CHAPTER IV EX-DIVIDEND DAY STOCK PRICE BEHAVIOUR: EVIDENCE FROM INDIA* 4.1 INTRODUCTION A general belief among market participants about the behaviour of stock prices around ex-dividend day is that, in a perfect capital market with no taxes and transactions costs, the stock prices should drop by approximately the full amount of dividend on the ex-dividend day [Campbell & Beranek (1955)]. However, in reality, capital markets are not perfect owing to taxes, transactions costs, etc. In such a market set-up, empirical research has shown that stock prices do not fall by the full amount of dividend on the ex-dividend day; rather, the fall is largely found to be less than dividend. The pioneering work on ex-dividend day stock price behaviour by Campbell & Beranek (1955) for US market has shown that, on the ex-dividend day, stock prices on an average drop by about 90% of the amount of dividend. On the other hand, Durand & May (1960) has shown that the American Telephone and Telegraph capital stocks drop by approximately the amount of dividend and the discrepancy, if any, was not very great. Subsequently, Miller & Modigliani (1961) argued that in a perfect capital market setup, investors would be indifferent as to whether they receive income in the form of * A paper based on this chapter is published in Prabhandan (Ramachandran et al., 2011) 76
dividend or capital gain and that, on the ex-dividend day, the stock prices should fall by the full amount of dividend. Following this, the behaviour of stock prices around ex-dividend day has been extensively investigated and three major theories have been proposed as possible explanations for the stock price drop on the ex-dividend day being less than dividend. One strand of literature, known as taxation hypothesis, attributes the stock price drop less than dividend amount to the differential taxation of dividends and capital gains. The second explanation, short-term trading hypothesis, relates to the short-term trading by arbitrageurs and tax indifferent investors. The third group of literature bases its arguments on the market microstructure factors viz., price discreteness and bid-ask spread. The theoretical framework of these explanations and the supporting empirical evidences are discussed below. 4.1.1 TAX HYPOTHESIS The tax hypothesis proposed by Elton & Gruber (1970) attribute the fall in stock prices on the ex-dividend day being less than dividend to the differential taxation of dividends and capital gains. They state that, investors who sell shares before the exdividend day lose the right to receive dividend in favour of buyers. However, if the sale of shares takes place on the ex-dividend day, the seller would enjoy the right to receive dividend. But, this retention of dividend comes at the cost of drop in share prices on the ex-dividend day. 77
Elton & Gruber (1970) derive an expression between the ex-dividend day behaviour of common stock prices and the marginal tax rates, which is based on the idea that stockholders wish to maximize their after-tax wealth. P t ( P P ) = P t ( P P ) + D (1 t ) B c B c A c A c o (4.1) where, P B is the closing price of the stock on the day before the stock goes exdividend; P c is the price at which the stock was purchased; P A is the closing price of the stock on the ex-dividend day; t c is the capital gains tax rate; t o is the tax rate on ordinary income; D is the amount of dividend. The expression P t ( P P ) i.e. the price received for the stock cum dividend B c B c minus the tax payable on any capital gain, represents the per share wealth that an investor would receive, if he sells his stock before the ex-dividend day. The wealth per share received by selling on the ex-dividend day equals to P t ( P P ) + D (1 t ) i.e. it is a combination of the after tax return on the sale of A c A c o the share [ P t ( P P ) ] plus the dividend times one minus the marginal tax rate A c A c on ordinary income [ D (1 t o ) ]. For an investor to be indifferent as to the timing of the sale, it is required that the wealth received from these cases is the same, as represented by equation (4.1). Rearranging equation (4.1) we get, P B P D A 1 t = 1 t o c (4.2) 78
