The Signaling Effect of Dividends on The Market & Financial Performance of Listed Companies in Sri Lanka

This study aims to examine the Signaling effect of dividends on the financials and market performance of listed companies in Sri Lanka. This study facilitates information regarding current and future performances, such as dividend practices, future growth projections and insights into market behaviour for stakeholders. The total population was identified as the companies listed in the CSE, which was approximately 316 firms. The sample of the study was 102 companies listed between the years 2015 to 2019. Systematic sampling was used to select the sample. In analysis, panel regression was used as the data analysis technique. STATA version 16 Statistical Software Package was used in conducting the data analysis. Random effect regression model was recommended by the hausman test. The results concluded that dividend decisions were impacted insignificantly on subsequent market performance. Dividend per share was positively and significantly impacted on both return on assets and return on equity. Dividend pay-out ratio was positively and insignificantly impacted on return on equity and also negatively and insignificantly impacted by return on assets. The current study recommends that further research could be done by focusing on dividend announcement dates and calculating the subsequent market performance, three to five months following the dividend announcement date.


Introduction
Corporate finance consists of three main areas, namely dividend decisions, investment decisions and financing decisions. Through of them, dividend decisions have proved to be the most complex and puzzling area in corporate finance. Dividend decisions involve deciding how much of the earnings of the companies should be distributed among their shareholders while retaining the balance partial amount of earnings with them for future growth of the companies. Therefore, dividend decisions from an integral part of overall decision making for companies as they generate the resources and furnish information regarding company's prospect of development. In addition to that, dividend decisions act as a communicator to transmit significant information about company's potential expectations to stakeholders. This dividend decision changes over time between developed and underdeveloped economies. In current study, Dividend Per Share (DPS) and Dividend Payout (DPR) ratios are considered as the determinants of dividend decisions. Through DPS, investors could take proper aware of firm's both past and current financial stability. The DPR indicates what quantity will be returned to shareholders from the earnings and what amount will be retained to pay off debts or reinvest in major business operations. The greater DPR will be highly valuable to shareholders because of the higher level of return on shares which were held. According to the signaling effect, dividends provide signals for stakeholder decision makings. For instances, if a company declares an increment in dividends, consequently the investors might earn high returns. Therefore, potential investors will pay much more attention to invest in best options to derive better earnings in future and government can have a clear idea of the company performances. Miller & Rock (1985) hinted that the unforeseen changes in dividends, the information related to current and future cash can be identified through signaling effect. In present study, financial performance is measured through Return on Equity (ROE) and Return on Assets (ROA). Through ROE, it represents how efficiently a company can utilize shareholder's money. ROE could also be contemplated as return on net assets, since the ratio shows the amount of profits that the entity generates in relation to its shareholder's equity. ROE is equivalent to the subtraction between total assets of a company and its debt. Whereas, ROA is a profitability ratio that shows the amount of profits that an entity could earn in relation to its total assets. In other words, ROA provides the efficiency in which the company is generating profits using the assets or the economic resources available to the company. Regarding to the market performance, it may fluctuate due to different variations. Those variations will directly impact on every industry in the country. Inconsistency of the market price of shares will conclude that how much pressure has put by dividend decisions on company's performances. Market based rate of return and Sharpe Ratio were used to measure market performance for this study. The market-based rate of return pinpoints the investors whether it is worth to invest in the company or exit. Furthermore, Sharpe Ratio is one of the best key indicators to measure the market performance which assist stakeholders to take awareness of investment return related to its risk. The reason of conducting this study is to explore the impact of dividend decisions and signalling effect of dividends on financial and marketing performance of all sectors of listed companies in Sri Lanka.

Originality Value
There was hardly any research that have utilized the Sharpe ratio to determine the market performance relating to determining the signaling effect of dividends. Therefore, the novelty of this research was using the Sharpe ratio.

Main Objective •
To investigate the signaling effect of dividend decisions on financial performance and market performance of listed companies in Sri Lanka.

