Dividend Announcement Effect On Stock Price In Case Of An Emerging Market-India
Dividends are basically a form of paying back excess wealth generated by the firm to its shareholders which is either paid in form of cash dividends or stock dividends. In other words, it's the profits made by the firm which is paid back to the shareholders in form of dividends. Cash dividends refers to the excess wealth generated being paid in form of cash while stock dividends also known as bonus issue refer to issuing the shareholders with additional shares in proportion to the shares already held by the shareholders. The decision whether to pay dividends in form of cash or stock is solely the decision of the management of the firm. However, this paper focuses and does its study on cash dividends as cash dividends is the most common dividend payout method used and specially in emerging market India, stock issue is rare and not much understood by common investors and mostly undertaken by companies with large market capitalisation. Also the investors are bias towards cash dividends and prefer investing in companies paying cash rather than issuing additional shares.
Many companies retain a part of the profits earned for further investment and pay the remainder earnings to its shareholders in form of dividends .Dividend payout policy is a very crucial decision which management of every firm makes as dividends reflect to certain degree the financial well being of the company. Also many investors predict their returns and future profitability of the company to the extent of dividend the firm pays. Thus, it is said that dividends do have informational content.
The relevance and irrelevance of dividend announcement and shareholder reaction has been a debate over long time which has been shown in prior empirical studies. Miller and Modigliani (61),showed the irrelevancy of dividend announcement on shareholders wealth after which many studies were undertaken by various researchers like being Bhattacharya (79), Aharony and Swary (80), Asquith and Mullins (83) Miller and Rock (85) etc to study the relevance of dividend announcement on shareholders wealth. Along with empirical studies showing that there is an effect on dividend announcement on shareholder wealth, some empirical studies like Watt (73), Black and Scholes (74), Penman (83) etc showed that there is no effect of dividend announcement on stock price, thereby no impact on shareholders wealth. In spite of such debates, one of the most important and crucial decisions made by management of a firm is that of the dividend payout policy because dividends act as an important mode of communication between the investors and the firm's management. It's believed that managers have inside information about the company which investors do not have and therefore they use dividends as a way of signalling these inside information to investors. Although investors doubt whether information provided by managers are reliable as managers will not provide the truth about the financial status of the firm, they do rely on dividend announcement as a signal about the financial status of the firm as cash dividends involve payment through cash which is difficult to manipulate. If this is the case, when a firm announces dividend, there should be an impact on the stock price of the firm as investors will invest in the firm announcing dividend leading to a change in the share price. The extent of effect on stock price i.e. whether the share price falls or rises depends on how the investor views the dividend announcement as positive or negative change. Positive change is when the management announces an increase in dividend to be distributed as compared to the dividend distributed in the previous year and negative change is when the management announces a cut in dividend in comparison to the dividend distributed in the previous year.
This paper aims to study this informational content of dividend, whether do dividends really signal something about the financial status of the firm thereby affecting the firm's stock price. Empirical studies conducted prior, mostly have shown that there is an effect on stock market return when a firm announces dividend irrespective of positive or negative change reflecting the that dividends do signal information about the future prospects of the company. Also proven is that change in stock price is positively correlated with dividend change. Mostly these studies have been conducted on firms which are listed in the stock markets of developed countries. The general focus of these studies depicting an effect on stock price with a dividend announcement is shown in markets of developed countries whereas the same effect on stock price of the firms listed in markets of emerging markets is not very well established. There can be variations in the effect of dividend announcement on stock prices across markets in developed and emerging countries.
Analysing a market of an emerging country will augment to the Dividend Puzzle (Black 1976).Hence; role of cash dividends in an emerging market, India is studied in this paper. Analysing the effect of dividend announcement on stock price or shareholders wealth can be a prolific empirical study for emerging market like India, which can vary from markets of developed countries. Selecting Indian stock market for the study is interesting as there are not many empirical studies cited on the Indian stock market, especially on the effect of dividend changes on the stock price. Also, over the years the Indian market has grown and become one of the popular emerging markets for investment by multinationals. Another reason for selecting Indian stock market is that till few years back investors had various apprehensions to make investments in shares of a firm as there was lack of information about the companies due to lack of transparency of information and also due to a major scam which took place during 1992.However the investment perception has changed in recent years with proper legal framework by Securities and Exchange Board of India, 1992 (SEBI) and awareness through media. In this context, with such major changes taken place over the years, it is interesting to study effect of dividend announcement on stock prices in Indian Stock market thus contributing to the prior empirical works on dividend announcement and its impact on stock price or shareholders wealth.
