The Determinants of Dividend Payout: Evidence from Private Banks in Ethiopia

The main objective of this study was investigating factors that determine the dividend payout of private commercial banks in Ethiopia under the study period covering from 2009 to 2017. A balanced panel data set from national bank of Ethiopia and ten private commercial banks annual report was used and analyzed through fixed effects panel data regression technique. Dividend payout was used as a dependent variable and the independent variables covered under this study were both bank specific (profitability, leverage, liquidity, size and last year dividend payout) and macroeconomic factors (inflation rate, real GDP growth rate and foreign exchange rate). The finding shows that from bank specific factor; profitability, liquidity, bank size and last year dividend payout have statistically significant positive effect on dividend payout while leverage shows insignificant positive effect for dividend payout. From macroeconomic variables inflation rate have a positive effect, but the real GDP growth rate, and foreign exchange rate have a negative effect and foreign exchange rate is the only significant macroeconomic factor for dividend payout.The results of this study have delivered some insights on the effects of both bank specific and macroeconomic factor for dividend payout of private commercial banks in Ethiopia and managements and board of directors of those banks need to consider these variables while designing their dividend payout. In addition, government body specially financial sector should have to consider the effct of macroeconomic variables on dividend payout when they are making a policy about macroeconomic issues.


Introduction
Banks in Ethiopia are an essential part of our economic system. The owners of those banks are Shareholders, and the bank distributes a portion of its earnings to those shareholders which is called dividend. Ishtiaq (2016) described it as the best way of communication for banks to its shareholders.
In banking business, finance managers mostly face two operational decisions at the beginning. What real assets the firm should acquire? (Investment decision) and how these assets should be financed? (financing decision). However, another decision is raised when the firm begins to generate profits which is concerned with whether the firm distributes proportion of earned profits in the form of dividends or should reinvest back into the business (Al-Malkawi, Rafferty, & Pillai, 2010;Alam & Hossain, 2012). Such a decision of the firm about how much earnings should be distributed, how stable the distribution should be, and how much should be retained is the a dividend payout decision (Chekole, 2016).
Different shareholders have different interest in relation to dividend payout. According to Dhanani (2005) certain shareholders may prefer cash dividends while others may prefer capital gain. Therefore, based on the shareholders' preferences, managers and board of directors of the company should decide carefully that how much amount of earnings should be distributed to shareholders and how much portion of earnings should be reinvested in the business. Researchers in corporate finance have developed several theories to explain the dividend payout of firms and their determinants. Profitability, liquidity, risk, leverage, age, financial performance, size, previous dividend and capital adequacy are included in most of prior literatures. But the finding relation to the effect of those variables on dividend payout is inconsistent across different researchers.
Lintner (1956) who found that dividend payout are the function of a firm's profit and last year dividend. Who was also supported by different researchers ( Pruit &Gitman 1991, andGill et al., 2009). But in contrary to this Elias (2015) found the profitability effect for dividend payout is insignificant and Chekole (2016) said it has negative significant effect. Muhammed (2012) founds liquidity as a significant positive factor for dividend payout, but Simegn (2013) found it in the other sign and Elias (2015); Chekole (2016) and Henok (2016) stated it as a negative and insignificant factor for dividend payout of banks. Higgins (1981) state that firm size has a significant positive effect, but Elias (2015) and Henok (2016) found out its effect is negative and significant.Collins, Saxena & Wansley (1996) found that financial leverage affects firms' dividend payout decision, while Elias (2015) found it has a negative insignificant effect.
Since there is continuous change in globalization, regulation, parallel competition and volatile market dynamics, the factor affecting dividend payout in today might differ from the factors for yesterday. So conducting a research on such issue is still important. In Ethiopia different academicians conduct their research on the www.iiste.org ISSN 2222-1697 (Paper) ISSN 2222-2847 (Online) Vol.12, No.3, 2021 24 determinants of commercial bank dividend payout such as; Theodros (2011); Simegn (2013);Elias (2015) and Chekole (2016). But as to the researchers' knowledge all of them concentrate on bank specific factors of dividend payout. That is why the researcher in this research tries to fill such a gap in the literature by incorporating additional macro-economic factors such as inflation rate, real GDP growth rate and foreign exchange rate.