Sub objectives •
To determine the impact of dividend decisions on subsequent market performance • To determine the impact of dividend decisions on current financial performance • To explore the alternative methods of assessing the impact of dividend signaling on firm's subsequent market performance

Problem Statement
Sri Lankan firms faced many complications when making dividend decisions over the last five years (CSE, 2019). There were considerable fluctuations on the behavior of dividend payout from 2015 to 2019 due to various situations such like Easter Sunday attack, political uncertainty, impacts of foreign intervention, fluctuations of foreign exchange rates, economic collapse. The current study is expected to determine how the fluctuations of dividend have been impacted on both subsequent market performance and financial performance. Figure  1.1 clearly illustrates the problem statement.

Literature Review
Prosperous companies are earning profits and these profits may be invested in firms' regular operations, securities such as bonds, stocks, certificates on deposits, commercial papers, pay off debts and finally distribute dividends for shareholders (Tamrin, Mus, Sudirman, & Arfah, 2017). Usually, dividends are determined by the board of commissioners by evaluating the financial position of last periods during a meeting. According to the available data of developing and emerging markets, it has been proven that the earnings of companies who pay dividends have better quality than the earnings quality in non-dividend paying companies (Nguyen & Bui, 2019). This research will be mainly focused on earnings quality. Earnings quality can be categorized into two major parts such as accounting based earnings quality and market-based earnings quality. Nguyen & Bui (2019, p.305) hinted that "Earnings quality usually includes accruals quality, persistence, predictability, conservatism, timeliness, smoothness, etc." Among the above dimensions, accruals quality, persistence, predictability and smoothness are useful to measure accounting-based earnings quality. Value relevance, timeliness and conservatism are the dimensions which are helpful to measure market-based equity quality (Nguyen & Bui, 2019).

Signaling Theory
Signaling theory states that the dividend policy acts as a communicator and is able transmit significant information to the investors about the company's future expectations. Announcements of cash dividends help the shareholders to convey significant information about the company's future profitability to the investors. There were circumstances that signaling theory breached. One of the circumstances was proved by (Wijekoon & Senavirathne (2019) research, firm's dividend policy issues complicated signals to the investors. Managers are not always leading dividend policy as value-maximizing for shareholders.

Agency Theory
Agency theory arises when the interest of shareholders and interest of managers do not coordinate. Managers are not always leading dividend policy as value-maximizing for shareholders and they mostly think to conduct dividend policy as maximizing their own private benefits such as salary and bonuses. However, it should be noted that the argument exists when some managers are making dividend decisions in respect of their interests, instead of focusing on shareholders' wealth maximization (DeAngelo & DeAngelo, 2006)

Bird in Hand Theory
This theory described that dividends are less risky than capital gains as dividends are certain. Therefore, investors prefer dividends rather than capital gains (Amidu, 2007). This theory proves one of the most popular proverbs that are "a bird in the hand is worth two in the bush,". Due to this, each firm has a duty to maintain better dividend policy to maximize the company's stock price (Al-Malkawi, Determinants of Corporate Dividend Policy in Jordan: An Application of the Tobit Model, 2007). On the other hand, the past literature Hussainey, et al. (2011) has proven a disagreement of Bird in Hand in Theory by the following statement "instead of the tax drawback relating to the dividends, better way of managing dividends of a firm issues some green signal for investors about the firm".

Clientele Effect
This is a concept when dividend policy changes, it will directly impact on the behaviours or reactions of various types of investors and eventually it may lead to volatility of market share prices. Furthermore, reactions and behaviors of investors may differ according to the changes of dividend policy, taxes and other policies. Primary assumption of this specific theory is, that the investors usually get attracted to the different company policies at first and make efforts to adjust their stock holdings when company policies have changed (Investopedia, 2018).