The remaining part of the paper is organized as follows: Section II gives a brief framework and history on the Indian Stock Market, section III reviews relevant literature on the prior empirical works conducted, section IV provides details about the data used for the study, section V states the hypotheses to be tested, section VI describes the methodology undertaken for the study followed by section VII in which the empirical results are presented and discussed. The last section i.e. VIII provides conclusion for the study undertaken.
Over View Of Indian Market
This section gives an overall idea of the Indian market and how it functions. The section is further segregated into two parts. The first part gives synopsis of the Indian stock market and its structure while the second part provides information on the corporate sector in India.
Brief On The Indian Stock Market
In India, there are 19 recognised stock exchanges. However, as per the legal structure, these stock exchanges are divided into two main groups. One group consist of 16 stock exchanges namely Madras Stock Exchange, Cochin Stock Exchange, Bhubaneswar Stock Exchange, OTCEI, Inter-connected Stock Exchange, Guwahati Stock Exchange, Bangalore Stock Exchange, Vadodara Stock Exchange, Delhi Stock Exchange, Calcutta Stock Exchange, Pune Stock Exchange, Jaipur Stock Exchange, Uttar Pradesh Stock Exchange, Ludhiana Stock Exchange, National Stock Exchange (NSE), Coimbatore stock exchange that were started by companies which were limited by shares or by guarantees. The second group consist of 3 stock exchanges namely Bombay Stock Exchange (BSE), Madhya Pradesh Stock Exchange and Ahmedabad Stock Exchange which earlier were established as an association of persons and then transformed into companies later.
In India, all the stock exchanges are supervised and regulated by the Securities and Exchange Board of India (SEBI), a regulatory authority. SEBI was set up in April, 1992 as a regulatory authority to protect investor's interest and to regulate and develop the securities market in India. As per SEBI's regulations all the stock exchanges needed to complete the process of demutualisation and corporatisation.However, except for Coimbatore stock exchange the remaining 18 stock exchanges are operating as profit companies limited by shares which have been demutualised and corporatized.
Although there are 19 stock exchanges, the BSE and NSE are the two main and important stock exchanges functioning in India. For the study undertaken in this paper, firms listed in BSE 100 are taken into consideration.
In Asia, BSE is the oldest stock exchange by being in existence for 133 years starting in 1875.However, in 1956 as per the Securities Contracts (Regulation) Act (1956), BSE received a permanent recognition making it to become the first stock exchange to get such recognition in the country by the Government of India. As of today, among exchanges it ranks at the first position for the number of companies listed and for number of transactions it ranks at the fifth position. As on 31st December 2007, BSE had a market capitalization of about USD 1.79 trillion. BSE has aided in the growth of the corporate sector in India by enabling the firms to take to their advantage and efficiently benefit from the various resources offered by the BSE. Majority of the companies have raised resources using the services of BSE from the capital market. The Bombay Stock Exchange provides 21 indices out of which 12 are sectoral indices.
BSE 100 is one of the indices offered by the Bombay Stock Exchange. It was launched on 3rd January 1989 as was earlier called BSE National Index.BSE 100 consist of 100 scrips i.e. companies from various industry sectors and trading on different exchanges. The BSE 100 index covers movement of share prices of companies on national level. Since 2004 it has adopted the free float methodology for computation. This is a method to construct an index which is based on the concept of free float where the shares that are readily available for trade in the market in the normal course. It excludes the shares that are held by any strategic investors, government holdings and the promoter holdings. Such a way of calculation thus reduces the market capitalisation of a company to only the number of shares that could be traded in the market every day. This helps to trace the current trends in the market easily. It does not allow only a few companies with large market capitalisations to dominate the index thus making the index a very broad based one. Free float index is beneficial for investors having both active and passive style of investment. An active investor can compare his or her realised returns to a benchmark which is a closer representation of an investible index. In case of a passive investor it reduces the tracking error. Free float mechanism has been notably adopted by various developed market index providers like MSCI, FTSE, S&P and STOXX. In this regard, firms listed on BSE 100 made an ideal choice to be the sample for this study.
Overview Of The Corporate Sector
As compared to corporate governance in developed countries, Indian corporate governance is quite different. In India, the corporate sector is not homogenous but contains a mixture of companies from government and private sector. The private sector again is a mixed bag of firms where the ownership ranges from multinationals to family business or individually owned firms. Various activities of corporate like share holders' rights, company's administration and governance, dividend announcements and various other disclosures etc take place under the legal framework of the Companies Act (1956).This legal framework has been fairly stable over the years.