Research objective 1.General Objective
The jeneral objective of the study is to investigate the major determinants of the dividend payout in case of private commercial banks in Ethiopia.

Specific objectives
 To investigate the effect of bank specific factors over the dividend payout of private commercial banks in Ethiopia.  To investigate the effects of macroeconomic factors over the dividend payout of private commercial banks in Ethiopia.

Hypothesis of the study
Research hypothesis are developed mainly by reviewing different literature so as to support the theory by empirical evidence. The researcher reviewed several imperical literature which is presented in the literature review part of this paper Therefore, as per the review of prior literature related to the study area, the researcher has formulated the following research hypothesis.
Ho. There is a significant positive relationship between profitability and dividend payout of private commercial banks in Ethiopia.
Ho.There is a significant negative relationship between leverage and dividend payout of private commercial banks in Ethiopia.
Ho.There is a significant positive relationship between liquidity and dividend payout ofprivate commercial banks in Ethiopia.
Ho.There is a significant positive relationship between Bank size and dividend payout of private commercial banks in Ethiopia.

Ho.
The last year dividend has a positive significant relationship with a dividend payout of private commercial banks in Ethiopia.
Ho.There is a significant negative relation between Inflation rate and dividend payout of private commercial banks in Ethiopia.
Ho.There is a significant positive relationship between real GDP growth rate and dividend payout of private commercial banks in Ethiopia.
Ho. There is a significant negative relationship between Exchange rate and dividend payout of private commercial banks in Ethiopia.

Research Methodologys 2.1. Research Design
The major objective of this paper is to investigate both bank specific and macroeconomic factors affecting dividend payout in Ethiopian private commercial banks. So the researcher follow quantitative approach and a primarily explanatory type of research is used to establish a relationship between a number of explanatory variables and dividend payout. On the other hand before testing the relationship between the dependent and independent variables, all the variables included in the study need to be identified and presented. Therefore descriptive type of studies is combined with explanatory study. According to Marshal (1996) the goal of quantitative research is to test a pre-determined hypothesis and to produce general results. By means of statistical methods, the results of quantitative analysis can prove or disprove hypotheses. Conclusions made from the analysis of quantitative data show how many are affected and where is the greatest area of impact.

Population and sample size
In this study a total population are all private banks in Ethiopia and a non-random sampling technique was applied, which is purposive or judgmental sampling based on the availability of 9 years data from all private commercial banks. List of banks that are included in the sample are presented in the following table clearly.Those sample present 62.5% of total population. Zemen BankS.C 2008 10 10 Birhan International Bank S.C 2009 9 Number of sample(10) /total population(16) =62.5% 0f total population Source: National bank of Ethiopia , 2017

Data Types and Source
In this study, the researcher use 10 private commercial bank financial data and national bank of ethhiopia (NBE) annual reports for subsequent 9 recent years, which covers the period from 2009 to 2017. Therefore, the researcher used secondary data only. Published secondary data were collected from selected private banks and from national bank of Ethiopia. In this the annual audited financial report of national bank of Ethiopia and some older private commercial banks are collected from their head office.

Method of Data Analysis
Descriptive statistics like mean, minimum, maximum and standard deviation were used to organize, summarize, and describe observations and to compare variables numerically. Second Correlation analysis is used to measure the degree of association between the dependent and independent variable. It also used to indicate the direction and the strength of association between independent and dependent variables (Simegn, 2013). Finally the regression analysis is used to see the relationship between the explanatory variables and the dependent variables and to test the hypothesis. The analysis was accomplished by using quantitative data analysis tool STATA version 12 software.