Tax Effect
This theory illustrates that lesser dividend payout ratios will reduce the cost of capital and enhance the market share price. Further, it describes lesser dividend payout ratios will be helpful to maximize the firm's value. This theory is completely opposite to the bird in hand theory. Since it argues against the assumption of bird in hand theory, this theory is built on different assumptions such as dividends will be taxed under the higher tax rates than the capital gains, dividends could be taxed immediately than the capital gains and investors might be highly focused on companies with low dividend payout (Al-Malkawi, Rafferty, & Pillai, Dividend policy and share price volatility:UK evidence, 2011).

Gordon's Theory
According to Gordon's theory, current dividends are more essential to evaluate market value of a firm. It is one of the famous calculation models used to evaluate market value by using the dividend policy. Further, Gordon's theory explains about "the relationship between rate of return and the cost of capital relating to market price per share" (efinancemanagement.com, 2020). But this model could not be used for companies with low growth rates and non-dividend paying companies (Acheampong & Agalega, 2013).
According to Abdi (2010), this research study extracted the following statement from Abdi (2010,p.10). "The theory stipulates that payment of dividend conveys information to the market with respect to expected future earnings of the company". Further, this study stated, increasing of dividends would be a promise towards the positive trend of future earnings and this research study stated the most prominent fact such as, if a company decided to declare more amount of dividend payment than the market.  Healy & Palepu (1988) research study was focused on determining the signaling effect. It mentioned that the firms which stopped dividend payments expereince in decrease of profits. It indicates that if a company pays more dividends, profits will be declined within the particular year. Modilgani & Miller (1961) and Chowdhury, et al. (2014) studies discussed about the positive impacts of signaling effect. Deeptee & Roshan (2009) stated, even though, the managers apply dividends to convey the information, dividend changes might not issue the perfect signal. In addition to that, Easterbrook (1984) mentioned that the enhancing of dividends will issue an ambiguous signal to recognize a developing firm or a poor firm. Modigliani and Miller (1961) research study stated that there is no relationship between firm's market value and its dividend policy while having the perfect stock market. In reference to this theory, there was some research which proved that there is no or little impact of signaling effect on future earnings (Watts, 1973;Gonedes, 1978;Penman 1983). According to Chowdhury, et al. (2014), however, in the Sri Lankan context, dividend announcement date will not be useful as the investors have an irrational behaviour. This is because Sri Lanka is not an emerging market. Basically, the signaling effect of dividends could not be practiced due to the probability of greater managerial incompetence, poor governance practices and inadequate financial forecast skills among investors. Masum (2014); Abdullah Al-Hasan, et al. (2013); Baskin (1989) studies discussed about the determining the impact of dividends on market performance. Masum (2014) research paper has focused on dividend policy and its impact on stock price. PAT, EPS and ROE were considered as moderating variables and only share market price was considered as a dependent variable which was useful to measure market performance of firms. Results of Masum (2014)  According to Baskin (1989), DPR was chosen as the independent variable and stock price volatility was chosen as the dependent variable. From the results and findings section the particular research, it has found that there was a positive but not significant relationship between dividend policy and stock price volatility.  2019) research paper defined dividend policy as "a compensation payable to shareholders for risk tolerance. DPS, EPS, DPR and Price Earnings Ratio (PER) were considered as dependent variables and ROA and ROE were considered as dependent variables. There is no significant relationship between DPR and ROA.
Priya & Nimalathasan (2013) conducted a research on the "Dividend policy as the decision to pay out earnings against retaining and reinvesting them", which concluded that an improvement in financial well-being of a company has a positive impact on company's dividend policy.
According to Amidu (2007) research paper, dividend policy and Payout Ratio were considered as the independent variables and ROA, ROE and Tobin's Q were considered as dependent variables of this study. In the results section, there was a positive and significant and negative relationship between ROA and Dividend policy.
Following Al-shattarat, et al. (2018) and Chowdhury, et al. (2014) studies discussed about the way of assessing the signaling effect. Al-shattarat, et al. (2018) research study mainly focused on event study methodology to determine the signaling effect of dividend decisions. The ulimate objective of event study management is to determine whether the shareholders earn excess returns or not when asscociating with special events. Applications of event study methodology within this research were described as follows. Firstly, the general assembly meeting date was considered as the event date. Then, Al-shattarat, et al. (2018) study was mainly focused on 100 daily trading observations. Selected event period was from day T to time T (T= -11 & Time T = -110). Further 11 trading days were considered.
Relating to Chowdhury, et al. (2014)  In present study, signaling effect of dividend decisions was determined through the subsequent market performance and current financial performance. Adjusted market rate of return and Sharpe ratio were recommended to measure the signaling effect.