As observed in various other emerging markets, the structure of ownership of corporate firms in India is also categorized by large shareholders. Generally the largest shareholders are given incentives and also the right to manage key decisions (e.g. dividend payout).In India also, the discretion of how much dividend to announce or whether not to announce any dividend only is solely dependent of the company's directors.
The Companies Act (1956), section 205 has set up certain rules and guidelines for all corporate firms which announce or pay dividend during any financial year. Under the section 205 (2),dividends announced or paid by firms must be out of its profits for that year or previous year after taking depreciation into consideration for that year or previous year respectively or could be out of both the profits. Another option under section 205 allows companies to pay or announce dividends from the money which the Central or a State government provides towards the dividend payment in order to fulfil any guarantee which the Government has given. Under section 205 (A),any company declaring dividend should deposit the dividend amount within 5 days from the dividend declaration date in a separate bank account and within 30 days form dividend announcement date should pay the dividend declared.
Generally in India, companies announce and pay dividends two times in a year which is known in form an interim and a final dividend. As seen companies pay dividend out of its profits and is like giving shareholders part of its earnings. Hence for shareholders dividends act as a source of income. In India, to actually understand amount a company pays out of its profits and the amount it retains to plough back for enhancing the business, dividends announced are generally conveyed on basis of ‘per share' like for example Rs 4 per share. To better understand the above mentioned sentence let's take an example. For instance, a company has Rs 8 earnings per share in the year and declares and pays Rs 4 dividend per share which implies that the company is distributing to its shareholders half of its profits and retaining other half to plough back into the business. This acts as a signal to the shareholders that the company is prospering leading to further investments and leading to increase in share price when a company announces to pay dividends. Hence in case the company declares a decrease in dividend per share, the market reacts looking at the stock as a growth opportunity for their investment since the company is now investing an increased surplus of funds in much more profitable ventures. Therefore in India, amount a company declares as dividend on per share basis is very crucial as investors judge the future prospects of the company on the amount it retains for reinvestment into profitable business activities based on the dividend per share declared.
Relation between dividend announcement and shareholder reaction has been an interesting topic in financial research over many years. Over the years there have been many studies conducted to test this relation. Of the early works are of Miller and Modigliani (61) demonstrating that under perfect capital market conditions (i.e. absence of taxes, perfect information, rationality) there is no effect of dividend on shareholders wealth. Though Miller and Modigliani demonstrated in perfect market that dividend policy is irrelevant, it is needed to act as an intermediary to send out information to the market participants i.e. it has informational content.
This informational content of dividends concept was applied by various researchers of earlier ones being Bhattacharya (79), Miller and Rock (85), John and Williams (85) to develop the dividend signalling theory. As studied by Miller and Modigliani that in perfect capital market dividend has no effect on the firm's value, however in the real world there exists no perfect capital market. In the real market world investors do not have access to all information and managers take advantage of this by using inside information to signal about the earnings of the firm by way of dividends.
Bhattacharya (79) showed that since investors have imperfect information and managers have more information, dividend policy's size does depict a signal for expected cash flow as it also involves cost. In his paper, Bhattacharya speaks about the cost structure involved in dividend payment. Dividends are taxed at higher rate than capital gains which result to making dividend payment process expensive which make the investors believe the firm is doing well. With this belief investors invest in the firm leading to changes in future earnings which depict that dividends do signal future earnings. Aharony and Swary (80), Asquith and Mullins (83) took their study further to asymmetric information and cited that there is a positive relation between dividend announcement and stock returns as dividends do convey some information about a firm's future earnings. Aharony and Swary (80) in their paper tested whether dividend announcement give more signal over current earnings about the expected future earnings. For this they used quarterly dividend announcement dates that vary from the quarterly earnings announcement dates at least by 11 days. Their findings showed a remarkable increase of 1% of average excess return over the 2 day announcement period. Their studies show that dividends provide more information about the future earnings than the current earnings announced by the management. Moreover, their results at the same time also support the semi form of efficient capital market in which the stock price efficiently adjusts to the newly acquired information. Asquith and Mullins (83) studied the dividend announcement and the dividend increment and arrived to the same conclusion as that of Aharony and Swary (80) that dividends do convey information about future earnings and impact the stock price. However, Asquith and Mullins do not take into account the current earnings; they based their study conclusions by analysing the announcement of dividend and dividend increase in subsequent time. For their study they tested 168 firms which either pay dividend or initiate once to pay dividends. Though the result is that dividend does signal information, the notable part is that the signal is stronger for firms which paid dividend for the first time as compared to firms which paid dividends regularly or increased the dividend in the subsequent time. Their result was an evident that dividends do affect stock price which is the share holder's wealth indirectly. Miller and Rock (85) extend their study of asymmetric information by amalgamating a firm's investment and financial policies. In their study, they also acquaint that managers benefit from trading of shares and use dividends as a false signal of good earnings and growth. Since investors do not have access to inside information, they purchase and invest in shares of firm leading to demand and prices to go up. However eventually the price do drop and level out but managers are benefitted who get bonus on the basis of performance.
