was stated by Mahesti Purbandari & Mujilan (2013) which states that debt and dividend payout have a significant and positive impact. Akhyar et al. (2014) indicated that growth significantly affects the dividend payout ratio. According to Imran (2011) a positive effect was found on dividend payouts. In contrast to the research produced by Simanjuntak (2016) argues that company growth negatively affects dividend policy. Rosdini (2009) describes free cash flow as the available discretionary cash flow to firms. It can be utilized for any discretionary uses including capital and acquisition expenditure with growth orientation, shareholders' payments as dividends, and debt payments. The greater the free cash flow the firm has, the healthier the firm will be since it has cash available for growth, debt payments and dividends. So, the higher the free cash flow the company has, the higher the dividend the company can pay, because the company has enough cash available. Research conducted by Rosdini (2009) indicated that free cash flow positively affects the ratios of dividend payout. Kouki and Guizani (2009) and Lucyanda and Lilyana (2012) who also found free cash flow positively affects the dividend policy. Jensen and Meckling (1976) indicate that the agency theory is based on a relationship which is known to be an agreement. In this agreement the principal "owner" implements another party that is known to be "the agent" acting on his/her behalf. However, the controlling power is in the managers' hand but the ownership itself is with the principal. Accordingly, the agency problem can be arisen between shareholders and managers/agent when these agents take some decisions which are seen not to be beneficial to shareholders but based on the self-interests Research Journal of Finance and Accounting www.iiste.org ISSN 2222-1697 (Paper) ISSN 2222-2847 (Online) Vol.11, No.12, 2020 they have. For example, these managers "agents" might decide to expand the firm size beyond the optimal capacity the company has since this gives good impressions about the performance of managers as the rewards they gain are tied to it. That is to say, this will hinder the dividend payment to the principals "shareholders". In contrast, the dividend payment to the common equity shareholders reduces the excess of the available free cash flow with managers, in so doing, mitigating agency problems between both shareholders as well as agents (Jensen & Meckling, 1976;Rozeff, 1982: Easterbrook, 1984. Moreover, Jensen (1986) confirms that companies are expected to paying more dividends when these companies are highly monitored. It could be apparently observed that firms having large holdings monitor agents well and this leads to a higher payout. Accordingly, the interests' conflict between the principals and agents could be solved by dividend policy payout.

4 Agency theory
Based on this, organizations consist of a nexus of contracts between agents and owners of those economic resources (Jensen & Meckling, 1976). This agency theory postulates that principals have less information than agents and this information asymmetry influences adversely the principals' abilities to supervise whether or not their interests are being duly served by agents. Additionally, an agency theory premise is that agents as well as principals rationally act and utilize a contract to maximizing their own wealth. Consequently, Jensen and Meckling (1976) reveal that the assumption might be the "moral hazard" problem, meaning that in an attempt to maximizing their wealth; agents might confront the dilemma of working against their principals' interests. This is because the principals have no access to the most of available information when decisions are made by an agent; they are incapable to decide whether the practices of managers are deemed to be in the best organization's interests.
By referring to the ownership concentrations, the agency theory builds up two important perspectives on the ownership concentrations' as well as the dividends' relationship as follows: the first perspective as stated by Harada and Nguyen (2011) shows that the interest's conflict appears between the owners and the managers of a company could be lessened thru a concentrated ownership. These kinds of predicts shows that concentration ownership positively impacts dividends (e.g., Abdullah, Ahmad, & Roslan, 2012). The second perspective of such relationship displays that owners' concentration prefers to have more private benefits from existing free cash flow and this leads to lowering the payments of dividends (Harada & Nguyen, 2011). Hence, a negative relationship between concentration ownership and the dividend payments is predicted.

Hypothesis and Methods
In this study the decision to accept or reject the hypothesis is based on statistical significance. Statistical significance considered as a description for a result or experiment when the probability is less than the significance level. The current study aims to examine the following hypotheses: H1: Ownership structure type will have a significant effect on the type of the dividend policy applied H2: Control variable type will have a significant effect on the type of the dividend policy applied

Procedure Methodology
This study aims to measure the ownership structure impact on dividend policy with paying attention to the control variables presented in the model. This study is going to adopt explanatory research approach as this study aims to clarify the causal relationship of research variables thru testing particular hypotheses. Sources of research data employed will be secondary data or indirect data. The population of the present study is public and private firms in Jordanian Kingdom and the sample size of the present study is 191 private and public firms listed on the Jordan Research Journal of Finance and Accounting www.iiste.org ISSN 2222-1697 (Paper) ISSN 2222-2847 (Online) Vol.11, No.12, 2020 Amman Stock Exchange, in the observation period from 2015 to 2019. The sample included the all companies registered.

1 Model Specifications
The analysis technique implemented in the present study is simple and multiple linear regression by using SPSS. Analysis of the data used is the descriptive statistics, simple and multiple regression analysis, and hypothesis testing. Regression analysis is used to predict of dependent variable based on the scores of the independent variables, also Correlation Matrix shows the direction of the correlation between the dependent variable and the independent variable (Ghozali, 2011). Simple and multiple Regression models are as follows: DP = α + β1 Family Ownership + е (1) DP = α + β1 Government Ownership + е (2) DP = α + β1 Private Ownership + е (3) DP = α + β1 Foreign Ownership + е (4) DP = α + β1 Family Ownership + β2 Government Ownership + β3 Private Ownership + β4 Foreign Ownership + β5 Firm Size + β6 Future Growth Opportunity + β7 leverage + β8 free cash flow + е (5) As for descriptive statistics, it will be used to present the descriptive analysis of the dependent variables and independent variables. As it aims to see the average of companies' distributions for dividend, additionally, to see the minimum, a maximum, the standard deviation of data collected.