Methodology
This section involves in research methodology that is supportive to accomplish the research objectives. The current study attempts to determine the signaling effect of dividends on subsequent market performance and current financial performance.

Figure No.1: Conceptualization Framework Independent Variable Dependent Variable
.

Sample of the Study
When selecting the sample, there were 316 total companies on 12 th July 2020 along with 21 sectors in the CSE website. The current study used systematic sampling technique to extract the sample for each sector category. Moreover, according to systematic sampling the current study selects every second company from the listed order (12 th July 2020) in the CSE. After conducting the systematic sampling method, the sample size was 158 companies. Out of 158 companies 56 companies had to be omitted as there was no sufficient information available and therefore the remaining 102 companies were selected as finalized sample

Moderating Variable
• Firm Growth

Sources of Data
Both market performance and dividend per share were calculated by using the market data obtained from CSE. Relevant data regarding to Dividend Payout Ratio, Return On Assets and Return On equity were gathered from published annual reports on CSE.

Analytical Tools & Techniques
Panel regression model with fixed effect model and Random Effect model was used to achieve the objectives of the study as the major analytical technique. Panel regression model was built by using STATA version 16 software.

Diagnostic Tests
The current study followed three diagnostic tests through STATA. Which diagnostic tests are multi-collinearity test and heteroscedasticity test.
Through the descriptive statistics, normality test was done to determine whether the data set is normally distributed or not.
Multi-collinearity test was done to explore whether there are similarities between independent and dependent variables of the study or not. Under the heteroscedasticity test, Breusche-Pegan-Godfrey test was followed to determine whether the errors or disturbances happened in linear regression models or not. Table 6.1 illustrates the equations for calculating the dividend decisions indicators, financial performance indicators and market performance indicators.

Multi-collinearity Test
There is a statistically insignificant impact of the DPOR_C on SR_S which is seen, as the p value of 0.458 is greater than the significant level of 0.05 and z value (t test) is lower than 1.96. Furthermore, there is a statistically insignificant impact of the DPS_C on SR_S as the p value of 0.128 is greater than the significant level of 0.05 and z value (t test) is lower than 1.96. Statistically insignificant impacts indicate that the changes of the dividend decisions won't convey adequate signals about the subsequent market performance of a firm. Furthermore, Table 5 illustrates the negative relationship between DPOR_C and SR_S due to the negative value of the coefficient statistics relating to DPOR_C. The negative relationship indicates that the increasing of DPOR_C minimizes the value of SR_S. In addition to that, the positive coefficient statistics of DPS_C indicates a positive relationship between DPS_C and SR_S, which implies that an increase of DPS_C would maximize the SR_S. AMRR t = α + β 1 (DPOR t − DPOR t−1 ) + β 2 (DPS t − DPS t−1 ) + β 3 F _ SIZE + ε The results illustrate the statistically insignificant impact of DPOR_C on AMRR_S which is seen from the generated p value of 0.867, that is greater than 0.05 and the z value (t test) that is lower than 1.96. Furthermore, Table 5 shows that there is a statistically insignificant impact from the DPS_C on AMRR_S since the p value of 0.996 is greater than 0.05 and z value (t test) is lower than 1.96. Statically insignificant impacts indicate that DPOR_C won't convey the adequate level of signals about the market performance of a firm. In addition, table 5 represents that both DPOR_C and DPS_C have negative relationships with AMRR_S, due to the negative value of the coefficient statistics relating to both of DPOR_C and DPS_C. The negative relationship indicates that the increasing of the changes in dividend decisions would reduce the AMRR_S. The overall results relating to Objective 1 show that the current study reports the statistically insignificant impact of both DPS_C and DPOR_C on subsequent market performance. ROA t = α + β 1 DPOR t + β 2 DPS t + β 3 F _ SIZE + ε The results indicate that there is a statistically significant impact from the DPOR_C on ROA which is seen, as the p value of 0.000 is lower than the significant level of 0.05 and z value (t test) is greater than 1.96. Furthermore, there is a statistically insignificant impact from the DPS_C on ROA as the p value of 0.667 is greater than the significant level of 0.05 and z value (t test) is lower than 1.96. Statistically significant impacts indicate that the DPOR could convey significant information about the financial performance of a firm. Furthermore, these results show a positive relationship between DPS and ROA due to the positive coefficient value. It indicates that paying a higher amount of dividends would improve the ROA ratio. However, there is a negative relationship between DPOR and ROA due to the negative value of the coefficient relating to DPOR. It indicates that paying a higher amount of dividends would decline the ROA ratio. ROE t = α + β 1 DPOR t + β 2 DPS t + β 3 F _ SIZE + ε

