Antecedents of Dividend Payout Among Listed Non-Financial Companies listed in Nairobi Securities Exchanges, Kenya

: The objective of this study was to establish the antecedents of dividend payout among listed non-financial Companies listed in Nairobi Securities Exchanges, Kenya, with four specific objectives; to establish the effect of profitability, capital expenditure, leverage and liquidity on dividend pay-out of listed non-financial companies in NSE. The study was founded on Modigliani and miller hypothesis, signalling hypothesis, birds in hand fallacy, Agency and clientele policy. Using correlation research design, the study examined the antecedents of dividend pay-out of listed non-financial companies listed in Kenya for the period 2008 to 2019. The result showed a positive and significant relationship between dividend pay-out ratio and profitability, liquidity and leverage. It also shows negative association between dividend pay-out and capital expenditure. It was concluded that capital expenditure significantly influences dividend paid out negatively. Interest payment for long term debt takes priority as a charge on the profit made by the company. An improvement of a company's liquidity would lead to a better compensation to shareholder inform of dividend distributions. Listed companies therefore are expected to pay dividends when the companies are performing well because otherwise shareholders may question the proceeding of the announced profit or results to signaling effect. Listed companies should arrange the financing of capital expenditure so as ensure shareholder remains at an advantage and enable the company to recover its cost on capital and expected returns.


I. INTRODUCTION
There has been persistent interest among finance scholars on the role of dividend pay-out in corporate companies (i.e. see Lintner, 1956;Miller and Modigliani, 1961;Baker, Farrely and Edelman, 1985; Allen and Michaely, 1994) as cited in Rafique (2012). According to Oruru (2010) the amount of earnings which is distributed amongst the shareholders is known as dividend. The dividend payment is an important component in an organization since it aims at providing a solution to dividend puzzle on whether the payment of dividend increases or reduces firm's value and more so it's a financing strategy whereby if a company distributes huge amount in form of dividend, then there is need to look for alternative sources of finance when pressed by some needs in future. Dividend pay-out is of concern to both investors and management since it can influence company's stock valuation which will impact the resultant firm's value. More so the amount of annual dividend can influence the asset pricing, capital mix and investment decision. Past scholars such as Myres (1977) as cited in Ardestani et al., (2012), perceived the amount of dividend to be received by the ordinary shareholders to be influenced by profitability, risk, ownership, size, investment opportunity and firm growth opportunities to significantly influence the amount of dividend paid by any corporate company. Finance theorists and corporate experts are in congruence that the main theme of all firms is shareholder's wealth and profit maximization (Ndili & Muturi, 2015). This calls for development of vibrant and robust financing, working capital, investment and dividend decisions. Earlier theoretical proponents purported an existence of perfect market though it was refuted by Gordon (1963) who perceived investors to be more concerned by both dividend and capital gain as stipulated in bird in hand theory. To Gordon dividend policy impacted positively corporate valuation and most investors were risk averse thus they opted cash payment instead of capital gains. These claims were refuted by the proponents of tax preference theory whereby investors had preference for capital gains rather than pay corporate tax rate. Therefore, there is no cemented conclusion on how, why, and how much ought to be paid as dividend to investors in Nairobi securities exchange.
and their policies that discriminated against them (Ngugi, 2013). The market was fully operationalized around 1954 after formation of a society of stock brokers. Later Capital Market Authority (CMA) was formed with the main purpose of developing and promoting securities exchange market in Kenya. This was through development of regulations and trading platforms that would enhance investor confidence and minimize the risk exposure. In 2000 Central Depository System (CDS) was introduced so as to enhance liquidity in stock market. In 2011 NSE changed its name to be Nairobi Securities Exchange in preparedness for trading of different securities and provide clearing services, derivative platforms and other related securities (NSE, 2013). NSE acts as platform for raising capital, provision of liquidity of financial assets and source of economic performance information. Through it there has been notable growth in equity, debt and derivatives trading. Currently there are 68 listed companies grouped into 11 sectors that are financial, utilities, consumer discretionary, energy, healthcare, industrials, technology, telecommunication, materials and real estate. The current study was limited to investigation of determinants influencing dividend pay-out of Non-financial Companies listed companies in NSE.

