AbstractWe test the relation between the dividend announcement date and themarket reaction to the dividend announcement by Thai listed firms.
We find the significant difference in positivemarket reaction on the announcement date and 4 days after the announcement dateand negative market reaction on the 9 days before and 8 days after theannouncement date. The evidencesuggests that investors can gain the short-term return by day trade investing. 1.
IntroductionMany researchesin the past shows that there are many measures that can assess the informationof cash dividend announcements. The two methods that most commonly use are theunusual or abnormalin stock return and the stock’s variability.Nevertheless,the stock price volatility does not have any theory that can significantlysupport the hypotheses of the tests so the conclusions will be mainly based onthe literature (Acker, 1999)unlike thestock’s abnormalreturns which can be calculated by using events studies.One of theimportant factors in stock price changes is uncertainty of information whichreflect investor’s expectation of return and risk.The paper hasutilized variability of returns as the measurement to find out the risk relatedto volatility of the stock return, for both long term and short term, on andaround the date of announcement of dividends which included only cash dividendby using methodology event study.The ‘risk informationhypothesis’ proposes that cash dividend announcementconducts a new change in risk of the firm.
The decreasein risk arise because the volatility of the firm and firm’s earningssurprises have been reduced (Dyland Weigand, 1998).The lower the volatility is, the lowerthe opportunity of loss or gain in the shortrun. Therefore,the volatility of the equity returns is going to boostup around the date of announcement ofthe dividend in cash form only. Thevolatility of equity returns is one of the essential measurement for measuringthe level of risk that the investors are exposed to (Guo, 2002), by nature investors are risk averse.
This paperdoes not focus on the investigate of abnormal return related toonly cashdividend announcement; alternately, it examines variability of returnsas the important aspect of risk. 2.LiteraturereviewThestudy of the relationship between the announcement of cash dividendannouncement and the volatility of stock returns has been conducted by severalresearchers. The arrival of newinformation would affect the stock volatility which is the use to measure risk (Ross,1989).Inthe US, the studies related to dividend and stock price volatility wereconducted by many researchers but the result are mixed.Theresult of Baskin (1989)concludethat the relationship between dividend yield and stock price volatility isinversely related but on the other hand Friend and Puckkett(1964);John and Williams (1987)resultcame out positive. Based on signalingtheory, the dividend announcement may use as a tool to signal quality of thefirm to the public (Allen and Michaely,1995).
Moreover,an unexpected change in dividend would influence the change in stock price in apositive direction. Gordon(1963)proposethat shareholders prefer a certainty from cash dividend rather than anuncertainty of capital gain that have more risk in future cash flow.Inaddition, frim believe that the announcement on dividend would make a positiveimpact to value of the firm. Onthe other hand, the studies in different countries both developed and emergingcountries show the different results.
In Australia, accordingto Allen and Rachim (1996),There is no supporting evidence that dividend yield would impact stock pricevolatility. The same result occurs inBangladesh (Rashid and Rahman,2008).Mesteland Gurgul (2003)proposethat the bad news announcement cause the price of stock more volatility sue tothe increase in uncertainty.
The lower dividendpayment shows the greater business risk (Jensenet al.,1992). 3. Objectives and research hypothesesTo identify the effects ofannouncing the cash dividend decisionson volatility of the equity returns, thispaper has also determinedthe size and which direct that the risk will be changedforboth short-run cash dividend announcement risk andlong-runcash dividend announcement risk during the cash dividend announcement date inThailand.
The following is the null hypotheses:H1: There is no significant difference inshort-termrisk before and after the cash dividends announcement date.H2:There is no significant difference inlong-termrisk before the cash dividends announcement date and after the cash dividendsex-date.H3:There is no abnormal return duringperiod before and after the cash dividends announcement date.
H4: There is no cumulative abnormal returnduring period before and after the cash dividends announcement date. 4. Research methodology 4.1.