The statistic, ( P P ) / D is known as the ex-dividend day price drop to dividend B A ratio. The right hand side statistic (1 t ) /(1 t ) captures the differential taxation of o dividends and capital gains and is termed the ex-dividend day tax preference ratio. c Elton & Gruber (1970) state that when dividend tax rate is more than capital gains tax rate, the ex-dividend day price drop to dividend ratio should be less than one. They have tested this proposition for US market and found that the ex-dividend day price drop to dividend ratio was less than one. From this, they inferred that, on an average, the stock prices on the ex-dividend day fell by less than the amount of dividend. Following Elton & Gruber (1970), a no. of studies have examined the ex-dividend day stock price behaviour and have found evidence consistent with tax hypothesis. Such studies are: Litzenberger & Ramaswamy (1979, 1980, 1982), Green (1980), Auerbach (1983) and Elton et al. (1984) for US, Booth & Johnston (1984) for Canada, Hietala (1990) for Finland, Crossland et al. (1991) for UK, Lamdin & Hiemstra (1993) for US, Hietala & Keloharju (1995) for Finland, Kato & Loewenstein (1995) for Japan, Lasfer (1995, 1996) for UK, Michaely & Murgia (1995) for Italy, Bartholdy & Brown (1999) for New Zealand, Bell & Jenkinson (2002) for UK, Elton et al. (2003), Graham et al. (2003) and Dhaliwal & Li (2006) for US. 4.1.2 SHORT-TERM TRADING HYPOTHESIS The second explanation advanced for the ex-dividend day price anomaly is shortterm trading hypothesis. Kalay (1982) asserts that short term traders trade around ex- 79
dividend day to earn profits and that their trading activity influences the ex-dividend day price behaviour. In a subsequent study, Lakonishok & Vermaelen (1986) state that, if short-term traders have major impact on the ex-dividend day stock prices, there would be net increase in trading volume around ex-dividend day. They have empirically tested this proposition for US and have found evidence in support of short-term trading hypothesis. They observe increased trading volume both before and after the ex-dividend day, and abnormal returns around the ex-dividend day. Support for short-term trading hypothesis is also found in Karpoff & Walkling (1988), Grammatikos (1989), Karpoff & Walking (1990), Robin (1991) and Stickel (1991) for US, Athanassakos & Fowler (1993) for Canada, Kadapakkam (2000) for Hong Kong, Bali (2003) for US, Bauer et al. (2006) for Canada and Dasilas (2009) for Greece. Evidence in support of short term dividend-capture trading by corporate traders is also reported in some studies. Corporate traders, a subset of short-term traders, are taxable investors who favour dividend income over capital gains and their trading activity involves the purchase of stock prior to the ex-dividend day and sale of the same soon after the ex-dividend day. Eades et al. (1994) for US found that the exdividend day returns are influenced by corporate dividend capturing. Similar result is evident in Bowers & Fehrs (1995), Siddiqi (1997), Koski & Scruggs (1998) and Naranjo et al. (2000) for US. Further, evidence of both short-term traders and corporate traders dominating the price determination on ex-dividend day is provided by Michaely (1991) for US and Walker & Partington (1999) for Australia. 80
4.1.3 MARKET MICROSTRUCTURE FACTORS The third explanation for the ex-dividend day price behaviour is market microstructure factors viz., price discreteness and bid-ask spread. Bali & Hite (1998) were the first to put forth the discreteness hypothesis as possible explanation for the price drop on ex-dividend day being less than the amount of dividend. They argue that the price drop on the ex-dividend day will be less than dividend, but greater than or equal to the dividend minus one tick 18. They have empirically tested this hypothesis for US and have found that on an average, the stock prices on exdividend day drop by an amount less than dividend. Similarly, Jakob & Ma (2004) for US report that the average price drop is less than average dividend and that tick size is an important factor in the ex-dividend day price drop anomaly. Hardin et al. (2007) for US also show that discreteness influences the ex-dividend day price behaviour. Frank & Jagannathan (1998) find evidence of price drop to dividend ratio being less than one for Hong Kong market and attribute this to bid-ask spread. They argue that investors who consider dividend collection and reinvestment as a nuisance would prefer to sell stocks on cum-dividend day and buy on ex-dividend day. On the other hand, market makers, who are active market participants for whom dividend collection and reinvestment is not cumbersome, will buy the stocks on cum-dividend day and sell on ex-dividend day. In such situations, most trades tend to occur at the bid price on cum-dividend day and at ask price on the ex-dividend day. This results in an increase in the stock prices on an average on the ex-dividend day and causes 18 It is the minimum price change allowed by stock exchanges. 81