II. MAIN OBJECTIVE
The aim of the study was to establish antecedents of dividend payout ratio of listed non-financial companies in NSE, Kenya. A. Specific objectives i) To establish the effect of profitability on dividend payout of listed non-financial companies in NSE, Kenya. ii) To determine the effect of capital expenditure on dividend pay-out of listed non-financial companies in NSE, Kenya iii) To find out the effect of leverage on dividend pay-out of listed non-financial companies in NSE, Kenya iv) To establish the effect of liquidity on dividend pay-out of listed non-financial companies in NSE, Kenya

III. THEORETICAL REVIEW A. Modigliani and Miller Hypothesis
This hypothesis was documented by Modigliani and Miller in 1958. The hypothesis argues that in an efficient market there are no transaction, taxes, bankruptcy and taxes since all stakeholders have equal access to information that may guide in decision making. The value of firm is not dependent on leverage because every firm's investment decision is solely dependent in the choice of their asset class. To attain financing optimality then a balance between interest costs and floatation cost associated with issuing new debt. Further, they purported that profitability and risk are firm value determinants and not capital structure. Indeed, investment decision is mainly determined by the arbitrage opportunities which exist in any viable investment opportunity. Therefore, investors will tend to dispose share of highly valued entities and invest in under-priced companies (Mwangi, 2016). Due to investors' rationality behaviour Modigliani and Miller (1958) purported that there exists an inverse relationship between cost of equity and gearing ratio and the investors are unwilling to take any risk which they cannot be compensated. In case when tax rate is zero, then it will be hard for any firm to obtain optimal capital structure. Further, Alifani and Nugroho (2013) argued that there are high chances of using debt financing because of the advantages associated with corporate taxes. The theory is not void of criticism more so because of the assumptions in which the theory is based on. In fact, it is so hard to have an operating environment void of transaction costs, bankruptcy costs and agency conflicts (Mwangi, 2016). An increase in market to book value ratio of listed companies may trigger increase in firm value and consequently enhance firm's ability to borrow. Debt financing would demand for regular commitments in servicing of loans that may have effect on available cash for payment of dividends.

B. Signaling Hypothesis
Signaling hypothesis theory was developed by Spence (1973). The theory proposes that dissemination of positive information publicly is meant to portray current and future positive performance. Corporation information is shared by respective companies to achieve desired objectives that may be short, medium or long term (Bini, Giunta & Dainelli, 2013). Regular flow of information is meant to optimize information access costs and minimize level of information asymmetry. There are cases of level of information asymmetry signaling changes in stock market. According to Spence (1973) the level of information hoarding is dependent on level of influence an institution controls on investors decision making. Moreover, there is need for regular information update so as to enhance investor confidence through quality of decision making. This is a theory which asserts that announcement of increased dividend payments by a company gives strong signal of positive trend. Companies use dividends to share profits with stockholders and when doing so, they can decide to issue a dividend when ploughing profits back into the company for development and growth is not the best option, is not necessary or not practical. At the time officials make the decision to offer a dividend, they usually make an announcement, providing information about the amount and date so shareholders know what to expect (Spence, 1973). These announcements are closely anticipated and followed because investors believe they can provide information about the company's financial health. Generally, dividend signalling is done by the company when it changes the amount of dividend to be paid to shareholders. Miller and Modigliani (1961) work sustain that, in a perfect capital market, a firm value is independent of the dividend policy. However, some years later, Bhattacharya (1979), John and Williams (1985) and  developed the signalling theory classic models, that linked dividend policy to information asymmetry. So, a dividend increases signals an improvement on a firm's performance, while a decrease suggests a worsening of its future profitability. Consequently, a dividend increase (decrease) should be followed by an improvement (reduction) in a firm's profitability, earnings and growth. Moreover, there should be appositive relationship between dividend changes and subsequent share price reaction. One of the most important assumptions of the signalling theory is that dividend change announcements are positively correlated with share price reactions and future changes in earnings. According to  signalling theory is also banked by the assumption that information is not equally available to all parties at the same time, and that information asymmetry is the rule. Information asymmetries can result in very low valuations or a suboptimum investment policy. Signalling theory states that corporate financial decisions are signals sent by the company's managers to investors in order to shake up these asymmetries. These signals are the cornerstone of financial communications policy. In this case, managers know more than investors, so investors will find "signals" in the managers' actions to get clues about the firm. So, the theory simply suggests that firm's announcements of an increase in dividend pay-outs act as an indicator of the firm possessing strong future prospects.