Data description and sample size The study has presented five yearsinformation during a period of 2012 to 2016.The secondary data of cash dividendannouncements was collected from SET High Dividend 30 Index are in accordancewith those used for SET Index calculation. To derive the Index value, means ofmarket capitalization and weight with dividend yield were used to calculate.The highest dividend yield used for thecalculation is capped at 15 percent. There were 222 cash dividendannouncements of the 30 companies listed on SET High Dividend 30 Index as on June 30, 2011.Out of the 30 companies listed, around11 companies were Resources Industry, 8 companies were Financials Industry andthe remaining companies were in other industries like Property &Construction, Technology, Services, Agro & Food and Industrial Industry.However, there were cash dividendannouncement data that was not available for 3 companies, thus reducing thesample size to 27 companies. 4.
2 Event study methodology4.2.1. Analysis of returns:Event study methodology was employed toexamine the stock price reactions (Brown and Warner, 1985). The eventstudy can be used in both clinical studies which is used to investigate the effectof an event of a single company’s stock prices and large sample studies which is used toexamine the effect of an event on prices of different sample stocks. Moreover,researchers have also examined the impact of an event on stock volatility (DeFusco etal.
, 1990; Engle and Ng 1993; Jayaraman and Shastri, 1993), on stocktrading volume (Benkraiem et al., 2009; Karafiath,2009), or on accounting performance (Barber and Lyon,1996). Also, notonly company stocks can be examined in the event study but also other types ofsecurities. Daily and monthly returns are commonly found in manyliteratures. Daily data is often considered as short-term eventstudies (Lummer and McConnell, 1989; Small et al., 2007), whereasmonthly data is normally chosen for long-term studies (Ritter,1991; Teoh et al.
, 1998b). Bowman (1983)described the 5 main steps inconducting an event study as follows: v Identifying the event of interest and selectingsample firm or stocks.Thisstep requires to specify the date of this event.
The event date can be called as theannouncement date of the event (whencash dividend is announced for the first time in the public newspapers), or day zero.v Model the security price reaction oridentifying the time line of an event study.Thisstep has to identify the test period (the event window)and the estimation period (EP).The event window of this study was 31days during a period of 15 days prior to the announcement date to 15 days afterthe announcement date along with the announcement day itself.The estimation window was in a range ofday 16 to 166 days prior to the window) consisting of 150 trading days as shownin Figure 1: v Estimating the expected return for eachsample stock over a period.Theexpected return, (Ri,t)is considered as the benchmark returnin the normal circumstance to compare with the actual return during the eventwindow. Thisbenchmark return represents the return unrelated to the event of interest whichcan be calculated by using CAPM formula as can been seen from Equation 1 below: Note:The parameters like and areestimated by running regression.
Rm,t is the corresponding return on theSET index and Ei,t is the error term. ?v Computing abnormal or excess returns.The abnormal return is the differencebetween the actual return and the expected return on a particular day.The abnormal return (AR)for each day for each firm can beobtained as following formula: v Analyze theresults by testing the significance of abnormal returnsIn order totest the significance of abnormal returns, most event studies use a parametrictest of t- statistics (e.g. Brown and Warner, 1985; Barber and Lyon, 1997). For thispaper, we use one – sample t-test to calculate the abnormal return.
4.2.2. Clean event window period To ensure that the event window study onlyfocused on pure cash dividend, thus, any other type of announcement like stockdividends and stock splits, bonus issue and share repurchase mergers,acquisitions, amalgamation, joint venture, capital investment, substantialorders from prestigious customers or any other such financial events during theevent window were not considered as a part of the sample (McWilliams andSeigel, 1997).
4.3 Analysis of short-term and long-term variance research designsThe experiment ofpre-testing and post testing research design is used tomeasure the change in the volatility of returns.The analysis is classified into twoperiods in short-termand two periods in long-term.The short-run analysisis evaluating based on a changing in variance over a period of 15 days for bothbefore date of cash dividend announcement and after date of cash dividendannouncement.