the price drop to dividend ratio to be less than one. Evidence in support of bid-ask spread is also reported by Yahyaee et al. (2008) for Oman market. Previous empirical studies have also reported evidence in support of both taxation hypothesis and short term trading hypothesis. Such studies are Fedenia & Grammatikos (1993) for US, Menyah (1993) for UK, Han (1994) for US, Athanassakos (1996) for Canada, Wu & Hsu (1996) for US, Espitia & Ruiz (1997) for Spain, Bhardwaj & Brooks (1999) and Zhang et al. (2008) for US. Apart from the studies discussed above that provide evidence of the price drop to dividend ratio being less than one, there are also studies that provide evidence of price drop to dividend ratio being equal to one. For instance, Poterba (1986) for US find that, for stock dividends, the price drop is nearly equal to the full amount of dividend, whereas for cash dividend, the price drop is less than dividend. Barclay (1987) and Boyd & Jagannathan (1994) for US have found that stock prices fall by the full amount of dividend on the ex-dividend day. Further, Manakyan et al. (1993) for US market exhibit that, for high yield stocks, the ex-dividend day price drop is equal to the full amount of dividend, while for low yield stocks, prices drop by less than the dividend amount. From the review of literature, it is evident that no study has attempted to examine the ex-dividend day stock price behaviour in the Indian market. The present study attempts to fill this gap by examining the ex-dividend day stock price behaviour of dividend paying Indian stocks. 82
4.2 METHODOLOGY In order to examine the ex-dividend day stock price behaviour, various ex-dividend day price ratios are computed. This is followed by an examination of the price and volume behaviour of stocks around the ex-dividend day using event study methodology. 4.2.1 STOCK PRICE BEHAVIOUR ON EX-DIVIDEND DAY To examine the behaviour of stock prices on the ex-dividend day, the ex-dividend day price drop to dividend ratio or raw price ratio is computed. This ratio measures the price change from cum-dividend day to ex-dividend day in terms of dividend and is of the form given in equation (4.3). = 1 1 (4.3) where is the closing stock price on cum-dividend day; is the closing stock price on ex-dividend day; is dividend per share; is the tax rate on ordinary income; is the capital gains tax rate. In India, since neither dividends nor long-term capital gains are taxed in the hands of investors, the raw price ratio (RPR) given in equation (4.3) becomes, = 1 0 1 0 i.e. = = 1 (4.4) 83
As shown in equation (4.4), the RPR should be equal to one in the Indian market i.e. on the ex-dividend day, the stock prices should drop by the full amount of dividend ( = ). RPR is computed using three methods. Under the first method, RPR is computed using the closing price of stock on cum-dividend day ( ) and the closing price on ex-dividend day ( ) as = (4.5) The second method computes RPR using closing price of stock on cum-dividend day ( ) and opening price on ex-dividend day ( ) as = (4.6) The third method uses the closing price of stock on the cum-dividend day ( ) and the closing price on ex-dividend day, after adjusting the closing ex-dividend price for the stock market movements [ /(1 + )]. The ex-dividend day closing price is adjusted using the daily market return ( ) of Nifty Index. The RPR so computed is termed the market adjusted price ratio (MAPR) and is of the form given in equation (4.7). = [ /(1 + )] (4.7) To test whether the raw price ratio equals one in the Indian market, the following null hypotheses are framed: 84
H 0a : Mean value of RPR cc-ec = 1 H 0b : Mean value of RPR cc-eo = 1 H 0c : Mean value of MAPR = 1 A shortcoming of the ratios discussed above is that they suffer from heteroscedasticity [Eades et al. (1984), Barclay (1987), Michaely (1991), Dasilas (2009)]. This is because the stock price change from cum-dividend day to exdividend day is scaled by dividend that is different for different firms; hence, the weightage given to the change in price where dividend is low tends to be excessive. To overcome this problem, the price change on ex-dividend day is scaled by closing price on cum-dividend day and the resulting ratio is known as raw price drop ratio (RPDR), which is of the form given in equation (4.8). = (4.8) RPDRs are computed in three ways. The first method uses the closing price of stock on the cum-dividend day ( ) and the closing price on the ex-dividend day ( ) as shown in equation (4.9). = (4.9) In the second method, closing price of stock on cum-dividend day ( ) and opening price on ex-dividend day ( ) is used to compute RPDR as = (4.10) 85
