JIUBJ Market Reaction Around Ex-Dividend Date

. A dividend is a component of returns for the investors besides capital gains. Hence, its payment becomes attractive for investors to own the stocks. According to the informational content dividend theory, the market reacts positively around the ex-dividend dates when they enhance their amount and vice versa. Motivated by the debatable evidence from many previous studies, this research wants to prove this theory by classifying the companies into two groups. Group one consists of the companies going up dividends per share. Group two contains the opposite ones. This research population is the consistent companies becoming members of the LO45 index between 1 February 2015 and 31 January 2019, where the number is 21. To calculate the total samples, we use the Slovin formula and get 17 companies. To take these companies, we use the simple random sampling method. Then, we obtain 49 and 29 ex-dividend dates for the companies elevating and decreasing dividends, one-to-one. To examine the proposed hypotheses, we utilize the one-sample t-test. Generally, this study disputes the informative content of dividend theory. In its place, this study supports the informational market inefficiency in the semi-strong form, evidenced by the lengthy market reaction. The adverse market reaction still occurs for the companies elevating dividends on the fourth day after the ex-dividend date. However, the positive reaction still happens for the companies reducing them on the third day after this date.


INTRODUCTION
Dividends are a company incentive for investors because of the willingness to let their money on a risk (Black, 1976). When the company pays cash for dividends, the ex-dividend date will become their attention (Nasir & Sari, 2014). Investors purchasing stocks on or after this date do not obtain them. Instead, they have to buy shares before this date to acquire dividends (Hanafi, 2017). The research about market reaction around the ex-dividend date is attractive for the previous scholars (Sularso, 2003;Virda & Karlina, 2009;Noviyanti, Medyawati, & Yunanto, 2013;Yang & Wu, 2015;Ozo & Arun, 2019;Tamara, Munir, & Maria, 2020). Excitingly, market reaction around the exdividend date of the group consisting of firms increasing dividends can be positive (Virda & Karlina, 2009 negative (Sularso, 2003;Noviyanti et al., 2013). Similarly, the market reaction of the group consisting of companies reducing dividends can be positive (Sularso, 2003), negative (Virda & Karlina, 2009;Ozo & Arun, 2019;Tamara et al., 2020), and insignificant (Yang & Wu, 2015). Regarding the various outcomes of those studies, this study wants to prove the companies' market reaction in paying dividends by utilizing the consistent members of the LQ45 index between 2015 and 2018. Additionally, the utilization of the firms forming the LQ45 index is due to the speedy adjustment of their price (Utama, 1992), as the efficient capital market theory investigation requires (Hartono, 2017).
The informational content of dividend theory enlightens the change in dividends is associated with the view modification of managers on prospective earnings (Miller & Modigliani, 1961). According to the investor perspective, the companies elevating the dividends deliver a positive signal in the future; therefore, investors boldly buy the stocks at the premium price. Conversely, if the companies diminish these rewards, the investor will have a negative image. Hence, investors sell their shares; thus, their price decreases (Gitman & Zutter, 2012). This theory is supported by Virda & Karlina (2009) by studying the market reaction around ex-dividend dates in Indonesia. They find that when companies cut the cash dividend, an adverse market reaction occurs. Similarly, Oman, Al-Yahyaee, Pham, & Walter (2011) confirm this theory by showing that when companies raise the cash dividend, a positive market reaction happens, and vice versa. Correspondingly, Ozo & Arun (2019) prove that the increase and the decrease in dividends make the market react positively and negatively, one-to-one, during ex-dividend dates. Partially, the empirical evidence backing this theory comes from (Sularso, 2003;Virda & Karlina, 2009;Al-Yahyaee, 2011;Ozo & Arun, 2019), showing a positive market reaction of the companies elevating dividends. Conversely, the study of (Virda & Karlina, 2009;Ozo & Arun, 2019) finds that the decreasing market reaction happens when the companies cut them. Besides, when investigating the market reaction of the firms elevating the dividend cut only, Tamara et al. (2020) find that an undesirable response. Based on this explanation, the first hypothesis formulated is in this way: H1: A desirable market reaction around the ex-dividend date exists when the firms increase dividends. H2: An undesirable market reaction around the exdividend date exists when the firms decrease dividends.