Data Analysis
The methods of data analysis implemented in this study are the descriptive statistics including, maximum, minimum, mean and standard deviation of variables. In addition, regression analysis is used to see the relationship forms of free variables included in structure ownership, namely, family ownership, government ownership, private ownership, and foreign ownership towards the "dividend policy" as a dependent variable and the control variables.

Results and Discussion
The descriptive statistics is being used to describe some important characteristics of variables (see table 2).
Research Journal of Finance and Accounting www.iiste.org ISSN 2222-1697 (Paper) ISSN 2222-2847 (Online) Vol.11, No.12, 2020 Table 2 illustrates that the dividend has a mean as 28.9 among Jordanian companies. The minimum value of dividend is 0.0 and the maximum one is 193.23. As for the family ownership, the mean is 21.6. The minimum is 0.0 and the maximum is 94.0. The government ownership mean is 4.38, with 54. As a maximum and 0 as a minimum. The private ownership mean is 13.87, while the maximum is 86 and the minimum is 0.0. The foreign ownership mean is 7.69. The maximum is 98.7 and the minimum is 0.0. In addition, the firm size mean is also presented. It is 17.03. The free cash flow mean is .917, and the future growth opportunity is 14.50; while the leverage is 35.44.

5.2.1Independent Variables and Dependent Variables:
In this section, the study aims to present the relationship between each independent variable and the dependent variable. This relationship is explained by employing the regression test of each variable. 5.2.1.1 Family Ownership and dividend per share: Table 3 explains the relationship between these two variables. This test is used to accept or reject the hypothesis. In fact, table 4 is implemented for the same reason. However, implementing each independent variable with the dependent variable could keep the same finding or it may have other changes. Accordingly, table 6 does not support the hypothesis and it does accept the null hypothesis which indicates that there is not significant relationship between family ownership and dividend per share.  Table 4 explains the relationship between these two variables, the result supports the hypothesis as it shows that P-Value is below P<0.05 indicating to a significant relationship between government ownership and dividend per share.  Table 5 explains the relationship between these two variables, the result illustrates that there is no significant relationship between private ownership and dividend per share.
Research Journal of Finance and Accounting www.iiste.org ISSN 2222-1697 (Paper) ISSN 2222-2847 (Online) Vol.11, No.12, 2020 Table 6 explains the relationship between these two variables, the result demonstrates that foreign ownership and dividend per share have no statistical relationship. Consequently, these dependent variables still have provided the same result even though they are tested individually with the dependent variable. The model in this study has been estimated by implementing a random effect method. Table 7 displays the relationship's findings between dividend per share as the "dependent variable" and the independent variables, namely, family ownership, government ownership, private ownership, and foreign ownership, and the control variables is also included containing firm size, future growth opportunity, free cash flow, and leverage, in order to achieve the main research objective. This model is seen to be statistically significant (see table 7) and it fits at the 0.05 % level with the P-value = 0.000. The R 2 = 0.182 in which it indicates that 18.2 % of the depended variable could be explained by the independent variables as well as the control variables chosen in the present research. The next section of regression analysis explains this kind of relationship of variables.  Table 8 shows the relationship of variables. Based on this table, the study shows that there is no significant relationship between dividend per share and family ownership because the P-value is seen to be above p<0.05, it is 0.343. This kind of finding rejects the hypothesis which says that there is a relationship between the family ownership and the dividend policy. On the other hand, the government ownership has a significant relationship with dividend policy as the P-value is below p<0.0.5. It is 0.03, hence, the hypothesis that says there is a relationship between the government ownership and the dividend policy in the private and public companies which are listed in the local stock market in Jordan; is significant. As for the private and foreign ownerships, the results show that there is no significant relationship between these two independent variables private ownership and foreign ownership and the dividend per share. Consequently, the two hypotheses of these variables are rejected.
Research Journal of Finance and Accounting www.iiste.org ISSN 2222-1697 (Paper) ISSN 2222-2847 (Online) Vol.11, No.12, 2020 Furthermore, this table illustrates the relationship of control variables and dependent variable. In doing do, it is seen that the future growth opportunity and dividend per share has a significant relationship, hence, the hypothesis, which says there is a relationship between the future growth opportunity and the dividend policy, is significant; while the other control variables, namely, firm size, free cash flow, and leverage do not have a significant statistical relationship with dividend per share. Accordingly, the three hypotheses of these variables are rejected since the P-value is above 0.05.
Understanding that relationship is useful because we can use the value of one variable to predict the value of the other variable. Accordingly, table 9 displays the results of Pearson's correlation for the independent and dependent variables included in this study, these results reveals that there are negative significant between independent variables, however, is no high correlated.

Conclusion
The study aimed to investigate the structures of ownership impact dividend policy of Jordanian firms that are listed in Amman stock exchange. The findings show that the independent variables and control variables have a significant relationship with the dependent variable. This was proven by the regression analysis which shows that P<0.05 and this proves such significant. The following two hypotheses are only accepted "there is a significant relationship between the government ownership and the dividend policy" and "there is a significant relationship between the future growth opportunity and the dividend policy". Not to mention that dividend policy is a significant factor in companies and more attention should be always paid to such factor. That is to say, future studies are recommended investigating how being this variable treated by having many factors around its own stability.  191 191 191 191 191 191 191 191 191 **. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).