The third method uses closing price of stock on cum-dividend day ( ) and exdividend day, after adjusting the closing ex-dividend day price for the market movements [ /(1 + )]. The market return ( ) of Nifty index is used to adjust the ex-dividend day closing prices. This ratio is termed the market-adjusted price drop ratio (MAPDR) and is computed as = [ /(1 + )] (4.11) Theoretically, the price drop ratios have a value equal to dividend yield. Dividend yield is computed as = (4.12) where is dividend per share and is the closing price of stock on cumdividend day. To test whether price drop ratios equal dividend yield, the following null hypotheses are framed: H 0d : Mean value of RPDR cc-ec = Dividend yield H 0e : Mean value of RPDR cc-eo = Dividend yield H 0f : Mean value of MAPDR = Dividend yield The ex-dividend day return or raw return is computed as = + = 1 (4.13) 86
In India, since dividend and long term capital gains are not taxed in the hands of investors, the ex-dividend day raw return should be equal to zero as shown below: = + = 0 0 = 0 1 0 (4.14) To examine whether the ex-dividend day raw return is equal to zero, the following null hypothesis is tested: H 0g : Mean of ex-dividend day raw return = 0 4.2.2 PRICE AND VOLUME BEHAVIOUR Event study methodology, as described in Wilkens & Wimschulte (2005), is employed to examine the price and volume behaviour of stocks on and around the ex-dividend day. This method enables to estimate and draw inferences about the impact of a particular event on the behaviour of stocks under consideration. The basic terminology of event study is discussed below: Event Day: Event day is the day on which the event under consideration occurs. In this study, ex-dividend day is the event day and is denoted as day 0. Event Period: Event period is the period over which a particular event is expected to have an impact on the stock under consideration. Event period includes the event day and some days prior to and after the event day. The event period for this study consists of the event day, and 5 days prior to and 5 days after the event day i.e. -5,,-1,0,+1,,+5. 87
Estimation Period: Estimation period is the period that is used to obtain the values of expected return (volume). The estimation period for this study starts from 170 days prior to and ends at 51 days prior to the announcement day (AD) of the event i.e. AD-170 to AD-51. 4.2.2.1 Abnormal return on and around ex-dividend day To examine how the prices of stocks behave on and around the ex-dividend day, abnormal returns are computed. Abnormal return of a stock is the difference between the stock s observed return and its expected return. Observed return is the actual return observed, while expected or normal return is the return that should have been observed had the event had not taken place. In this study, abnormal returns are computed employing three different methods viz., mean adjusted model, market adjusted model and market model. In each of these methods, the observed return of stock i on day t (, ) is computed as, =,, (4.15) where, is the price of stock i on day t;, is the price of stock i on day t-1; i = 1,2,..,N; N = no. of sample stocks. Each of the three methods differs the way in which the expected return is computed. The computation of abnormal return under each of these three methods is discussed below. 88
i) Mean Adjusted Model Under this model, abnormal return of stock i on day t (, ) is computed as, =, ( ) (4.16) where, is the observed return of stock i on day t; ( ) is the expected return of stock i; i = 1,2,.,N; N = no. of sample stocks; t = -5,,0,,+5. The expected return of stock i [ ( )] is computed as the average of the observed returns of stock i during the estimation period, ( ) = 1 120, (4.17) where, is the observed return of stock i on day s; i = 1,2,.,N; N = no. of sample stocks. The abnormal return specified in equation (4.16) thus becomes,, =, 1 120, (4.18) ii) Market Adjusted Model The expected return of stock i on day t [ (, )] in this model is equal to the market return on day t (, ) i.e. (, ) =, (4.19) where i = 1,2,.,N; N = no. of sample stocks; t = -5,,0,,+5. 89
Therefore, the abnormal return is,, =,, (4.20) iii) Market Model The market model assumes a linear relationship between return of a stock and the market return, as shown in equation (4.21)., = +, (4.21) where, is the return of stock i on day t;, is the market return on day t; is the market model constant; is the slope coefficient. For each stock i, using the stock s return and the market return during the estimation period, the relationship in equation (4.21) is estimated and the parameters and are obtained. Then, using these parameters, the expected return of stock i on day t [ (, )] is computed as, (, ) = +, (4.22) where and are the parameters estimated using the returns of stock i and that of the market during the estimation period;, is the market return on day t; i = 1,2,.,N; N = no. of sample stocks; t = -5,,0,,+5. Abnormal return is then computed as,, =, ( +, ) (4.23) 90