However some empirical studies by Watt (73), Gonedes (78), Penman (83), Benartzi, Michaely and Thaler (97) note just the opposite that dividends do not signal future earnings. Watt (73) was among the first few who said that dividends do not signal much about the future earnings. Like Aharony and Swary, Watt studied if current and past earnings are better indicator for future earnings or current and past dividends are a better signal for future earnings. For this, he used 310 firms to study the relationship between current and past dividend and future earnings by regressing future earnings (t+1) on the current year's dividend (t).Though results showed that dividend do have informational content about future earnings, the change in future earnings as a result of dividend is little. Although the change is small, the informational content of dividend leading to the change is minuscule. The reason for the same is that though management possess more information, the cost involved in the dividend process and the difference between the actual dividends paid and dividends announced lowers the informational content of dividends. Later on Penman (83) in his paper went a step further to analyse management's forecast on earnings and its effect on future earnings along with the informational content of dividends. In his paper, Penman tested whether management's forecast on earnings have same effect on the future earnings as the dividend announcements made by management have on the future earnings. He concluded that though dividends and management forecast have informational content for future earnings, dividends do not have more informational content as management forecast have. Management forecast is a more powerful tool for the future earnings as compared to dividends. Like Watt even Penman's reason for it is that management do not adjust dividends to the earnings stated in forecast which tends to show low adjustment of dividends resulting to the poor informational content of dividends for the future earnings. However Penman's result is limited as the study is done on well established firms which have good earnings which limit the impact of dividends. Dividend increase does not always mean a signal of good news was studied by John and Lang (91).In their study, they analysed relation between insider trading and dividend announcement. Insider trading refers to trades done by the management before or on the date of dividend announcement. Unlike other models where insider trading is exogenously stated, in their model it is indigenized. They show insider trading interacts with the dividend policy choice and affects the effect of dividend announcement in different ways. They tested their model on companies which made first time initiation of dividends and insider trading activity data was tried to match to the dividend initiation dates. Their result was that not all dividend announcements signalled good news, a firm has announced dividend but it has insider selling the stock, it will lead to negative effect and the reverse holds true when there is insider buying the stock with the dividend announcement. Therefore, they concluded that an investor must observe the insider trading activity in conjunction to dividend announcement to understand the effect of future stock price and earnings. Again, in their study they do not take into account for firms that pay ongoing dividend so their study is partially right and also insider trading by management is subjective as one cannot know the actual reason behind the action taken by a particular manager for selling and buying. Therefore insider trading can be taken as a guideline but be fully dependant on.
Recent studies by Benartzi, Michaely and Thaler (97) also support Watts (73) evidence that dividends play a role in slight change in future earnings but do not contain the informational content as proclaimed by earlier studies. They conduct categorical and regression analyses. In categorical analyses they undertake firm specific events and divided their results as per the extent of dividend changes in groups called quintile 1 to 5 where quintile 1are firms experiencing lowest dividend increase and quintile 5 being with highest dividend increase. Then unexpected earnings are calculated for year 0, 1 and 2.Further they use regression analyses where changes in year 0, 1 2 are regressed on various variables like past and current changes in dividends. Their study concludes that there is a link between dividend and future earnings but it's minimal. Its more to do with past and current earnings as increase or a dividend announcement depict that the firm had a good financial year and excess returns it distributes as dividend. This generally makes an investor believe that that the firm is doing well which leads to investment and thus the change in future earnings. However in their test they do not include firms which initiate dividend only once and they fail to explain the difference of how the signal change if a firm increases dividend or initiates for the first time. However the most recent study by Nissim and Ziv (2001) show that dividends do convey informational element. They used a different modified regression model addressing measurement error and omitted correlated variables related to earnings. Unlike prior researchers, in their study they used linear mean reversion in earnings.
As seen there are mixed result for dividend as a signalling element for future earnings and impacting share price which is indirectly the shareholders wealth.