Table2.2:Variable description and their expected sign No
Variable Symbol Description Ex-sign 1 Profitability PRO PAT/shareholders' equity +ve 2 Leverage LEV Total debt/total asset -ve 3 Liquidity LIQ Current asset/current liability +ve 4 Bank size SIZ Natural logarithm of total asset +ve 5 Last dividend payout LDP Last year dividend payout rate +ve 6 Inflation rate INR Annual inflation rate by NBE -ve 7 Real GDP growth rate GDPGR GDP last-GDP this/GDP last year +ve 8 Exchange rate EXR annual exchange rate Birr/USD -ve Source: prior researchers (Mundati, 2011;Theodros, 2011;Elias, 2013, knife, 2015and Chekolle, 2016

Model specification
The panel regression equation differs from a regular time-series or cross-section regression by the double subscript attached to each variable. The general form of the panel data model can be specified more compactly as: From this i, represent the cross-sectional dimension t, representing the time-series dimension Yi,t represents the dependent variable ratio Xi,t contains the set of explanatory variables; and αi is constant over time t and specific to the individual cross-sectional unit i. Thus dividend payout rate is expressed as a function of bank specific and macro-economic factors.
, , , , , , Where PRO, represents profitability LIQ, represents liquidity LEV, represents leverage SIZ, represent bank size LDP, represents last year dividend payout INR, represent inflation rate GDPGR, represent real GDP growth rate FER, represents foreign exchange rate β1 -β8, represents regression coefficients for each explanatory variable ai,represents the intercept of the regression equation i, represents the cross-section dimention (banks) t, represents the time series dimention (years) represents the error term which accounts for variables affecting DPO that are not included in the model. Table 4.1 provides a summary of the descriptive statistics of the dependent and independent variables for ten private commercial banks from year 2009 to 2017 with a total of 90 observations. The table shows the mean, median, standard deviation, minimum and maximum for the independent and dependent variables used in this research. It shows the average indicators of variables computed from the financial statements of each bank and national bank of Ethiopia annual report. The above table shows the degree of correlation/association between one dependent variable (dividend payout) and eight explanatory variables (profitability, leverage, liquidity, size, last year dividend, inflation rate, real GDP growth rate and foreign exchange rate). The result shows that all bank specific variables are positively correlated with dividend payout. This indicates that when bank specific variables increase dividend payout moves in similar direction. In the other case from macroeconomic variables inflation and real GDP growth rate are negatively correlated with dividend payout. This indicates that when those macroeconomic variables increase, dividend payout goes in the opposite direction. In the other side foreign exchange rate shows a positive correlation with dividend payout. There for as indicated from the table the p-value is above the 5% significant level. There for we have no reason to reject the null hypothesis stated as there is no serial correlation.

Test for multi-collinearity
The values of the correlation coefficient are always between -1 and +1. A correlation coefficient of +1 indicates that the two variables are perfectly related positively; where as a correlation coefficient of -1 indicates that two variables are perfectly related in a negative linear sense. A correlation coefficient of 0, on the other hand indicates that there is no relationship between two variables (Gujarati, 2004).  Furthermore, the variance inflation factor (VIF) can show the existence of a high degree of multicollinearity. According to Gujarati (2004) the benchmark to say there is a problem of multicollinearity among explanatory variables is when the mean value of variance inflation factor for each explanatory variable should be above 10. So, according to the result the measure of VIF shows that each variable has quite below 10 and mean variance inflation factor is 1.68. See the table below.

