Study on the Market Response to Stock Splits
Disclaimer: This work has been submitted by a student. This is not an example of the work written by our professional academic writers. You can view samples of our professional work here.
Any opinions, findings, conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of UK Essays.
Published: Thu, 01 Mar 2018
The market response to stock splits is investigated with the dataset from an emerging country – India for period 2006 March 2009. study reports significantly positive abnormal returns on day of split execution and next trading day. regression analysis suggests that the reaction can be attributed to liquidity hypothesis. The postsplit period experiences abnormally high negative which wipes out any positive gain during split execution. This seems mostly explained by presplit price increase and size of firms suggesting that the have experienced a in period are ones suffer worst returns.
In theory, stocksplits are cosmetic corporate events as they simply increase the number of outstanding shares and decrease the price of each outstanding share. Hence, there should be no significant effect on the value of the firm. A stock split does not change the revenue or assets of a company. So, stocksplit should cause no change in price other than the adjustment warranted by the split factor. There should also be no change in distribution of stock returns around exdates of stock splits. Exdate refers to the date on or after which a security is traded without a previously declared dividend or distribution. However, empirical evidence suggests that the market generally reacts favorably to stock splits. In different developed markets, for instance, UK and US, significant positive abnormal returns and increase in variance and volumes of trade have been documented around stock split announcements as well as exdates. The contradiction between theory, which expects no change in firm value consequent to stock splits, and the reality, with scores of evidence of significant market reaction, triggers the present study.
In February 1981, the Indian ministry of finance issued a guideline that denomination of equity shares be fixed uniformly at Rs.10, and that the denomination of the then existing shares other than Rs.10 be converted into denomination of Rs.10. In another guideline in January 1983, the Indian government clarified that denomination of shares of Rs.100 need not be changed to denomination of Rs.10, i.e. shares of all companies were required to be in denominations of Rs.10 or Rs.100 only. Even so, several companies converted the denomination of shares of Rs.100 into that of Rs.10 on the grounds that it generated better liquidity, as also a higher value for the shares. However, in March 1999, the Securities and Exchange Board of India (SEBI) decided, “with the objective of broadening
the investors’ base,” to dispense away with the requirement of standard denomination of Rs.100 or Rs.10 and gave freedom to companies to issue shares of any denomination but not below Re.1. Companies that had issued shares of the face value of Rs.10 or Rs.100 were also permitted to avail of this facility by consolidation or by splitting their existing shares. To reap benefits of splitting, a number of existing listed companies having denomination of Rs.100 or Rs.10 have split their stocks into different denominations, e.g., Re.1, Rs.2 or Rs.5, etc. These recent changes in the India’s regulatory environment offer a unique opportunity to gain further insight into the stock splits with reference to their effects on variables like stock prices, return, volatility, and trading volume. With the increased integration of international markets in general and a wave of liberalization and globalization, the importance of understanding these stock events has increased dramatically. Further, there are different capital gains tax laws in India. Under these circumstances, splits may have different effects contrary to what has been reported in various literatures. Furthermore, compared to the world’s major stock exchanges, there are proportionally more small firms listed on Indian exchanges; consequently, many firms are thinly traded. Hence, these differences between Global and Indian markets necessitate studying split events in India.
The results of the present study shows significantly positive cumulative abnormal returns on and the next trading day after split execution, following which there is a major decline in share prices which wipes out most of the gain of the execution period. The signaling hypothesis and the trading range hypothesis do not seem to provide any explanation for the significant CAR around execution date, while the liquidity hypothesis seems to contribute significantly towards the positive CAR occurring on and immediately after the execution. The small firm hypothesis also
– 6 –
shows weak explanatory power for the change in wealth. The post
execution negative reaction is mostly explained by run up of stock prices preceding the execution, implying that the stock split has induced a revision of stock’s fundamentals, probably bringing prices to a more fundamental level.