To test the effect of dividends on stock prices, this paper undertakes the study on the 100 companies which are listed on the BSE 100 for 4 successive years 2005 to 2008.During these years, out of 100 companies 14 companies never announced nor paid any dividend. Hence have been excluded from the sample size thereby making the sample size to 86 companies for 4 successive years 2005 to 2008.For these 86 companies the dividend announcement dates have been collected across the 4 years which result to 323 events i.e. dividend announcement dates. The sample under study is diversified and not confined to a particular sector which means it is across all sectors like media, telecommunication, infrastructure, banking and financials etc. The period under study was chosen as India during those years was facing a high growth phase where gross domestic product (GDP) of India averaged at 7.5% - 8%.A high growth era usually represents a flourishing economy which depicts high liquidity, investment opportunities for the companies and active investor participation. During a booming economy, the investor participation in markets is generally high, therefore this period provides with an apt sample to test the investor reactions to dividend announcements leading to changes in stock prices of the company.
For testing the dividend effect, final dividend dates announced and dividend paid per share by the companies have been collected from Prowess database, managed by the Center for Monitoring the Indian Economy (CMIE) and BSE. India being an emerging market, not all the companies pay semi annual or quarterly dividend. Also, it is observed that companies with larger market capitalisation announce and pay dividends semi annually or quarterly while the companies with smaller market capitalisation announce and pay dividends annually. Hence keeping with this, the paper under study has chosen to take into consideration final dividends in order to test dividends effect on stock price. In this study, the dividend announcement date which has been used is defined as the date on which the company releases i.e. announces its intention to pay dividend in the near future. This piece of news appears on various leading financial newspaper and also on BSE, thereby making the shareholders aware of future dividends to be paid. Neither the dividend paid date nor the ex dividend paid date is taken into consideration. These dates have not been taken into consideration primarily because effect of an event is seen when the event is announced and not after the event is effective. In other words shareholders react when dividends are announced which make the stock price fluctuate giving abnormal returns and not when paid.
Closing stock price for the companies during the mentioned years also has been collected from BSE which is required for calculating daily share price returns which is used in the event study conducted for the purpose of abnormal return calculation . The closing share price is after adjusting for dividends and stock splits.
To check whether the share price is congruent with dividend announcement, any other announcements like merger & acquisition, stock split, earnings announcement etc have been considered for the event window from ± 10 days of the dividend announcement date. However, such an announcement co-occurring and having significant influence on stock price during the ± 10 days of the dividend announcement was very rare during the sample period chosen.
Lastly, BSE 100 daily market closing price for the 4 years (2005-08) also has been extracted from BSE to calculate daily market returns which aide in calculating abnormal returns in the event study conducted.
Hypothesis To Be Tested
This paper tries to understand if the dividend signalling hypothesis holds true. The hypothesis in this paper is conducted to test the implications of dividend announcements on stock prices. If the dividend declaration increases or decreases the stock price, it implies that the dividend must carry some information that causes a change in the stock price. In the absence of any information the stock price should remain unaffected.
Studies conducted by Aharony and Swary (80), Asquith and Mullins (83) show increase in dividend are associated with a sound company that have healthy cash flows, and strong future earnings. A dividend decrease is associated with poor future earnings growth. In developed markets it is generally observed that a positive dividend change (i.e. an increase in the dividend) generates an increase in stock price thus generating an abnormal (i.e. In excess) return for the investor and a negative dividend change (i.e. a decrease in the dividend) generates a decrease in the stock price thus again generating an abnormal return but this time on the negative side for the investor. This however is in stark contrast to what is observed in the emerging markets.
Emerging markets are high growth economies. Therefore they provide huge investable opportunities to not only the investors but also the investee (company) as it gives the investee (company) equally good opportunities to re-invest cash surpluses for enhancing its business opportunities. Therefore a positive dividend change should lead to a decrease in stock price as it sends across a negative signal to the investor that the company is probably slowing down since it is not able to reinvest into newer business opportunities and the implication should be vice-versa for a negative dividend change.
Dividends therefore tend to carry significant information content about a company to the outsiders as they turn out to be one of the sources for the management of the company to communicate to the investors about the real financial status of the company. Therefore this paper undertakes the study to test the signalling effect of dividends on stock prices in India.
The hypothesis to be tested is stated as below:
The Null Hypothesis:
H0 : The coefficients are not statistically different from zero.
For the study it shows that the change in stock price is not statistically related to a dividend announcement.
The Alternate Hypothesis:
Ha: The coefficients are statistically different from zero.
For the study it shows that the change in stock price is statistically related to a dividend announcement.
In this paper, to study the effect of dividend announcement and stock price the standard event study methodology (eg Campbell, Lo & MacKinlay (1997)) has been used. Firstly, abnormal returns are calculated which the companies under study generate during the event window surrounding the dividend announcement date. Then cross-sectional regression of the abnormal return calculated is carried out on dividend announcement factor to explicate the abnormal returns calculated from the event study. In other words, the abnormal returns are regressed on the amount of dividend change around the announcement date.