.4. Hausman specification test
After the basic classical linear regression model assumptions are satisfied, the researcher goes to select the appropriate panel data model by using the Hausman specification tests. The Hausman specification test is a test used to choose between the fixed effect model and random effect models. According to Hausman (1978) the null hypothesis is that the favoured model is a random effect that assumes unobserved variables are not correlated; the alternative hypothesis is that the preferred model is fixed effect. This test recommends the application of Random effect (RE) panel regression if the test statistics (chi2) p-value is above 5%. Otherwise, the fixed effect model (FE) also called within regression is a good estimation tool.  ISSN 2222-1697(Paper) ISSN 2222-2847(Online) Vol.12, No.3, 2021 29 As presented in table Hausman specification test was providing evidence in favor of the fixed effect model with p-value of 0.0000 which is less than 5% and this supports the fixed effects (FE) method is an efficient estimator for the panel data models.

Profitability
Profitability is found out as a significant explanatory variable for dividend payout at 5% significant level with pvalue of 0.0000, which has a positive coefficient of .7266785. This indicates holding other things constant a 1 birr increase in profitability of banks will lead to a 72.66 cents increase in current dividend payout and the reverse is true.This finding is supported by the work of (Pruitt & Gitman, 1991;Gill, et al., 2009;Muhammed, 2012 andSimegn, 2013)In the other side the results of this study is not consistent with the finding from (Ferris, et al., 2003;Al-Ajmi, 2008 andTheodros, 2011).The result is consistence with the researcher early expectation. Therefore the researcher haven't enough reason to reject hypothesis 1 stated as there is a significant positive relationship between profitability and dividend payout of private commercial banks in Ethiopia.

Liquidity
Liquidity has a positive and statistically significant relationship with dividend payout at p-value of 0.034. The coefficient for liquidity is 0.1193412. This indicates that holding other variables constant a 1 percent increase in liquidity leads to an increase by 0.12 percent to dividend payout of the sampled private commercial banks in Ethiopia and the reverse is also true.The result from this study is consistent with the researcher expectation. Therefore the researcher haven't got enough reason to reject hypothesis 3, which stated that there is a significant positive relationship between liquidity and dividend payout private commercial banks in Ethiopia.The finding of this study is consistent to Theodros (2011) and the Jensen's (1986) agency theory which stated that companies with higher free cash flow have higher dividend payout ratios. In the other side there are also a number of studies that have showed a negative or insignificant relationship between liquidity and dividend payout (Imran, 2011;Kinfe, 2011;and Maladjian & El Khoury, 2014).

Bank size
size is significant factor for dividend payout at 5% significant level with a p -value of 0.027. The coefficient of bank size is 0.0471476. This shows that holding other variables constant a one percent change in bank size leads about 0.047 percent changes in dividend payout of sampled private commercial banks in Ethiopia. The finding is consistent to the researcher's early expectations. Therefore there is no enough reason to reject the null hypothesis 4, which states that there is a positive relationship between bank size and Ethiopian private banks dividend payout. ISSN 2222-1697(Paper) ISSN 2222-2847(Online) Vol.12, No.3, 2021 30 The finding of this study is consistent with Theodros (2011) who found out firm size is statistically significant determinant of dividend policy. In the other side in contrary to this finding Elias (2015) suggests that firm size has a negative and statistically significant relationship with dividend payout. In addition to this Medhe et al. (2010) found out size is not a statistical significant factor for firm's dividend payout.

Last year dividend
last year dividend payout has a positive and statistically significant relationship with dividend payout at 5% significance level with p-value of 0.001. The coefficient for last year dividend is 0.2791758. This indicates that holding other variables constant a 1 birr dividend paid in previous year will lead to a 27.9 cents increase in current dividend payout. This indicate that the sampled private commercial banks in Ethiopia paid a consistent dividend for their shareholders. Therefore the researcher haven't enough reason to reject the hypothesis 5, stated that last year dividend has positive and significant relationship with dividend payout of private commercial banks in Ethiopia.The finding is consistent with Lintner, 1956;Bhattacharya, 1979;Javid, 2009;Theodros, 2011;Simegn, 2013 andHenok 2016) those researchers found out liquidity as a positive significant factor for dividend payout. It is also in line with signaling theory of dividend, where companies want to give a positive signal to the market that the company is in good condition by continuing paying dividends.