– 7 –
2. . Literature review
There have been numerous researches on the effect of stock splits on different parameters of capital markets. Fama etal (1969) has been the pioneering study to examine the share price performance of splitting firms. Although the economic literature has not yet found a definitive explanation for either the abnormal returns observed around the announcement and execution dates, or the reasons why managers decide to split, different explanations, not necessarily mutually exclusive have been proposed. The more prominent hypotheses are the signaling hypothesis, the trading range hypothesis, the liquidity hypothesis and the neglected firm hypothesis. One such research paper advocates considering the three different market efficiencies (weak form, semistrong form, and strong form) that the investor can make an above normal return by relying on public information impounded in a stock split announcement. This study agrees that according to the semistrong form market efficiency, the stock split announcement do impact the company stock price. The study done by Desai, Jain (1997) elaborates more on longrun performance of common stock following stock splits announcement and hence concludes that the capital market doesn’t fully react to the information conveyed in the stock split announcement. Considering the ignored studies of small firms, the paper examined firm portfolio of different sizes and more diversity in terms of industries. Taking a large sample of stock information for a period of 1976 – 91, the research paper concluded that the market does not incorporate the full effect of the stock split announcement in the month of announcement. It is evident that managers believe that stock split results in optimal trading price of a stock that attract small investors and hence enhances liquidity. Joshipura (2008) studied the price and liquidity effects associated with stock split surrounding its announcement and execution
– 8 –
dates in Indian stock exchanges. The results suggested that though there were some positive abnormal return associated surrounding announcement and execution dates of the stock split, but it reverses in just a few days after the event dates, and ultimately generates significant negative abnormal return in slightly longer postexecution window. It also found that there was a significant improvement seen in liquidity surrounding announcement and execution dates of stock split. Desai and Nimalendran (1998) examined the effect of the change in trading activity after stock splits on volatility and spread. The results of the study show that the increase in volatility cannot be attributed solely to microstructure biases arising from the bidask bounce and price discreteness. Even after correcting for these two biases, the study found a significant increase in volatility after the split. The study also found an increase in the number of trades after the split, and the increase in the biasadjusted volatility was positively related to this increase in the number of trades. The study decomposed volatility into transient and permanent components and found that both components of volatility increase after the split. Attributing transient volatility to noise traders and permanent volatility to informed traders, the study suggested that trading by both types of traders increases after the split. Ikenberry, etal (1996) discusses that splits are used to move stock prices into a trading range to increase liquidity and that they are used by management as a signal of positive private information. The study found evidence that is consistent with the view that splits are typically used to realign stock prices to a normal trading range. The study also confirmed that splits convey favorable information, thereby validating the signaling hypothesis. It was found that market reaction was greater for small firms, low booktomarket firms and firms splitting to low share prices. The study also found an inverse relationship between the presplit run up and postsplit excess returns, suggesting that the results were not attributable to momentum.
– 9 –
There are various studies devoted to studying the effect of stock splits in specific geographies. Asquith, etal (1989) examined stock splits in the US market and found that stock splits do convey earnings information. The results indicated that firms split their shares after a significant increase in earnings. Before the stock split announcement, the market expects these earnings increases to be temporary. The split announcement leads investors to increase their expectations that the past earnings increases are permanent. The study also found that the market’s reaction to the split announcement cannot be attributed to expectations of either future earnings increases or nearterm cash dividend increases. Elfakhani, etal (2003) examined the market behavior surrounding stock split announcements in the Canadian market for the 1977–1993 period and the effect of the 2year before compared to the 2year after the announcement. Using the event study methodology, the findings indicated that positive abnormal returns exist on both the announcement days (0,1) and the 11day period surrounding stock split announcements. The results also showed that following the split event, bidask spreads decrease, while both trading volume and the number of transactions increase thus suggesting that split events enhance liquidity. Further, the study observed that earnings grow in the 2year period following split events, thus implying that split events signal future performance of the firm. Wulff (2002) investigated market reaction to stock splits using a set of German firms and in line with the US findings, found significant positive abnormal returns around both the announcement and the execution day of German stock splits. The study also observed an increase in return variance and in liquidity after the exday. The study found that abnormal returns around the announcement day are not related to changes in liquidity, but (negatively) to firm size, thus lending support to the neglected firm hypothesis. Despite noting a substantial increase in liquidity after the split, the study did not find support for the liquidity
– 10 –
hypothesis. Menendez, etal (2003) analyzed the motivations and valuation effects of stock splits in the Spanish market. The findings of the study suggest that splitting firms present a presplit stock price above the normal trading range, and that, after the split, the number of transactions and the average transaction size increase significantly. Moreover, positive abnormal returns are observed around the announcement dates and around the exdate. For the latter, however, these positive wealth effects are outweighed by the negative abnormal returns observed closely afterwards. The study found that liquidity, or the optimal trading range hypotheses prevailed over other hypotheses as an explanation for stock splits in the Spanish market. The findings of the study suggest the main reason behind a stock split and for the positive market reaction around the stock split announcements is a higher share price than the normal trading range. The reduction of this higher price seems to attract small investors and thus significant increases in the number of transactions and reductions in the trading volume per transaction after the split are observed, without there being any significant variation in the volume of shares traded. This adjustment of the firm’s stock price to a normal trading range is valued positively by investors. Most of these studies are concentrated mainly around market reaction at the announcement date. In a study on the UK equity market, specifically concentrating on the exsplit date, Staikouras etal, (2009) has documented positive abnormal returns on and around the exsplit date which are partially predictable using the publicly available information prior to the exsplit date. The study also observed a persistent increase in the post split volatility of the stocks in the UK equity market with this increase being better explained by the daily trading volume. This is in contrast to the US findings where the daily number of trades was found to better capture the increase in volatility.