Calculating Abnormal Returns
In order to access the abnormal returns generated by the companies which announced dividends from 2005 to 2008 standard event study method (e.g. Campbell, Lo & MacKinlay (1997)) has been used. In this study, event window taken is ± 10 days around the event date (dividend announcement date) which means 10 days prior to the event (-10,-9,-8....-1) and 10 days after the event (+1,+2,+3) and 0 refers to the event date which is the day the company announces dividend. Estimation period taken is 180 days which is -200,-20 days of the event date. Firstly the daily returns for the company is calculated using simple returns.
Rit= Pit - Pi (t-1) (1)
In the above formula Pit is the price per share of the company i at the end of time t, Pi (t-1) is the price per share of the company i at the end of time t-1.Using the same formula mentioned above, the daily market returns for BSE 100 corresponding to the dividend announcement date for the company is also calculated.
Next abnormal returns are computed for each of the companies under study for the event window under consideration. There are various models available however for in this study it is assumed that each company's returns follow the single-factor market model,
Rit = αi + βi.Rmt + εit (2)
Where Rit is the rate of return of the common stock for the ith firm on day
t, Rmt is the rate of return for the (equally-weighted) market index (m) on day t, and εit is a random variable that is expected to have a value of zero. The abnormal return (AR) during the event window for the ith common stock on day t is given by
ARit = Rit - (αi + βi.Rmt) (3)
Where the coefficients αi and βi are Ordinary Least Squares estimates of αi and βi, estimated from regressing the daily security returns (Rit) on daily market returns (Rmt) from t = -200 to t = -20 (where t = 0 is the event date).
The regression used above applies the returns observed for a time span prior to the event time in order not to alter or modify the estimation with the effect of the event under the study.
After computing abnormal returns, average abnormal returns for 3 day are calculated for every dividend announced by the firm. By 3 day it means 1, 0, +1 days of the event i.e. announcement date of dividends by a company. The reason behind taking 3 days is to capture all the effect of announce of a dividend by a company. Day 0 is the event date that is the day the company announces dividend to be paid in future which appears in the leading financial newspapers & the BSE. However some companies unofficially announce the previous day i.e. -1 day and report the news the next day. Also if the announcement news is declared on t=-1 before the stock market closes, the investors response to the dividend announcement news on the same day which precedes the announcement by a day's time. Also if the announcement news reaches the market after the market closes on day 0, the shareholders response to the news next day i.e. t=+1, which post dates the announcement day to next day. Also if the company announces on day 0 which is the announcement date during the market time, market reacts to the news on the same day, making the announcement day at 0.As seen, in reality there is a 3 day announcement day (t-=-1, t=0, t=+1).This 3 day average abnormal return is calculated by using cross sectional average abnormal returns for each security class. Cross-sectional average abnormal return is taken as its better to observe average impact of the dividend announcement on that day rather than observing impact on individual companies.
In the above formula ARit is the abnormal return the company at time and N is the no of events in the sample.
Lastly Cumulative abnormal return (CAR) is calculated for the 3 day event window i.e. -1, 0, +1.
Where T is a specified number of event days prior to the event day t.
CAR is further used as the dependant variable for the regression to be followed to test the effect of dividend announcement on stock price.
Calculation Of Dividend Change
After calculating abnormal returns using standard event study method, dividend change for each company for the 4 successive years are calculated. To calculate the dividend change, dividend per share for each company under study was collected for the years 2005 to 2008 from BSE and Prowess data base. As stated earlier, final dividend per share paid by each company has been taken into consideration for this study. The dividend change for each company for each year has been calculated as follow
Dividend Change=Current dividend per share-last year's dividend per share (7)
After calculating the dividend change three possible outcomes are observed which are no dividend change, positive dividend change and negative dividend change. No dividend change implies that the company didn't increase nor decrease its current dividend announced compared to the previous year it announced to shareholders which results to zero dividend change. Positive dividend change implies that the company increased its dividend announced to the shareholders in the current year as compared to last year's dividend it announced resulting to a positive increase in dividend change. Lastly, negative dividend change connotes a decrease in the current years dividend announced by the company as compared to its announcement for the previous year resulting in a negative dividend change. Next, based on the results obtained from the above calculation, the companies are divided into these three groups
Dividend Change= 0
Dividend Change= +1
Dividend Change= -1
Based on this segregation, 39 companies fall under the zero dividend change category, 215 companies under positive dividend change and 69 companies under negative dividend change making it total 323 data points. The dividend change calculated is employed as the independent variable for the regression undertaken.