After calculating the abnormal returns, the mean abnormal return (MAR) is computed to draw an overall inference about the impact of the event on stock prices. MAR on day t ( ) is obtained as the average of abnormal returns of the sample stocks on day t, i.e. = 1, (4.24) where,, is the abnormal return of stock i on day t; i = 1,2,.,N; N is the no. of sample stocks; t = -5,,0,,+5. The null hypothesis being tested here is H 0h : mean abnormal return on ex-dividend day = 0. 4.2.2.2 Abnormal volume on and around ex-dividend day To examine the trading volume behaviour on and around the ex-dividend day, abnormal volume is computed. As a proxy for trading volume, no. of shares traded is considered. The trading volume of stock i on day t (, ) is log transformed as (1 +, ). Similarly, the trading volume of the market on day t (, ) is log transformed as (1 +, ). Two methods are used to compute abnormal volume viz., mean adjusted model and market model. i) Mean adjusted model Under this model, the expected volume of stock i [ ( )] is computed as the average of the trading volume of stock i during the estimation period ( ) = 1 120 (1 +, ) (4.25) 91
where, is the volume of stock i on day s; i = 1,2,.,N; N is the no. of sample stocks. The abnormal volume of stock i on day t (, ) is then computed as, = 1 +, ( ) i.e., = 1 +, 1 120 (1 +, ) (4.26) where, is the volume of stock i on day t; i = 1,2,.,N; N = no. of sample stocks; t = -5,,0,,+5;, is the volume of stock i on day s. ii) Market model Abnormal volume of stock i on day t (, ) is computed as, = 1 +, ( + (1 +, )) (4.27) where, is the volume of stock i on day t; i = 1,2,.,N; N = no. of sample stocks; t = -5,,0,,+5;, is the market volume on day t; and are the parameters obtained by estimating the following relation: 1 +, = + (1 +, ) (4.28) where, is the volume of stock i on day s; i = 1,2,.,N; N = no. of sample stocks; s = AD 170 to AD 51;, is the market volume on day s. Next, to draw an overall inference about the impact of the event on trading volume, mean abnormal volume on day t ( ) is computed. is obtained as the average of the abnormal volumes of sample stocks on day t, i.e. 92
= 1, (4.29) where,, is the abnormal volume of stock i on day t; i = 1,2,.,N; N = no. of sample stocks; t = -5,,0,,+5. The null hypothesis being tested is that H 0i : mean abnormal volume on ex-dividend day = 0. The statistical significance of under each of the models is tested using the following test statistic = where ( ) (4.30) = 1 120 119 (4.31) where is the standard deviation of MAR s of estimation period; is the on day u; is the on day v. The test statistics in equation (4.30) follow student t-distribution with 119 degrees of freedom. The test statistic to test the statistical significance of MAV is analogous to that of MAR. 4.3 EMPIRICAL RESULTS As initial sample, dividend payments made by the firms listed in CNX Nifty Index during the period January 2005 to December 2007 were considered. There were a 93
total of 211 dividend payments (events) during this period. Of this sample, only those events were retained that met the following criteria: i) The stock undergoing the event should not have undergone any other corporate events, such as mergers, rights issue, bonus issue, stock split or buy-back of shares, in the respective event and estimation period ii) For a stock that has paid dividend more than once during the sample period, it is required that the event period and estimation period corresponding to each of its events do not overlap with either the event period or estimation period of its other events. These criteria reduced the initial sample to 126 events. For each of these events, the corresponding stock prices, dividend per share, and number of shares traded and the price and number of shares traded of Nifty index over the event and estimation period were sourced from Prowess database and the official website of NSE 19. Those events for which the required data was not available were dropped, which yielded a final sample of 100 events 20. The observed mean values of the various ex-dividend day price ratios and raw return are presented in table 4.1. From table 4.1, it is evident that the mean RPR cc-ec, RPR cceo and MAPR are 0.16, 0.19 and 0.47 respectively. Further, the mean values of RPDR cc-ec, RPDR cc-eo and MAPDR are found to be 0.01, 0.01 and 0.01 respectively. The mean raw return is 1%. The ex-dividend day stock price behaviour is examined by testing whether the mean values of these measures significantly differ from their 19 www.nseindia.com 20 List of the ex-dividend day events is provided in Appendix IV-A 94