foriegn exchange rate
Foreign exchange rate has p-value of 0.047. which indicates it is significant at 5% significance level. The coefficient of foreign exchange rate is -0.7012238, which shows for one percent change in foreign exchange rate, keeping the other things constant will result -0.7012238 percent changes on dividend payout of sampled private commercial banks in opposite direction.The finding is consistent with the researcher's early expectation Therefore, the researcher haven't got enough reason to reject the predetermined hypothesis 8, which states that there is a significant negative relationship between foreign exchange rate and dividend payout of private commercial banks in Ethiopia.The finding is consistent with Mundati(2013) who found out foreign exchange rate has a negative significant relationship with dividend payout.

Conclusion
The general objective of this paper is to investigate the determinant of private commercial banks dividend payout by using nine years balanced panel data set covering from 2009 to 2017 across 10 private commercial banks operating in Ethiopia. For this purpose bank specific variables that contain profitability ,leverage, liquidity, bank size and last dividend payout; and macroeconomic variables like inflation rate , real GDP growth rate and foreign exchange rate are incorporated as explanatory variables for dividend payout. To see the effects of the identified explanatory variables over the dependent variable (dividend payout), the researcher applied fixed effect panel data regression models and the results shows that all bank specific variables have a positive effect in determination of dividend payout of private commercial banks in Ethiopia.
In general from bank specific explanatory variables that are incorporated in this study profitability, liquidity, bank size and last year dividends are found out as a positive significant factor for dividend payout of private commercial banks in Ethiopia. However leverage is not a significant factor for dividend payout of private commercial banks in Ethiopia. From macroeconomic variables foreign exchange rate is found out as a significant variable but inflation and real GDP growth rate are not a significant variable for dividend payout of private banks in Ethiopia. Therefore, the major determinants of dividend payout of private commercial banks in Ethiopia are profitability, liquidity, bank size, last year dividends and foreign exchange rate.

Recommendations
In this part the researcher provides the possible recommendation for all stakeholders, those who should consider the result of this study including both actual and potential investors, Ethiopian private commercial banks and the regulatory body of Ethiopian government or national bank of Ethiopia. Based on the finding of this study the following recommendation is forwarded by the researcher to the above identified stakeholders.
 The finding related to profitability shows there is a positive significant relationship between profitability and dividend payout. Therefore it is better for private commercial banks, if they make an investement in profitable benture to earn higher profit and to pay a better dividend for their shareholders.  The finding in relation to liquidity shows that it has a positive effect for dividend payout. This implies that shareholders expect more dividend from the banks that has higher liquidity position. Therfore it is better for the banks, if they maintain their liquidity position in order to be enough to pay higer dividend for their shareholders.  The finding in relation to bank size shows there is a positive significant relationship between bank size and dividend payout. this indicates that the larger the bank size the more dividend payout in Ethiopian private commercial banks. So it is better for banks, if they pay dividends proportionally to their sizes.  Last year dividend is found out as a significant factor for dividend payout and this shows that it is considered as a good signal for banks performance on the side of shareholders. Therefore, in order to keep their loyality on the side of their shareholders, it is better for Ethiopian private commercial banks, if they pay dividends consistently.  Most of the time potential investors who want to invest in the banking sector are seeking information about the bank's profitability and dividend payout rate. But there are also other factors that should be considered in deciding the decision whether to invest or not in that particular bank .So it is better for those investors, if they understand the effect of other bank specific variables like liquidity, size and macro-economic factors like foreign exchange rate.  The finding related to macro economic factors shows that exchange rate has a significant effect on dividend payout. So the government body,especially the financial sector regulatory body should consider the effect of macro-economic factors on dividend payout so as to make Ethiopian private commercial banks enough to pay dividends to their investors.  In general the result of this study has identified the significant factors that affect the bank's dividend payout. Therefore, it is better if the management and bank's board of directors take consideration of those factors affecting dividend payout.