To study the effect of dividend announcement on stock price, statistical tool like regression analysis has been used for the study. Under this, the cumulative abnormal return calculated in equation 6 is regressed on an independent variable which is the dividend change calculated in equation 7.In other words, it tries to explicate the abnormal returns obtained in equation 3 in relation to the dividend change.
The result obtained explains if dividend announcement have an impact on stock price. Regression f the following type has been used to test the effect of dividend announcement.
CARi=α + β. dividend change i +εi
Where CARi is the cumulative average return for the 3 day announcement for the firm, dividend changei is the changes in the dividend announced by the firm.
Regression is carried out for all the three dividends groups and for all the dividend change observed during the study period.
This section brings out the results for the methodology conducted above which shows the effect of dividend announcement on stock prices which eventually impacts the shareholders' wealth. It has been observed that dividend increases and decreases are an effective signalling component. Changes in dividend payout proportions actuate abnormal stock returns through increase or decrease in stock prices as they convey important information regarding the managements' vision of the future prospects of the company.
Average Abnormal Returns
The study conducted for Average abnormal returns shows that there is no positive relationship between an increases in dividend on the stock prices of Indian firms which is shown in the table below.
From the above table, It can be observed that during the 3 day event window (0 is the event date,-1 is the previous day and +1 is one day after the event) out of 284 firms, 215 firms have shown a negative average abnormal return of -3.20% in spite of an increase in dividend announced by the company and 69 firms have shown a positive average abnormal return of 0.10% on a decrease in dividend announced by the company. This is in contrast to investor reactions seen in developed markets as shown in Asquith and Mullins (83) paper. This different behaviour seen in Indian markets can be attributed to the fact that emerging markets like India have seen a significantly high growth period during the years 2005-2008, the period which has been used to select the firms for the sample data.
A drop in AAR post a positive increase in dividend announcement was interpreted by the investors as lack of management's ability to reinvest firm's cash surpluses in its ongoing business or other profitable businesses. As per the Equation, ROE=RR * g, where ROE is return on equity, RR is the retention ratio and g is the sustainable growth rate, a higher dividend payout ratio implies a lesser retention ratio which reduces the ROE of the firm, indicating that the firm has no further growth opportunities to increase its profits. Increasing dividends therefore proved to be a negative signal in a high growth economy which otherwise would have been a positive signal in a mature economy. In case of a decrease in dividend, the investors viewed it as a positive signal because a lower dividend announcement implied a higher retainment of the firm's earnings to further up its growth. Therefore the investors interpreted a lower dividend announcing company to have a higher growth rate in the future which led to an increment of the stock price of the company, generating a positive average abnormal return.
The regression analysis includes a study of the cross sectional data involving many observations on Y variable (CAR) and X variable which is the dividend change announced by the Indian companies for the time period 2005-2008.The cross sectional model comprises of data from 323 companies to test whether a change in dividend announcement by the company generates an abnormal return or not, thereby indicating the informational content of dividend. Here the cumulative abnormal returns are regressed against the announcement of an increase or decrease in dividend.
This paper observes and concludes the results obtained by analysing it at 1 % significance level i.e. at 99 % confidence interval. The reason behind this is that if it is accepted at 99 % confidence interval which is the smallest then it is obvious it will be accepted at 90% and 95% confidence interval level also.The p value is the minimum significance level at which the Null Hypothesis can be rejected. Here the significance level is α=0.01.If the p value is less than the significance level then the study fails to reject the null hypothesis showing that the coefficients of regression do not have any statistical significance. If the p value is greater than the significance level then the study rejects the null hypothesis showing that the coefficients of regression are statistically significant.
The regression analysis for an announcement of a positive dividend change shows both α and β are statistically significant, thus proving a strong relationship between both dividend change (independent variable) and CAR (dependent variable).
The p value for α=0.080 is greater than the significance level (0.080> 0.01), thus implying that it is statistically significant. Similarly the p value for β=0.595, which is again greater than the significance level (0.595 > 0.01) thus implying that it is also statistically significant.
The same can also be proven by checking if the t stat for the coefficients lies within the confidence interval at the given significance level. The above table shows that for both the coeffiecients, the t stat does not lie in the calculated confidence intervals i.e. 0.2887≤ -1.7537≤0.05607 is not true. Thus the study rejects the null hypothesis and fails to reject the alternative hypothesis, implying that the coefficients are statistically significant thus proving that there is a strong relationship between the observed abnormal returns and announcement of a positive dividend change.