theoretical values. The theoretical and observed mean values of price ratios, price drop ratios and raw return are reported in table 4.2. Table 4.1: Observed Mean Values Ratios Mean RPR cc-ec 0.16 RPR cc-eo 0.19 MAPR 0.47 RPDR cc-ec 0.01 RPDR cc-eo 0.01 MAPDR 0.01 RR 0.01 Note: RPR cc-ec is the raw price ratio computed using closing price of stocks on cumdividend day and ex-dividend day; RPR cc-eo is the raw price ratio computed using closing price of stocks on cum-dividend day and opening price of stocks on exdividend day; MAPR is the market-adjusted price ratio computed using closing price of stocks on cum-dividend day and the market adjusted closing price on ex-dividend day; RPDR cc-ec is the raw price drop ratio computed using closing price of stocks on cum-dividend day and ex-dividend day; RPDR cc-eo is the raw price drop ratio computed using closing price of stocks on cum-dividend day and opening price on ex-dividend day; MAPDR is the market-adjusted price drop ratio computed using closing price of stocks on cum-dividend day and market adjusted closing price on ex-dividend day; RR is the raw return on ex-dividend day. 95
Table 4.2: Ex-Dividend Day Stock Price Behaviour Ratios Theoretical value Observed mean value t-statistic p-value RPR cc-ec 1.00 0.16-1.96 0.05 RPR cc-eo 1.00 0.19-3.73 0.00 MAPR 1.00 0.47-1.46 0.15 RPDR cc-ec 0.01 0.01-2.75 0.01 RPDR cc-eo 0.01 0.01-3.59 0.00 MAPDR 0.01 0.01-2.01 0.05 RR 0.00 0.01 2.95 0.00 Note: DPS is dividend per share; DY is dividend yield; RPR cc-ec is the raw price ratio computed using closing price of stocks on cum-dividend day and ex-dividend day; RPR cc-eo is the raw price ratio computed using closing price of stocks on cumdividend day and opening price of stocks on ex-dividend day; MAPR is the marketadjusted price ratio computed using closing price of stocks on cum-dividend day and the market adjusted closing price on ex-dividend day; RPDR cc-ec is the raw price drop ratio computed using closing price of stocks on cum-dividend day and exdividend day; RPDR cc-eo is the raw price drop ratio computed using closing price of stocks on cum-dividend day and opening price on ex-dividend day; MAPDR is the market-adjusted price drop ratio computed using closing price of stocks on cumdividend day and market adjusted closing price on ex-dividend day; RR is the raw return on ex-dividend day. As shown in table 4.2, the mean values of RPR cc-ec and RPR cc-eo are 0.16 and 0.19 respectively. Both these values are statistically significant and reject the null hypotheses H 0a and H 0b, implying that RPR cc-ec and RPR cc-eo is less than one. The mean value of MAPR is found to be 0.47, which is statistically insignificant and fails to reject the null hypothesis H 0c. 96
The mean values of RPDR cc-ec, RPDR cc-eo and MAPDR are 0.01, 0.01 and 0.01, respectively. All the three mean values are statistically significant and reject the null hypotheses H 0d, H 0e and H 0f. The mean value of raw return on ex-dividend day is 0.01, which is statistically significant and rejects the null hypothesis H 0g, implying that raw return on ex-dividend is significantly different from zero. These findings from table 4.2 reveal that the price drop to dividend ratios are less than one and that the ex-dividend day return is significantly different from zero. This implies that, in the Indian market, the drop in stock prices on the ex-dividend day is not equal to the full amount of dividend; rather, the stock prices drop by less than dividend. The tax hypothesis cannot be an explanation for such a price drop, since neither dividends nor long term capital gains are taxable in the hands of investors in India. Also, price discreteness cannot be attributed as possible explanation for the price drop, since the dividends paid are an integer multiple of tick size 21. Bid-ask spread is also ruled out as an explanation for such a price drop, since electronic trading system of NSE dispenses with the cumbersome process of dividend collection. Having ruled out these explanations as cause for the price drop on ex-dividend day being less than dividend, the next step is to investigate whether short-term trading hypothesis could explain the ex-dividend day price drop to dividend ratio being less than one. To this end, the stock price and trading volume behaviour on and around the ex-dividend day is examined. The behaviour of stock prices is examined by 21 If minimum dividend paid is an integer multiple of tick size, then price discreteness cannot explain the ex-dividend day price drop to dividend ratio being less than one (Borges, 2008). 97