METHOD
A market reaction, measured by an abnormal return, becomes the research variable. By definition, the abnormal return gets calculated by subtracting the actual return in the window period from the expected return (Hartono, 2017). Moreover, this expected return is calculated by mean return in the estimation period. By referring to Dasilas & Leventis (2011), the number of days used in the estimation period is 100, started from t-105 until t-6. Additionally, by denoting (Marfuah, 2006;Virda & Karlina, 2009;Dasilas & Leventis, 2011;Do & Hieu, 2018), the total days utilized in the window period is 11, started from t-5 to t =+5 (see Figure 1).  (2017) The population of this study is the companies consistently becoming members of the LQ45 index, having an ex-dividend date from 1 February 2015 to 31 January 2019. After observing the data, this study finds 21 suitable companies with the names in Table 1. This study exhausts a formula created by Slovin in equation one with the error margin (em) as high as 10% to get the total representative samples (TRS) of the population (TP).
The variable utilized in this research is single: market reaction measured by abnormal return; therefore, following Hartono (2012), the hypotheses constructed are descriptive. Thus, by mentioning Ghozali (2016), this study employs a one-sample t-test to check them based on the output of IBM SPSS 19. To statistically verify a positive or negative average abnormal return in this output, we compare the probability (1-tailed) of the t-statistic with the relaxed significant level as high as 10%. a. If this probability is the same or above 10%, the positive or negative sign is meaningless; therefore, the abnormal return does not exist.
b. If this probability is below 10%, the positive or negative sign is meaningful; therefore, the abnormal return is available according to its sign. As the requirement, the normality of the abnormal return has to be achieved to support a one-sample t-test (Hartono, 2012). Following Ghozali (2016), we use the Kolmogorov-Smirnov Z-statistic to examine this assumption by comparing the asymptotic significance (2-tailed) of Z-statistic with the 1% tightened significance level. Furthermore, to test the null hypothesis declaring that the abnormal return follows a normal distribution, we refer to this instruction: (a) If the significance (2-tailed) is higher than or the same as 1%, the null hypothesis is acknowledged; and (b) If the significance (2-tailed) is lower than 1%, the null hypothesis is refused. Table 2 demonstrates the Kolmogorov-Smirnov testing result examining the normality of abnormal return before, at, and after the ex-dividend date for the companies elevating dividends. Furthermore, the asymptotic significance (2-tailed) for abnormal return in this table is between 0.013 and 0.394. These values are still higher than the significance level of 1%; therefore, the null hypothesis stating abnormal return following normal distribution is accepted.  Table 3 shows the Kolmogorov-Smirnov testing result examining the normality of abnormal return before, at, and after the ex-dividend date for the companies cutting dividends. Furthermore, the asymptotic significance (2-tailed) for abnormal return in this table is between 0.279 and 0.997. These values are still higher than the significance level of 1%; therefore, the null hypothesis stating abnormal return following normal distribution is accepted.  Table 4 present the result of a one-sample t-test for abnormal return before, at, and after the ex-dividend date for the companies enhancing dividends. It can be seen that a negative abnormal return exists on: (a) Day one before the ex-dividend date (see the negative mean of 0.005945 and probability (1-tailed) of 0.091, lower than 10%); (b) The ex-dividend date (see the negative mean of 0.011231 and probability (1-tailed) of 0.001, lower than 10% on day 0); (c) Day one after the exdividend date (see the negative mean of 0.006392 and probability (1-tailed) of 0.029, lower than 10%); and (d) Day four after the ex-dividend date (see the negative mean of 0.006216 and probability (1-tailed) of 0.079, lower than 10%).  By the facts presented in Table 4, we reject the first hypothesis stating that a desirable market reaction around the ex-dividend date exists when the firms increase dividends. Table 5 presents the result of a onesample t-test for abnormal return before, at, and after the ex-dividend date for the companies cutting dividends. It can be seen that a negative abnormal return exists on: (a) Day one before the ex-dividend date (see the positive mean of 0.00564 and probability (1-tailed) of 0.065, lower than 10%); (b) The ex-dividend date (see the positive mean of 0.01069 and probability (1-tailed) of 0.011, lower than 10% on day 0); and (c) Day three after the ex-dividend date (see the positive mean of 0.01540 and probability (1-tailed) of 0.000, lower than 10%). By the facts presented in Table 5, we reject the second hypothesis stating that an undesirable market reaction around the ex-dividend date exists when the firms decrease dividends. Based on statistical testing, an undesirable market reaction exists when the companies elevate dividends, which supports Sularso (2003) and Noviyanti et al. (2013). However, a desirable market reaction happens when the companies cut dividends, and this circumstance verifies the work of (Virda & Karlina, 2009;Ozo & Arun, 2019;Tamara et al., 2020). Also, this study confirms the study of Warrad (2017), Hadianto et al. (2020, declaring that the more dividends compensated for public shareholders, the more diminishing stock price. By this evidence, we refuse the informational content of dividend theory. Instead, we acknowledge Brealey, Myers, and Allen's perspective (2020), stating that the public shareholders prefer the companies cutting dividends to the firms increasing dividends. By reducing them, these shareholders believe that the companies can allocate their money to invest in promising projects to increase their value in the future. Additionally, this evidence supports the informational market inefficiency in the semi-strong form, reflected by the market reaction after the event: (a) For the first event, the companies enhancing dividends, the market reaction is still available on the fourth day after the exdividend date; and (b) For the second event, the companies reducing dividends, the market reaction is still is obtainable on the third day after the ex-dividend date.

CONCLUSION
This study goals to check the market reaction around the ex-dividend date for the companies chosen consistently to be the members of the LQ45 index from 1 February 2015 to 31 January 2019. By employing 49 and 29 ex-dividend dates for the companies that increase and decrease dividends, respectively, we test the abnormal return to measure the market reaction. After checking the market reaction, we infer that a desirable market reaction is available when companies cut dividends and vice versa. Also, the extended market reaction in two events, i.e., the companies enhancing and reducing dividends, shows that the market is informationally inefficient in the semi-strong form. Although this study result informationally supports the inefficient market, some restrictions are still available: Firstly, this study uses a small sample size: 17 companies with 78 observations of abnormal returns. Secondly, this study utilizes the average-adjusted model to calculate the expected return as one of the components of abnormal return determination. As the first restriction solution, the subsequent scholars are expected to use more companies forming the specific index: the KOMPAS 100. Moreover, as the second restriction solution, the subsequent scholars can use the expected return based on the single-index model and its adjusted one.