Conclusion
This study mainly focused on the signaling effect of dividends on subsequent market performance and current financial performance and revealed some important findings regarding to the signaling effect of dividends in the Sri Lankan context. According to the first hypothesis, the study disclosed that the subsequent market performance was insignificantly impacted on the dividend decisions in Sri Lankan context. Also, finding from the second hypothesis observed that there is a significant impact of DPS on the current financial performance. Third hypothesis investigated that there is an insignificant impact of firm growth on subsequent market performance. According to the fourth hypothesis, the study explored that there is an insignificant impact of firm growth on ROA and a significant impact on ROE. According to the findings, the current study recommended that dividend decisions were more significantly impacted on the current financial performance rather than the subsequent market performance. Also, among the two recommended indicators for measuring market performance, dividend decisions were more significantly impacted on Sharpe ratio rather than the adjusted market rate of return. Determining the signaling effect of dividends is one of the most challenging and unsolved issue in the corporate finance.

Contribution of the Study
Shareholders' wealth maximization is the ultimate goal of any business organization. To satisfy shareholders, companies establish a set of dividend decisions which are helpful to determine shareholders' investment decision. Through this study, companies can do changes relating to dividend decisions along with raising awareness of shareholders. This study is essential to identify the firm's current and future performances, for instance, dividend practices, future growth projections and insight of market behaviour. Furthermore, the particular research study might be helpful for stakeholders to take different decisions on various purposes such as making decisions about profit planning and investments by managers, decisions of educational and research purposes by students, making dividend payment decisions by board of directors. Apart from that, ratio analysis (ROA, ROE, DPS and DPR) of this study will be useful to gain a clear idea about determinants of dividend decisions. For example, DPS facilitates to indicate management's willingness to make consistent payout to shareholders and DPR assists to examine the company's stability and growth of future earnings.

Policy Implications
Provided that dividends had a statistically insignificant impact on the subsequent market performance, the current study recommended that a majority of the shareholders in the Sri Lankan context didn't place much concern on dividend announcements (Abdi, 2010). Further, Chowdhury, et al (2014) in his study, mentioned that since Sri Lanka is not an emerging market, testing the rational behaviour of investors on dividend announcement date would not be useful. Therefore, the dividend announcements won't adequately be reflected through the share prices. It further explains that since most investors in Sri Lanka don't immediately react based on the issued dividend announcements, dividend information won't be reflected in stock prices of a firm.
However, this study reveals that DPS had a significant and positive impact on the current financial performance. This particular impact recommends that the companies have favourable a dividend policy when they are earning more profits while having an unfavourable dividend policy when they are earning losses. However, according to this study, DPOR had a negative relationship with the current financial performance. This negative relationship reflects a puzzled nature and the negative correlation recommends that the most of the companies did not enhance their dividends even they were earning more profits. It seems that the most companies are maintaining constant dividend payments.