C. Birds in Hand Fallacy Theory
This theory was brought forth by Gordon and Linter (1963) it is based on certainty of dividends and capital gains among individual investors. They argued that failure to pay all earnings as dividends is precipitated by availability of investment projects which have net present value, an undertaking of such projects increases shareholders through capital gains. Moreover, the theory perceives that risk averse investors has more preference for dividends against capital gains, receipt of dividend income has certainty when compared with capital gains which can be received upon surpassing the risk in which the investment is exposed to and if a company is highly concentrated with investors who have preference for certain income then there are high chances of high dividend payment and vice versa. According to Easterbrook (1984), the bird in hand effect will occur if investors utilize their dividends for consumption or to acquire treasury bills, but if they reinvest the received dividends in the same or a different business, they subject their cash (minus taxes paid) to the same risks as if there had been no dividends, the bird in hand effect will not occur until the business alters its in-house policies. However, just because investors are ready to wait for their dividends rather than receive them right away does not mean they do not want them. Some investors (legally protected shareholders) are ready to wait for future dividends in the case of attractive investment possibilities, and this is common in fast-growing companies because fast-growing companies pay smaller dividends than slow-growing companies (La Porta, De Silanes, Shleifer, & Vishny, 2000). Furthermore, La Porta, et al., (2000) explained, on the contrary, regardless of investment opportunities, some investors (poorly protected stockholders) are want to take dividends as much as they can get, most possibly because the investors think, dividends (bird in the hand) are better than retained earnings (bird in the bush) that might be never to realize as future dividends (bird can fly away), if investment misallocation happened. This is similar with finding by Brennan and Thakor (1990) that, if the effective personal income tax rate on dividends is not too high, then the stockholders with low ownership, will prefer to take dividends.

D. Agency Theory
According to Jensen and Meckling (1976) all corporate organization are mostly run by management which acts on behalf of shareholders. There are different parties which are involved in running of companies and they must be involved in making dividend decisions. For example, the management is interested in satisfying their needs through huge payment which will increase the expenses and consequently reduce the amount entitled to the shareholders. In addition, regulators may increase the minimum capital requirements. This may alter the proportion held inform of capital. Providers of debt capital may increase interest charges that may decrease the amount of earnings attributable to shareholders. Agency theory is associated with creation of agency problem that is manifested in risk sharing among different cooperate entities (Arrow, 1971). Different stakeholders in an organization have different levels of risk tolerance hence their actions would be different. Principal stakeholders may be willing to undertake higher risks so as to achieve more economic gains as compared to agents who may have short term goals that may deter them from undertake high risk investments. These differences in risk sharing would amplify conflict in different groups. According to Ross (1973) agency conflicts are associated with incentives while Mitnick (1975) attributed it institutional framework. Ross alluded agency problems were associated with compensation. Institutional structure argument by Mitnick was crucial in development. According to this proposition organization are founded on agency and grows as they seek agency reconciliation. Alchian and Demsetz (1972) and Jensen and Meckling (1976) organization creates contracts that supports their production process. Corporate entities are legal entities based on contractual relationships. Principal agency partnership is a relationship based on self interest that causes conflicts. Thus, monitoring and agency costs are introduced to mitigate them. In this contracts, structure, level of information asymmetry and labour market share has significant role in achievement of ownership structure.
According to Fama and Jensen (1983) decision making process can be segregated into decision management and decision control and agents has crucial role throughout decision making process. They are segregated dependent in complexity and non-complexity of decision-making process.
In complex situations, agency problems arise since management decision process is not initiated by the real owners of corporate entities. Hence, the need to control decision making process to enhance survival of a firm.