The regression results above also show that both the coefficients are statistically significant hence indicating a strong relationship between the two variables, for an announcement of a negative change in dividend.
The p value for α=0.2761 is greater than the significance level (0.2761>0.01) thus showing its statistical significance. The p value for β=0.2787 is also greater than the significance level, thus showing its statistical significance as well. The table above also shows that the t stat for both the coefficients do not lie in their respective confidence intervals. For α, -0.1287≤1.0979≤0.03107 is not true and for β, -0.0015≤1.092012≤0.00366 is not true, thus rejecting the null hypothesis in both the cases. This statistical significance of the coefficients implies that the observed cumulative abnormal returns are strongly related to negative change in dividends.
Since the coefficients are statistically significant at 1% significance level, they will also be significant at 5% and 10% significance levels i.e. 95% and 90% confidence intervals respectively for both, a positive change in dividends and for a negative change in dividends.
The above table shows the regression results for all the 323 events i.e. dividend change which took place during 2005-2008 including the zero dividend change. The above regression results show that both the coefficients are statistically significant hence indicating a strong relationship between the two variables.
The overall p value for α=0. 0825 is greater than the significance level (p 0.0825>0.01) thus showing its statistical significance. As seen p value for β=0.8112 is also greater than the significance level, thus showing it's statistically significant as well. From the results seen in the above table also shows that the t stat for both the coefficients do not lie in their respective confidence intervals. For α, -0.1593 ≤ -1.7418 ≤ 0.03124 is not true. The same is not true for β, -0.0177 ≤ -0.23906 ≤ 0.01476 (not true), thus rejecting the null hypothesis in both the cases. From this, the analysis drawn are that irrespective of a positive or negative or no dividend change ,overall dividends announced do trigger abnormal returns for the stock price of the company during the event window.
Thus the above three empirical analysis aid in concluding that any changes in stock price of a firm are strongly related to changes in dividend. However, the regression does not aid in determining the direction of change in abnormal returns of stocks. It just provides with the understanding that the dividend announcements could lead to abnormal returns in either directions i.e. negative or positive.
This study demonstrates that dividends definitely carry an informational content as seen in the results of regression analysis that the paper has conducted. The abnormal returns on the company's stocks are related to a dividend announcement by the company as the board and management use the dividend as a signal to convey inside information which the investors have no access to. This hence reduces the asymmetry of communication between the investor and the management. The result of the paper is in coherence with the studies conducted by other papers like Bhattacharya (79), Miller and Rock (85), John and Williams (85), Aharony and Swary (80), Asquith and Mullins (83) which shows dividend do signal information about the company's future prospects.
The above papers have mostly conducted their studies on developed markets, where they show that a subsequent dividend increase sends a positive signal about the company, while a cut in dividend sends across a negative signal about the company. A positive signal is reflected by an increase in the stock price, thus generating a positive abnormal return. A negative signal is reflected as a decrease in stock price, thus generating a negative abnormal return.
This contradicts to the study conducted by this paper on Indian Markets, an emerging market. It has been observed that the abnormal returns generated on share price was inversely related to the change in dividend announced by the company, a dividend increase brought about a decline in stock prices and a cut in dividend brought about an increase in stock prices. This contrasting investor reaction can be contributed to the fact that the investor behaviour varies across different countries which is primarily a factor of the macro environment of that particular country.
The period under study (2005-2008) marked high growth era for the Indian markets. In these times not increasing the dividend is viewed as good news by the investors as it implies that the company is retaining its earnings for reinvestment into newer profitable horizons which will help in maintaining higher returns in the future as well. This will increase the shareholder wealth in the long term. The investor therefore invest in shares of such companies which causes the share prices to increase, thus realising a positive abnormal return for the investor
The implication for an increase in dividend signals a bad news for the investor, as the investor perceives it to be a result of lack of management's ability to generate higher profits for the company by using internally generated funds. Lesser retention by the company connoted a lower growth for the company in the future. The investor therefore showed lack of interest in such companies which caused a reduction in their share prices. Investor therefore realised a negative return.
Thus the paper concludes that although the dividends have an informational content, often the interpretation by the investor of a dividend declared is different in different countries. This difference is more evident in the behavioural characteristics between the investors of developed and emerging markets.
Further research could be extended to interpret the strength of dividend signal which can help to determine the absolute amount of movement in stock prices. Also research could be extended to the share repurchase ,which is another type of payout policy used by firms, to check whether does that have the same effect like dividends do on emerging market or are in symmetry with the developed markets.