computing the abnormal returns on and around the ex-dividend day, the results of which are reported in table 4.3. 98
Table 4.3: Abnormal returns on and around ex-dividend day N=100 Mean adjusted model Market adjusted model Market model Day MAR (%) t-statistic MAR (%) t-statistic MAR (%) t-statistic -5-0.08-0.37 0.13 0.81 0.19 1.16-4 -0.24-1.05 0.06 0.37 0.09 0.55-3 0.14 0.61 0.15 0.94 0.21 1.31-2 -0.01-0.05-0.28-1.71* -0.26-1.61-1 0.10 0.42 0.23 1.40 0.24 1.50 0 0.57 2.55** 0.45 2.74*** 0.47 2.95*** 1 0.22 1.00-0.07-0.40-0.08-0.49 2-0.12-0.55 0.00-0.02-0.04-0.22 3 0.07 0.29-0.11-0.65-0.04-0.24 4-0.07-0.29 0.02 0.13 0.05 0.30 5-0.46-2.05** -0.24-1.49-0.22-1.37 Note: MAR denotes mean abnormal return; day 0 denotes the event day i.e. the ex-dividend day; ***, ** and * denotes significance at 1%, 5% and 10% respectively 99
On the ex-dividend day, significant positive MAR of 0.57%, 0.45% and 0.47% is observed under the mean-adjusted model, market adjusted model and market model respectively. These values lead to the rejection of the null hypothesis H 0h, implying that there is significant positive abnormal return on the ex-dividend day. Next, the abnormal volumes on and around the ex-dividend day are computed to examine the trading volume behaviour and the results are presented in table 4.4. Significant positive abnormal volume of 0.18 and 0.22 are observed on the exdividend day under the mean-adjusted model and market model respectively. This result leads to the rejection of the null hypothesis H 0i. Significant positive abnormal volumes are also observed on days -1 and +1 under both the models. On day -1, MAV of 0.11 and 0.13 are observed under the mean adjusted model and market model respectively. On day +1, MAV is found to be 0.15 and 0.18 under the mean adjusted model and the market model respectively. These findings indicate that there is increase in trading volume on and around the ex-dividend day, which is consistent with the notion that there is short-term trading around ex-dividend day. 100
Table 4.4: Abnormal volume on and around ex-dividend day N=100 Mean adjusted model Market model Day MAV t-statistic MAV t-statistic -5 0.04 0.63 0.04 0.62-4 -0.09-1.31-0.02-0.27-3 -0.08-1.15 0.00-0.01-2 -0.10-1.48-0.05-0.74-1 0.11 1.67* 0.13 2.19** 0 0.18 2.62*** 0.22 3.60*** 1 0.15 2.18** 0.18 2.99*** 2-0.05-0.70 0.07 1.18 3-0.06-0.91 0.05 0.84 4-0.07-1.10 0.01 0.10 5-0.09-1.36-0.01-0.14 Note: MAV denotes mean abnormal volume; day 0 denotes the event day i.e. the ex-dividend day; ***, ** and * denotes significance at 1%, 5% and 10% respectively 101
Thus, the stock price and trading volume behaviour on and around the ex-dividend day lend support to short-term trading hypothesis as possible explanation for the exdividend day price drop to dividend ratio being less than one. 4.4 CONCLUSION In a perfect capital market, it is expected that stock prices should drop by the full amount of dividend on the ex-dividend day. Empirical examination of the behaviour of stock prices on the ex-dividend day has, however, found that stock prices do not drop by the full amount of dividend on the ex-dividend day; rather, the stock price drop is largely found to be less than dividend. Taking lead from this, an attempt has been made in this chapter to examine the ex-dividend day stock price behaviour in the Indian market. To this end, a total of 100 dividend payments made during the period 2005-2007 is considered. The results of empirical examination indicate that the price drop to dividend ratio is less than one, implying that the stock prices drop by less than dividend on the ex-dividend day. Of the available theoretical explanations, neither taxation nor market microstructure explanations could be attributed as a cause for such a price drop; hence, an attempt is made to examine whether short-term trading hypothesis could explain such a price drop. To test this, event study is employed. The results of event study reveal that there is significant positive abnormal return on the ex-dividend day. Significant positive abnormal volumes on and around ex-dividend day are also observed. These results lend support to the proposition that there is short-term trading activity around the exdividend day. Thus, short-term trading hypothesis could be attributed as an explanation for the ex-dividend day price drop to dividend ratio being less than one. 102