E. Clientele Dividend Policy Theory
According to Petit (1977) there are different groups of clients who are attracted to invest in a given company. These groups of individuals have different characterisation which will make them have unique preference of dividend policies for example among the retiree they may have more preference on receiving annual dividend as such to increase their source of meagre earnings. In contrast young investors would prefer reinvestment as such to minimize the amount of taxes which they ought to pay from their total earnings. According to Easterbrook (1984) by view of clientele effects, if some investors are in different tax positions from others (for example, some hold tax-sheltered funds while others are taxed at ordinary rates), the different groups will have different preferences for dividends, or it could be said that the taxed group would prefer to take profits as capital gains; the untaxed group would be indifferent. By Baker and Haslem (1974), in their study had found that, dividends is the most desired factor by investors, because dividends are assumed less risky than the capital gains expected from reinvested earnings. Also, there is an informational content behind dividends or dividend changes, because it can provide an indication of management's judgments concerning the firm's future earnings expectations. But Baker and Haslem (1974) also found that, investors cannot be classified in a single homogeneous class, because certain types of stocks prove attractive to particular types of investors, which was created clientele effect (Miller & Modigliani, 1961). Furthermore, Baker and Haslem (1974) classified the investors based on their behaviour and classes in two distinct types, those who seek dividends and those who seek capital appreciation. The investors who seek dividends are investors which have tendency for using dividends and financial stability as the basis for their investment decision, also, they are known as a risk-averse investor. Whereas, investors who seek capital appreciation are investors which have tendency for using future expectations as the basis for their investment decision, or in the other words, these investors are willing to sacrifice their current dividends for future price appreciation. Mehta (2012) investigated the determinants of dividend payout policy among companies listed in Abu Dhabi securities in 2005-2009. The study hypothesized that dividend policy is influenced by firm size, profitability, price earnings ratio, leverage and liquidity. Profitability was operationalized as return on equity, return on assets and Earnings per share. Census sampling was used to select all 44 companies which were listed in Abu Dhabi in 2005-2009. Secondary data was collected from audited annual financial statements. Both correlation and step wise regression analysis were used to analyse the data. Profitability had negative and significant effect on dividend policy. Moreover, the most profitability had the highest explanatory power of the changes in dividend policy among the companies listed in UAE. Even though the study applied step wise regression it ignored the period and group effects and it would have been appropriate to use panel regression analysis with pooled, random or fixed effects regression analysis. Lee, Isa and Lim (2012) evaluated dividend changes and future profitability among companies listed in Malaysia. Multiple regression model was fitted on secondary data. The study found that although dividend was influenced by firm earnings it was weakly related with earnings for one year and there was no relationship for earnings beyond one year. In addition, there was weak evidence that dividend influenced firm's profitability. Ahmed (2015) posited that profitability and liquidity had significant influence on dividend payout in the banking sector. Ahmed operationalized profitability as return on equity, return on assets and earnings per share. Bivariate and multivariate techniques analysed the data. Inverse significant effect of ROE, ROA and dividend payout was documented. Moreover, EPS positively affected dividend payout. Muchiri (2006) studied antecedents of dividend payout. Using questionnaires, data were collected form the respective companies finance directors. Among other factors, current and expected profitability was ranked first in determining whether to pay dividend which using multiple regressions was found to be significantly influencing the decision on whether to pay dividend. Wasike  Panel research design was adopted and secondary data was collected from annual financial statement from 1992 to 2008. In general, multi regression analysis revealed that although debt financing, equity financing and cost of debt had significant influence on dividend payout in New York they had no influence in Shanghai securities exchange. Moreover, there was positive and insignificant relationship between debt financing, cost of debt financing and dividend payout ratio. Equity financing had positive and significant influence on dividend payout ratio. It was recommended that those companies which were listed in China should diversify their financing alternative as such to boost investor's confidence. A Pakistan case studying antecedents of dividend payout was carried out by Ahmad and Muqaddas (2016). Panel research design was adopted; secondary data was collected over nineyear period from 2006 to 2014. The study hypothesed that financial efficiency, safety, risks and profitability. Financial efficiency was operationalized as interest ratio, safety was measurement investment to total assets, risk was measured as nonperforming loans to gross loans and profitability was measured as return on assets. Dividend payout policy was positively affected by financial efficiency and risk. Further, safety and profitability inversely contributed on dividend payout ratio. Musiega et al., (2013) investigated the determinants of dividend policies among non-financial companies listed in Nairobi securities exchange. Descriptive research design was adopted in the study and 30 non-financial companies which were listed from 2007 to 2012 were purposively selected. Profitability, growth opportunities, liquidity and current earnings were assumed to have influence on dividend payout. Firm size and business type were moderators. The study revealed positive and significant relationship between growth opportunities and dividend policies while liquidity had inverse and non-significant relationship with dividend payout ratio. Nnadi, Wogboroma and Kabel (2012) studied dividend payout ratio antecedents in African securities markets. The study adopted correlation research design. Purposive sampling was used to draw 1742 companies from 29 African countries. Panel secondary data was collected from 1998 to 2009. The study revealed an inverse and non-significant relationship between agency cost, leverage and dividend payout ratio. In contrast there was a positive and nonsignificant relationship between government ownership, age and dividend payout ratio.

D. Liquidity and Dividend Payout
Ahmed (2014) examined the impact of liquidity on dividend policy among 30 listed commercial Banks in 2012. Although, the data was single period it was assumed to be a true representative of the historical patterns. Multivariate approach was used for data analysis. The study found a positive but non-significant relationship between liquidity and dividend payout ratio. Kibet (2012) studied liquidity and dividend payout in Kenya using regression model. A positive but not significant relationship between liquidity and dividend policy was documented. From the findings it was implied that listed firms ought to maintain high levels of liquidity as such to be in a position to pay dividend when they fall due. Since the data was panel in nature it would have been appropriate to use pooled, fixed or random effects regression modelling as such to capture both time and group effects of each company. Olang

C. Data and Data Collection Procedures
The study adopted panel data that was sourced from NSE hand books, financial statements and specific company's websites. As indicated in the data collection sheets information sought was on profitability, capital expenditure, leverage, liquidity and dividend payout. Period under considerations was 2008 to 2019. Year 2008 was characterized by post-election violence that may have impacted on business operating practices.

D. Data Analysis
Collected data was processed and analyzed using Stata 14. A mix of descriptive statistics; mean, standard deviation, skewness and kurtosis and inferential statistics that include Product moment correlation and multiple regression were used. The resulting model is; Y it =α +β1x1 it + β2x2 it + β3x3 it + β4x4 it + έ Where. Y -Is the Annual dividend paid, α -is the regression constant term, β1, β2, β3 and β4 -Are the regression coefficients X1 = Profitability, X2 = Capital expenditure, X3 = Financial Leverage, X4 = Liquidity and έ= error term

E. Operationalization of the Variables Profitability
Profitability is the positive difference between revenue and operating costs of a corporation (Pandey, 2009). Corporate entities are formed with the intention of maximizing shareholder's wealth and profitability. Javeda, Raob, Akramc and Nazird (2015) argues that profitable organization should always exceed the cost of capital on their investment returns. Profitable corporations have higher odds of paying dividend as compared to non-profitable entities. Profitability may have positive or negative effect of dividend payout since some entities may retain more of it depending on their stages of growth. Profitability was operationalized as natural logarithms of earnings after tax.

Capital Expenditure
Capital expenditure is the annual budgetary allocation on acquisition and improvement of non-current assets. The amount allocated for capital expenditure on annual basis may minimize the dividend to be paid. This will depend on sources of financing for respective capital expenditure. Ardestan et al., (2013) document significant contribution of investment opportunity on dividend payout. Imran (2011) documented positive significant contribution of EPS, profitability, growth in sale, cash flows and firm size on dividend payout. Capital expenditure was operationalized as natural logarithms of annual capital expenditure.

Leverage
Leverage is the amount of financing in a corporate entity that is sourced from external sources. Debt financing will attract fees that are payable on regular basis and they may have negative implications if respective investment does not generate amount enough to cater for regular costs. Tamimi and Takhtaei (2014) a negative but insignificant relationship between financial leverage and dividend per share. The choice of OLS was not appropriate since they ought to have considered period and group effects jointly. Farahani and Jhafari (2013) found a positive but insignificant relationship between debt-to-equity ratio and dividend per share. Leverage was operationalized as natural logarithms of interest paid per annum.

Liquidity
Liquidity is the capacity of an entity to convert its resources into liquid cash. The faster the conversion period the higher the liquidity. Igan, Paulo and Pinheiro (2010) found a positive significant contribution of liquidity on dividend payout. Moreover, shareholder's concentration had the strongest positive moderating effect on the relationship between dividend policy and stock liquidity. Liquidity showed no significant association with dividend payment ratio, according to Komrattanapanya and Suntrauk (2013). However, dividend payout ratio was inversely related to investment possibilities, financial leverage, and sales growth.
Liquidity was operationalized as natural logarithms of (Net Income + Depreciation & Amortization + Other Noncash Adjustments + Changes in Non-cash Working Capital).

F. Diagnostic Tests
Panel modelling is based on several assumptions on heteroscedasticity, normality, serial correlation, Multicollinearity and Hausman test among other tests.

G. Stationarity
Stationarity is a situation in which statistical features of variables under examination remains constant within the period under examination. Classical models should be fitted on stationary data otherwise it spurious model will be fitted. In this study stationarity was examined through use of Augmented Dickey Fully (ADF) test. The tests assume that the data is stationary against an alternative of non-stationary (Stock & Watson, 2018).

H. Normality Test
According to Hansen (2020) normally distributed data portrays a bell shape with mean of zero and standard deviation of 1. Even though, there are graphical and statistical tests that can be applied for normality test the current study used Jarque Berra test. According to Baltagi (2005) Jarque Berra test assumes that data is normally distributed and if P value <0.05 the data is not normally and there is need for transformation so as to achieve normality. Data can be transformed through natural logarithms, use of inverse or square roots.

I. Multicollinearity
Multicollinearity is a condition in which explanatory variables are highly related (Wooldridge, 2012). Variance inflation factors (VIFs) and tolerance were applied for its examination. There is multicollinearity among variables if VIF is more than 10 and tolerance limits less than 0.1. To manage it then there may be need for model rectification and dropping correlated while modelling.

J. Heteroscedasticity
According to Gujarati and Porter (2009) heteroscedasticity is a condition in which the error terms of a regression model have no equal variance. The assumption can be examined through use of modified Wald test and if its p value < 0.05. Then there is need to fit Feasible Generalized Least Squares Model (FGLS) or ordinary least square model with robust standard errors (Baltagi, 2005).

K. Serial Correlation
According to Wooldridge (2012) autocorrelation is a situation in which current and past period error terms of variables under examination are related. In this study it may be examined through use of Wooldridge Drukker test. If p value > 0.05, then there was no serial autocorrelation (Torres-Reyna, 2007).

L. Hausman Test
Fixed-effects models are a class of statistical models in which the levels (i.e., values) of independent variables are assumed to be fixed (i.e., constant), (Greene, 2008). Random-effects models are statistical models in which some of the parameters (effects) that define systematic components of the model exhibit some form of random variation. Statistical models always describe variation in observed variables in terms of systematic and unsystematic components. According to Greene (2008) fixed effects model assumes different groups have different intercepts and random effects model assumes that the groups have different error terms. Hausman (2008) argued that there is always a conflict between the choice of random and fixed effects model which can be resolved through the use Hausman test which hypothesis that the study data has fixed effects.

VI. RESULTS AND DISCUSSIONS
The following are the results from the secondary data collected from 42 listed non-financial companies in NSE in 2008 to 2019. Descriptive analysis, diagnostic tests and panel data analysis is presented. Since the data had both cross sectional and time series characteristics panel data approach was used. Prior to the data analysis the data was transformed using log transformation.

A. Panel Data Descriptive Analysis
Results in Table 4.1 shows that the average annual dividend is 15.45 with an average deviation of 1.504. The average profitability is 13.407 with a standard deviation of 2.890 which is higher variability as compared to the dividend that was distributed. Capital expenditure averaged a value of 16.069 which is higher than the profit made. Leverage had mean of 15.077 whereas liquidity was found to have an average of 13.201 with a standard deviation of 0.985. Normality examination indicated that the variables under examination were not normal since their respective p values for Jarque Berra were less than 0.05. Non-normality was in agreement with Githira, Muturi and Nasieku (2019) who reported that financial data among listed companies in East Africa securities Exchanges (EASE) was not normal. This is mainly because of differing features of listed companies such as size which may have implications on financial decisions adopted. These results contravene Wanjau, Muturi and Ngumi (2018) who reported that transparency characteristics of listed companies in EASE were normally distributed. This was not in agreement with Mwangi, Muturi and Ngumi (2016) who reported nonnormality of financial structure of listed companies in EASE.

B. Correlation Analysis
Correlation analysis results are shown in Table 4.2.
Profitability positively contributed to annual dividend pay. This implies an increase in profit would lead to an increase the value of annual dividend paid. There was an inverse but significant relationship between annual dividend and capital expenditure in the Kenyan companies. This implies that an increase in capital expenditure would lead to fall in the amount of dividend paid. There was a positive and significant relationship between leverage and annual dividend paid. This implies that an increase in leverage increase amount of annual dividend paid; it can be deduced that Kenyan companies invest borrowed funds on activities which are increasing annual dividend.
There was a strong positive and significant relationship between annual dividend paid and liquidity, this implies that companies that are more liquid are expected to be large amount dividend as compared to those that are experiencing liquidity problems. Therefore, there is a need to examine the working capital decision as such to counter its impact on liquidity position of the companies. The odds of multicollinearity were low since none of predictors had correlation coefficient higher than 0.7 with each other.

C. Diagnostic Analysis
Diagnostic tests were carried out before modelling to evaluate the most appropriate model to fit in the study. Tests carried out include stationarity, Hausman, multicollinearity and heteroscedasticity among others.

Stationarity Test
Unit roots was carried to examine the stationarity of variables under examination. ADF stationarity was used. As shown in Table 4.3 the null hypothesis for non-stationarity or presence of unit roots was rejected since p values were less than 0.05 for all variables. Consequently, it was concluded that annual dividend paid, profitability, capital expenditure, leverage and liquidity were stationary. The findings supported Wairimu, Muturi and Olouch (2019) who found firm financial characteristics of listed non-financial companies in NSE to be stationary.   Table 4.5 shows the test results for time fixed effects. There were no time related effects when the dependent variable was annual dividend since the p value >0.05. It was not appropriate to introduce a dummy variable or use two-way analyses when annual dividend is the dependent variables.

Heteroscedasticity and Serial Correlation Test
Both heteroskedasticity and serial correlation tests were summarized as shown in Table 4.6. Heteroskedasticity was tested using modified Wald test whose results showed that there was no uniform variance since in the trio cases the P values were less than 0.05 therefore robust standard errors should be used to eliminate biasness. In addition, there was no evidence for serial correlation among the panels (p value > 0.05).

Multicollinearity Test
Multicollinearity was absent since VIFs did not exceed 10 as shown in Table 4.7. Hence, multiple regression model would be fitted to evaluate the influence of profitability, capital expenditure, leverage and liquidity on dividend payout among limited liability companies in Kenya.

Hausman Test
Hausman guided on choice between random and fixed effects model. To achieve this Hausman test was applied. Results in Table 4.8 revealed that the most appropriate model to fit when annual dividend was the dependent variable was random effect since the p value > 0.05.

VII. CONCLUSION
In the review of theories and empirical literature the study found that there exists an interconnection between the amount of dividend and four independent factors under consideration.
These factors were profitability, capital expenditure, leverage and liquidity. To start with, profitability as measured by total annual earnings after tax has similar co-movement with annual dividend payout. This means that an increase in the level of earnings after tax would result on more Dividend being paid. Another key observable determinant that affects the amount of dividend paid annual is the company investment in capital expenditure as measured by annual capital expenditure. Capital expenditure significantly influences dividend paid out negatively. This means that as the company management and shareholder choose to invest more then, they forgo dividend receipts since this reduces earning amount left to be shared. It can also be concluded as leverage as proxied by annual interest charges on long term debt had a positive and significant relationship with dividend policy as measured by the total annual dividend paid. Interest payment for long term debt takes priority as a charge on the profit made by the company, even before tax, and hence reduces the amount of profit after tax which is available for distribution to shareholders. Finally, it can also be concluded that liquidity as measured by combination of net income, depreciation, amortization, other non-cash adjustments and changes in non-cash working capital influence positively and significantly the level of dividend paid. This means that an improvement of a company's liquidity would lead to a better compensation to shareholder inform of dividend distributions.