Forthe long-run,the period regarded is beforeevent date of post announcement 120 days and 120 days after the ex-dividend date of the event date.To test the differences aresignificant or not, we use paired sample t-test to set against the mean of the before cashdividend announcement andafter cash dividend announcement of the specified time period.The fifteen days before and after theex-date are excludesto avoid the distortions because of a lot of trading activity is made during those periodsof time.Figure 2. theperiods of our investigating areconcerned both short run effects and long run effects of announcing a cashdividend.5. Empirical findings Thepurpose of this segment is to test whether is there a significant difference involatility of the sample companies before and after cash dividend’s announcement in both in short period and long period.
Moreover, this segment also tests theexcess return and cumulative excess return of the sample companies during theperiod of announcement of cash dividend. The result will be shown in table 1, table 2 and table3. 5.1 Short-term effect Asyou can see from table 1, the variance has increased for 15-day periods after the cash dividend announcement.The average risk of stock price afterthe declaration is higher than the mean risk of stock price before the declaration,but the result from the paired t-test show that they were not statically significant.The increased variance come from thesignificantly change in average abnormal returns after the announcement of cashdividend as shown in table 3. Also, theincreased abnormal return may be from investors perceived the companies thatpaying dividend as less risky and have enough free cash flow to repayshareholders as dividend. Therefore,companies that announce cash dividend gain popularity which lead to an upsurgein trading volume.
Also, theywill has higher volatility of return after the cash dividend’s announcement. 5.2 Long-term effect Table 2 showsthat the return volatility from the long-term impactof cash dividends of the samples companies.The 120-day meanvariance had decreased. Also, the 120-day standarddeviation had decreased. The supporting reasons to this result are thatin the before-cash dividends, investors had weighted highlyon information of dividend announcement, or another way, they had disturbed theprice reaction that lead to high volatility in the before-cashdividends.
And the price changes in the after-cashdividends period is less and its volatility is lower because the dividendpayment had occurred and there is no disturbance of the price from investors,or we can say that the market condition is normal.Also, theless volatility of the stock after-cash dividends period is that there is lowertrading volume of the stocks because the nature of the samples companies; highdividend stock, is flat, mostly this kind of stock is not involved withspeculative investors, that create high price volatility.5.3 AbnormalReturn and Cumulative Abnormal Return The finding in this researchare obtained from the event study methodology using the average abnormal return(AAR)andcumulative average abnormal return (CAAR) in dividend announcement date as the startingpoint (day0) that were obtained from the sample stocks forthe study period. The sample period covers the 15 days before (day -15)and after (day 15)the dividendannouncement date.
To test hypothesis 3, the result in the tableshow that the average abnormal return in the day of announcement (day 0)and 4 daysafter the announcement date (day +4) are 0.26%and 0.19%respectivelyand significant at 10% level with the positive t-statistic of1.68 (p-value=0.09)and 1.70 (p-value=0.09)respectively.
The resultalso shows that the AAR in the 9 days before (day -9)and the 8days after (day +8) of the announcement date are -0.28%andsignificant at 5% level with the negative t-statistic of -2.36 (p-value=0.02)and -2.14 (p-value=0.
03)respectively.To testhypothesis 4, the result in the table show that there is no significant at anylevel for cumulative average abnormal return (CAAR)on theannouncement date or the day before and after the announcement date. ConclusionWith therespect to the test of the hypothesis 1, we notreject the null hypothesis according to the p-value for two-tail is morethan significant level. However, the variance after the announcement ofcash dividend seem to be higher than before the announcement.
Therefore,investors perceived companies paying cash dividends as less risky than companythat no dividend payment.With the respect to the test of thehypothesis 2, we do not reject the null hypothesis according to the p-value for two-tail is more than significant level.However, the variance after theannouncement of cash dividend seem to be lower than before the announcement.Therefore, investors perceivedcompanies paying cash dividends as more risker than company that no dividendpayment.With therespect to the test of the hypothesis 3, we reject the null hypothesisaccording to the significant on the day -9, day 0, day4, and day 8.
It is possible to make the profit by daytrading on the day of the announcement and the 4 days after the announcementthen gain the return of 0.26% and 0.19%respectively.On the otherhands, it needs to rethink about buying the stock on the 9 days before and 8days after the announcement date since the result show the negative return of -0.28%.