– 11 –
In this study, the market response to stock splits is investigated with the dataset from an emerging country – India, which is distanced from the west in terms of geographical location, economic development, institutional and legal framework. Not much is available in the Indian context with a focus on the exsplit date, so far, except for the commendable work by Mishra (2007), which documents negative effect on price and return of stocks following splits. The study also reports a positive effect on volatility and trading volume following the split events. The present paper tries to provide a few additional insights on the issue and therefore, differs from Mishra’s (2007) study in the following ways. Firstly, an attempt is made to explain the significant cumulative abnormal returns around the split execution dates with the help of regression analysis. Secondly, the independent variables cover issues like small firm hypothesis, price run up, deviation of price from market average, which are unexplored in his paper. Thirdly, the data set of the present study covers the period post Mishra’s study, i.e., from 2006 to March 2009.
HHHHyyyyppppooootttthhhheeeesssseeees s ffffoooorrrrmmmmuuuullllaaaattttiiiioooon n
TTTThhhhe e ssssiiiiggggnnnnaaaalllliiiinnnng g hhhhyyyyppppooootttthhhheeeessssiiiis s The signaling hypothesis proposes that, in a scenario of asymmetric information between managers and investors, managers may use stock splits to signal positive information to the market about the firm’s future expectations. The presence of positive abnormal returns around the stock split announcement that is found in many empirical studies provides evidence for the signaling hypothesis.
– 12 –
Trading range hypothesis
According to the optimal trading range hypothesis, stock splits are used as tools to realign the share price to a desired price range so that it is more affordable for small investors to buy round lots of shares. If the presplit share price is at a high level, then a stock split is justified for improving the marketability of the shares. Empirical findings suggesting an increase in the daily number of transactions after the split do not reject this optimal range hypothesis.
TTTThhhhe e lllliiiiqqqquuuuiiiiddddiiiitttty y hhhhyyyyppppooootttthhhheeeessssiiiis s The management’s motivation to bring the share price to an optimal trading range arises from the desire to improve liquidity. According to literature there is an observed increase in trading volume during the postsplit period, and hence provide support for the liquidity hypothesis of stock splits. Staikouras etal, (2009) in their study of the UK equity market document a strong and positive relationship between the measures of trading activity and the returns’ volatility over the preand postsplit horizons.
SSSSmmmmaaaalllll l ffffiiiirrrrm m hhhhyyyyppppooootttthhhheeeessssiiiis s Small firm or neglected firm hypothesis suggests that since the smaller firms have fewer announcements published in the financial press, the split announcement is expected to create greater market interest than it would be in case of larger firms. So, small firms may have an incentive to adopt the stock splits to grab more attention.
Based on the discussion above, we can lay down the objectives of this study. The study proposes to, using data from the Indian stock market, examine the presence of positive abnormal returns over the stock split
– 13 –
period and, if found to be true, to study whether the returns can be
explained using any of the hypotheses mentioned above.
We formulate the following hypotheses:
HHHH1111: There is no significant abnormal return around the exsplit date.
HHHH2222: If H1 rejected, returns are identical for all firms in sample.
HHHH3333: and H2 the abnormal observed around event window [1,+1] can be attributed to publicly available information based on one or more of theoretical hypotheses above.
HHHH4444: If H3 is true, a similar explanation can be made using this data for different time horizons around the exdate.
An event study framework is employed to test the above hypotheses. An OLS regression model is used for determining the factors for the occurrence of abnormal returns across the event window.
– 14 –
3333. . DDDDaaaatttta a aaaannnnd d MMMMeeeetttthhhhooooddddoooollllooooggggy y
3333….1 DDDDaaaatttta a CCCCoooolllllllleeeeccccttttiiiioooon n
The basic sample is comprised of all Bombay Stock Exchange (BSE) equity stocks that have split between January 2006 and March 2009. The National Stock Exchange website was used to download list of stocks that have undergone a stock split in this period. There were a total of 151 stock splits during the period. All financial data series for these stocks like daily closing adjusted prices, market capitalization, trading volume and market indices were downloaded from Thomson DataStream.
The following criteria have been applied to include a company in the sample. i) The stock price data is available for 260 days prior to the exsplit date. ii) Data for 260 days are available for the postsplit period. iii) Where a stock has split more than once in the sample period, the first exsplit date was considered. iv) Other required financial information is available.
After filtering on the basis of the above criteria, the number of firms on which the analysis could be carried out was 99.
3333….2 DDDDaaaatttta a AAAAnnnnaaaallllyyyyssssiiiis s
Table 1A & Table 1B below show summary statistics of the sample stocks used for this study.
There is an even distribution of stock splits in each year of the sample period indicating normal stock split activity in the Indian equity market
– 15 –
for the given period. However, analysis on a monthly basis reveals that
August, September and October are the most active months for stock splits, possibly indicating a preference by firms to execute the split around that time. More than 40 percent of the firms in the sample have the stock split in this period.
5:1 split is the most common split ratio (57 firms) in the sample followed by 10:1 and 2:1 split ratios. For the 4:1, 6:1 and 5:2 split ratios, there is only one stock in the sample period. Therefore, summary statistics for these stocks were not calculated as any observations made would be a result of a very firm specific performance and not a general conclusion.
The average price for the 2:1, 5:1 and 10:1 split sizes are Rs. 229.99, Rs.
192.30 and Rs. 215.27 respectively. No conclusive relation between the stock price and the split ratio can be inferred from the maximum and minimum values shown below. The average marketcap for the 2:1, 5:1 and 10:1 split sizes are Rs. 13068.56 million, Rs. 57129.56 million and Rs. 87126 million respectively. The average market capitalization is observed to increase with higher split ratio possibly indicating that the largecap stocks are the ones that usually opt for the higher split ratio.
– 16 –
TTTTaaaabbbblllle e 1111AAAA: : SSSSttttoooocccck k SSSSpppplllliiiit t ssssaaaammmmpppplllle e ddddiiiissssttttrrrriiiibbbbuuuuttttiiiioooon n
TTTTaaaabbbblllle e 1111BBBB: : SSSSttttoooocccck k SSSSpppplllliiiit t ssssuuuummmmmmmmaaaarrrry y ssssttttaaaattttiiiissssttttiiiiccccs s
– 17 –
3333….3 MMMMeeeetttthhhhooooddddoooollllooooggggy y
EEEEvvvveeeennnnt t bbbbaaaasssseeeed d ssssttttuuuuddddy y Event studies start with hypothesis about how a particular event affects the value of a firm. The hypothesis that the value of the company has changed will be translated in the stock showing an abnormal return. Coupled with the notion that the information is readily impounded into prices, the concept of abnormal returns (or performance) is the central key of event study methods. How does a particular event affect the value of a company? We must be careful because at any time we observe a mixture of market wide factors and a bunch of other firm events. To correctly measure the impact of a particular event we need to control for those unrelated factors. The selection of the benchmark to use or the model to measure normal returns is therefore central to conduct an event study. The empirical model can be stated as follows: when an event occurs, market participants revise their beliefs causing a shift in the firm’s return generating process. For a given security, in non event periods,
Rt = xt B + et while in event periods,
Rt = xtB + FG + et
Rt is the return of the security in period t, xt is a vector of independent variables (for example the return of the market portfolio) in period t, B is a vector of parameters, such as the security beta, F is a row vector of firm characteristics influencing the impact of the event on the return process. G is a vector of parameters measuring the influence of F on the impact of
– 18 –
the event and et is a mean zero disturbance term possibly differing in
event and non event periods. Hypotheses usually centre on the parameters that measure the influence of the event (G) and most of the times F is set to unity. The null hypothesis is that such an event has no impact on the return generating process. Event study methods are the econometric techniques used to estimate and draw inferences about the impact of an event in a particular period or over several periods. The most common approach involves three steps: (1) Compute the parameters in the estimation period; (2) Compute the forecast errors (and obtain variance/covariance information) for a period or over an event window; aggregate across firms and infer about the average effect; (3) Regress crosssectional abnormal returns on relevant features of the stock supposed to influence the impact of the event.
In this study, the event is the split execution date, defined as day 0. The event splits the sample into two sets – the presplit period and the postsplit period. The presplit period considered in this study is a period of 260 days prior to the event date (260 days to 1 day) and the postsplit period is the period of 260 days after the event date (+1 day to +260 days). This leads to a total period of 521 trading days data for each stock (including the split date) centered around the event date for that stock. In essence, all stocks are aligned according to their event timeline. The estimation window is the 220 day period from 260 to 41 trading days. A similar event based alignment of data was performed for the other financial data namely market capitalization, market index and trading volume. In this study, the benchmark index chosen for running the regression for the market model is the BSE 100 index. The index price was also aligned according to the split date for each stock to obtain comparable market
– 19 –
return at and around the event date. The Brown and Warner (1985)
methodology is applied to test for the significance of abnormal returns.
Regressions to estimate the parameters of the market model
The standard single index asset pricing model is used to estimate the market parameters ( β and c). The market model used to estimate the parameters is given as below:
Ri = βRm + c
– expected return of stock i
– market return
constant of regression
Cite This Work
To export a reference to this article please select a referencing stye below: