The information content and stock return behavior around the stock splits--evidence from India.

1. Introduction

Stock splits have mystified the researchers in finance for long time. A split merely increases the number of shares outstanding by subdividing the existing number of shares into a greater number of units. It seems purely a cosmetic event with no real economic consequences. If stock splits were purely cosmetic it would be surprising to find real effects associated with them. The real effects are associated with both the announcement of the splits and the execution of the split events.

Stock splits are purely a book entry which does not increase a firm's cash flows, but it involves several costs. The costs include administrative cost, printing costs, legal expenses etc. These costs have to borne by the shareholders. Further after the splits the relative bid-ask spread increases, which lead to increase in true transaction costs. All these costs are inversely related to stock price.

Theoretically, this raises a valid question of why do companies splits their stocks, if there is no gains out of it. Moreover the company has to borne some costs. A large number of companies actively splits their stocks across the globe must expect some gains out of it otherwise why they do so?

Several theories in this area of research have been developed over last thirty years. Some researchers argue in favour of optimal trading range hypothesis Copeland (1979). As per this theory the managers of the splitting firms bring down the stock price to an affordable price range for the non institutional investors through stock splits. Lakonishok et al. (1987) provide some empirical evidence on the existence of an optimal trading range in the US. Conroy, Harris and

Benet (1990) in support of Copeland (1979) conclude that "managers seem to engineer splits to return their company' stock price to a particular level that is remarkably stable over time".

The next theory is liquidity hypothesis. According to this theory, the researchers hypothesised that the decrease in stock price to an optimal trading range would be helpful for the non institutional and wealth constrained investors. Hence, the activity and the number of small investors will increase after the split. Lamourexu and Poon (1987) first give the evidence in favour of increase of number of small investors after ex splits. Angel (1997) finds increase trading activity by small investors after the splits. Schultz (2000) also confirms the above findings and shows substantial increase in small orders following the splits.

Lakonishok and Lev (1987) have postulated the signaling hypothesis as a possible explanation of stock splits. According to signaling hypothesis, managers announce stock splits to communicate favorable private information about the current value of the firm. Managers are the actual decision makers about the future operations and the investment decision of the firm. The existence of positive abnormal returns surrounded to split announcements is consistent with the signaling hypothesis. [Grinblatt et al. (1984), Asquith et al. (1989) and McNichols and Dravid (1990)].

Previous studies have found that these noneconomic events do have effect on shareholders wealth. Grinblatt et al. (1984) document positive abnormal returns of 3.3% over the two days event period surrounding the split announcements. Researchers find evidence of significant abnormal returns for the splitting stocks in the long run. Desai and Jain (1997) show splitting firms on average earn significant abnormal returns of 7.05% after 1 year of split announcement. Similarly the firms earn 9.39% and 11.87% respectively after 2 years and 3 years period of the announcement month. In the same way 1 year, 2 years and 3 years abnormal returns following reverse splits announcements months are -10.77%, -20.62% and -33.90% respectively.

Ikenberry et al. (1996) examine 2-for-1 stock splits of NYSE from 1975 to 1990. They take quite a large sample of 1275 splits and report the excess returns of 7.93% in the first year after the splits and 12.15% in the first three years following the splits. Further, they notice, there exist an inverse relationship between firm size and announcement returns. The advantage of increased information particularly accrues for small firms. In other word, the additional information is much more valuable for small firms than for large firms, because for larger firms usually the information is abundant. These findings support the findings of Grinblatt et al (1984), Lamoureux and Poon (1987). On the contrary Byun and Rozeff (2003) examine a vast sample covering 12,747 split events from 1927-1996 and report that stock splits is a value neutral events.

In support of the signaling hypothesis, Conroy et al. (1999) find that excess returns after stock splits are considerably higher when shareholders are surprised by a larger than expected split ratio. Also the financial analysts increase their earnings forecast for the firm. Woolridge and Chambers (1983) report that reverse splits are associated with negative abnormal returns around announcement and ex date of splits.

This paper is divided into two parts. The first parts of this paper contribute to the existing literature. It extends the international empirical evidence on stock splits to the Indian capital market, which is growing at a very fast pace among all the emerging markets across the globe. It also provides additional insight into the relative explanation power of the existing theories.

In the second part of the paper, we analyse the abnormal returns cross-sectionally in order to test the impact of other factors that may affect them. Our regression model includes variables like, whether stock is in the option category or not, the market capitalisation of the firm, book-to-market ratio, average trading volume, promoters' stock holdings and total institutional holdings. Here, our hypothesis is the stocks which are actively traded in option market are more information efficient. More specifically, we hypothesise that the announcement of a stock split conveys less new information to the market for a stock that is optioned than for one that is not optioned. Therefore, it is expected that abnormal returns associated with stock splits announcements are lower for optioned than similar non-optioned stocks.

This essay contributes to the existing literature in several ways. It broadens the international evidence on stock splits to the Indian stock market. It provides the explanatory power of existing theories supporting mainly signaling hypothesis also analyses the abnormal returns cross-sectionally in order to control for several other factors that may affect them.

The remainder of this paper is as follows. Section 2 briefly reviews the existing literature. We frame the hypothesis in section 3. We describe the data and methodology in section 4. We elaborate the abnormal return testing procedure in section 5. We examine the abnormal return in multivariate settings. We provide finding section 7 and conclusion in section 8 respectively.

2. Review of Literature

The stock splits literature can be split among three categories: the first category deals with the explanation, why managers may resorts to stock splits. Ross (1977), Brennan and Copeland (1988) worked on signaling aspect of stock splits. Leland and Pyle (1977) showed the information asymmetry rational for stock splits. Baker and Gallagher (1980), McNichols and Dravid (1990) shown the management rationale for stock splits.

The second category of research investigates and documents their empirical results regarding the announcement effects of stock splits. These groups of researchers have done the event studies and try to find the stock price reactions around the announcement date and the execution date. The prominent research in this category mainly contributed by Grinblatt et. al (1984) Woolridge (1983) etc.

The third category of papers deals with the impact on different variables such as variance of return, volume, liquidity, betas across the pre and post split periods. Previous research found that splits have significant liquidity effects Copeland (1979), Lakonishok and Lev (1987), Conroy et al. (1990). The evidence suggests that splits are associated with a post-split risk adjusted drift in prices Grinblatt et al. (1984), Ikenbery et al. (1996). Increased volatility Ohlson and Penman (1985) and increased market betas Lamoureux and Poon (1987).

Sufficient empirical studies in the United States show that stock splits are associated with abnormal return on both the announcement day and the execution day. Fama et al. (1969) were the first to analyse the stock splits. They found significant positive abnormal return in the month of the announcement of splits but slowly it vanishes. The cumulative abnormal returns come down to zero after the 30th month of the splits. They also documented that the months immediately after a splits reflect a market anticipation of substantial increase in dividend. Further when dividend paying firms announce stock splits it is also more likely to announce a dividend increase afterward.

Two general theories have been developed to explain the announcement affect: the liquidity hypothesis and signaling hypothesis. As per liquidity hypothesis, stock splits bring down the stock price to a lower trading range and attract the wealth constrained new and small investors. This in turn improves the liquidity of the stock. As per signaling hypothesis there exist an information asymmetry between the managers and the investors. The split announcement is inferred as favourable signals.

Grinblatt et al. (1984) studied stock splits for the period 1967 to 1976 and they found an excess return of 3.44% during the three days surrounding the stock splits. Koski (1998) observed the similar results 3.4% excess return as Grinblatt et al. (1984) surrounding the split announcement days. Moreover he found excess returns surrounding the execution days also.

Ikenbery et al. (1996) found long tern positive abnormal return following the splits. They find 7.93% of excess return in the first year and 12.15% in the three years subsequent to the splits. Desai and Jain (1997) point out at similar direction. They find 7.05% and 11.87% excess returns respectively after one and three years of the split events. More recently, Byun and Rozeff (2003) examine the long run consequence of splits from 1927 to 1996, covering 12,747 stock splits event. They find negligible abnormal returns and concluded that the long term "stock splits evidence against market evidence is neither pervasive nor compelling".

McNichols and Dravid (1990) find a significant relation between excess announcement returns and the one-year ahead earnings forecast error. This finding suggests that excess announcement returns can be explained by management's private information about future earnings. Conroy and Harris (1999) hypothesise split announcement period abnormal returns to subsequent changes in earning forecasts by the equity analysts. They find significant relation between forecast revision and the expect split ratio of the firm.

In similar other literature Grinblatt et al. (1984) find that split announcement effects are higher for non-dividend paying firms compare to dividend paying firms. Asquith et al. (1989) find have superior earnings performance in the year before the splits. Further this superior earning performance is not temporary rather permanent for the subsequent years following the splits. Similarly, Ohlson and Penman (1985) and Lakonishok and Lev (1987) the average earnings growth has increased after the year of splits.

Similar findings have been reported from the other countries. Wulff (2002) finds evidence from German Stock exchange. He shows excess returns for consecutive four days immediately after the split announcement day as well as split execution day. Bechmann and Raaballe (2004) report a significantly positive split announcement return of 2.5% in Danish Stock exchange. Elfakhani and Lung (2003) find similar positive abnormal returns in the Canadian market. Among developing market Wu and Chang (1997) studied Hong Kong market and find a great amount of abnormal return associated with split announcement. He finds more than 18% abnormal return for three days surrounding the split announcement.

In India Gupta and Gupta (2007) studied split event from 1999 to 2004 taking 60 stocks and find a mixed announcement effect. They find significant positive abnormal returns on the very next day of announcement but the following days show significant negative abnormal returns. Specifically, they find significant positive abnormal returns on the execution day and the next to next day. In another paper Mishra (2007) studied the splits by taking a larger sample of 180 stock splits from 1999 to 2005 and finds abnormal return of around 2.45% surrounding the three days of split execution and the result is statistically significant. This finding suggests that stock splits on Bombay Stock Exchange exhibit similar ex-days price reaction as it is found in the German stock exchange Wulff (2002).

Many empirical researches show that the information efficiency increases for stocks which have option associated with them i.e. the stock is eligible in option trading. It is evident that the stocks which are eligible in option trading draw more investors and analysts in that particular counter. It is said that options provide additional cost effective ways in which market participants can trade on new information (both positive and negative).

Jennings and Starks (1986) find that prices of optioned stocks adjust more rapidly to earnings announcement than prices of non-optioned stocks. Damodaran and Lim (1991) find that after options introduction, there is significant increase in the number of analyst's coverage for the stock and also there is an increase in the number of articles that mentioned the stock in The Wall Street Journal.

Many researchers show that trading in options increases the informational efficiency for the underlying stocks. Options provide a means by which more information both positive and negative can come to the market and stock prices tend to adjust more quickly to new information. Ross (1976) suggests that stock options improves the informational efficiency and subsequently reduces the volatility of the underlying stock. Easley et al. (1998) analyse the linkage of price, volume and information between stock and option market. They find the change in option volume have significant predictive power for stock price movements in case of both positive news and negative news. Chakravarty et al. (2004) find direct evidence of significant price discovery in options and it plays an important informational role.

In the light of the above study, Chern et al. (2006) hypothesised that the split announcement itself reveals less new information to the market for the optioned socks. They concluded that the evidence is consistent with the hypothesis that the announcement effects of stock splits of NYSE/ Amex stocks that are optioned are lower and are completed more quickly than non-optioned stocks.

3. The Hypotheses:

We calculate average abnormal return for all observations over the full test period t = -10 to t = +10. We test the null hypothesis that the average abnormal return at time t is zero. Our alternate hypothesis is that the average abnormal return at time t is greater than zero. Based on the earlier empirical research, we expect a positive price reaction on the day of announcement of stock splits.

Further, we sum average abnormal return over the full test period t = -10 to t = +10 and three sub period: pre event period t = -10 to t = -1, event period t = 0 to t = +1 and post event period t = +2 to t = +10. We test the event period (t = 0 to t = +1) reactions as the financial press normally reports the stock splits on the day after the firms release this information to the market. We test the cumulative abnormal return (CARs) against the null hypothesis that CAR =0.

4. Data & Methodology:

Before 1999, it was mandatory to have the minimum par value for the stocks in India. This law inhibited many companies from splitting their stocks. Since March 1999 the Security and Exchange Board of India (SEBI) the capital market regulating body in India allowed companies to set the face value of their shares, as long as it is not fractional. This move enabled many companies to split their stocks. Keeping this in mind the study covers a time periods of 9 years from the year 2000 to 2008. The primary source of data for this study is collected from Centre for Monitoring Indian Economy (CMIE) Prowess database

During this period we found a total of 664 stock split events and 19 reverse stock split events. Again, out of the total sample a large number of stocks are eliminated because the announcement date of the split events was not clearly mentioned. To ensure the validity of estimates, we put some additional criteria for selecting the stock splits sample.

(i) The split shares must be ordinary common shares.

(ii) There should be no contamination effect on the split announcement within the 5 days of the stock splits announcement date or ex-dates. This means the split stocks are free from any other announcement effects such as cash dividends, bonus issues, right issues etc.

(iii) The splitting firms must have annually and quarterly stock holding pattern information available from the Prowess database.

Finally, we arrive at a sample of 302 announcement events and 295 ex-date events of stock splits. The data in table 1 shows that the largest number of stock splits occurred during the year 2005 and the uptrend continued in the following years as well. We check the industry-wise splitting events and find the stock splits are across the board and not specific to a small set of industries.

The daily closing price data is expressed in Indian Rupees. For stock returns the data were obtained from price series and the returns are computed as:

[R.sub.t] = log ([P.sub.t]/[P.sub.t-1]),

Where [R.sub.t] and [P.sub.t] are the return and the daily closing price, respectively for period t and [P.sub.t-1] is the prior day's closing price.

We calculate the daily market return in the similar fashion using the Bombay Stock Exchange's (BSE) sensitivity index "Sensex". We have referred the National Stock Exchange of India website to check whether or not a particular stock in the sample is eligible for option trading. The reason behind this is that, the majority of the derivatives trading in India is done through the National Stock Exchange platform.

5. Statistical test:

In this study we use the most commonly used event study methodology of Brown and Warner (1980). This is a simple methodology based on the market model. This method is well specified and relatively powerful under a wide variety of conditions.

There are two most important dates related to stock splits; the announcement date and the execution date. The announcement date is the date of the board meeting and the announcement of the split event for each sample company. Here it is marked as AD 0. The other event date is the execution date when, the stock splits actually takes place in the market. On that day the stocks are traded on split-adjusted basis. Here, it is marked as ED 0. In this paper we test the abnormal returns during the period 10 days before and 10 days after the AD 0 and ED 0 dates . These 21 days around split announcement and split execution dates would be the window for our study.

As mentioned earlier, we use Bombay Stock Exchange sensitivity index "Sensex" as a proxy for computing market return.

[R.sub.mt] = Log([I.sub.t]/[I.sub.t-1])

We employ the market model to compute the abnormal returns, which are determined by estimating the following Ordinary Least Square (OLS) equation.

[R.sub.it] = [[alpha].sub.i] + [[beta].sub.i] [R.sub.mt] + [[epsilon].sub.it]

Where [R.sub.it] = the daily return of security i at day t.

[R.sub.mt] = the daily return of Sensex at day t.

[[alpha].sub.i] and [[beta].sub.i] = OLS intercept and slope coefficients respectively.

[[epsilon].sub.it] = the regression error term for security i at day t.

[[alpha].sub.i] and [[beta].sub.i] are derived from the market model over -210 days to -11 relative to the announcement dates.

Then we compute the 'abnormal return' for each of the sample company for the window period. The abnormal return is defined as the actual return minus the expected return using market model.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

Where, [AR.sub.it] is the abnormal return for firm i at day t.

Next, we calculate the cross-sectional average abnormal returns for all firms in order to eliminate the effect of any one or group of securities on the abnormal returns. The ARs of individual companies are averaged for each day surrounding the event day (-10 to +10 days) using the following formula.

[AAR.sub.it] = [N.summation over (t=1)] [AR.sub.it]/N

Where N = Number of securities.

To understand the cumulative effect of AARs on the days surrounding the event we have to obtain the Cumulative Average Abnormal Return (CAAR) for event days [t.sub.1] (-10 days) through [t.sub.2] (+10 days) by summing the average abnormal returns for these days.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

Calculation test of significance:

Next we calculate the standard deviation of abnormal returns for the period -210 days to -11 days. Then, the Standardized Abnormal Return (SAR) for each company is obtained by dividing abnormal returns of the event period (-10 days to +10 days) by the obtained standard deviation.

The Z-statistic for the average abnormal return for N companies will be determined by

[Z.sub.t] = [N.summation over (i=1)] [SAR.sub.it]/[square root of N]

For testing the cumulative excess returns for N securities over T days (from [t.sub.1] through [t.sub.2]) the Z-statistic is

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

To test the statistical significance of the abnormal returns and the cumulative abnormal returns is that, whether the average abnormal returns of the splitting firms sample for day i are significantly different from zero.

6. Examination of abnormal returns in a multivariate setting:

We analyse the abnormal returns surrounding split announcement days and split execution days with a cross-sectional regression analysis. The cross-sectional regression model has abnormal returns surrounding the days of split announcement and execution as the dependent variable. The main objective is to determine whether the split event effects are different for optioned stocks than non-optioned stocks by including, an option dummy variable with a value of 1 for the optioned stocks; 0 otherwise.

The secondary objective of the regression model is to control for the effects of different variables such as market capitalisation, book-to-market ratio, average trading volume and promoters' holdings.

We calculate the average daily market capitalisation of the firm as a measure of the size. As our measure of size, we calculated as the natural log of average market capitalization in the event period (-90 days to -30 days). In case of size of the firm, we hypothesise that smaller the size, more the information asymmetry, hence higher would be the abnormal returns around the split announcement.

The second variable is the book-to-market ratio. Lakonishok et al. (1994) examine glamour and value stocks based on the book-to-market ratio. They have shown on average the low book-to market (glamour) stocks have an average annual return of 9.3% and the high book-to-market (value) stocks have an average annual return of 19.8 percent. They have taken 22 long years in their study period. This shows the firms having high book-to-market (value stock) are more likely to be undervalued.

Ikenberry et.al. (1996) examine the market reaction of split announcement by dividing the sample into book-to-market quintile. Contrary to general explanation they find the announcement reaction to glamour stock is 3.96 percent (t= 13.18) which is much larger than the announcement return for value stocks of 0.91 percent (t=1.57). This empirical evidence suggests that the market reaction to a split announcement is more positive for glamour stocks at least in the short run.

In the light of the above empirical evidence, there is a significant difference of average annual return and announcement return between high book-to-market (value stock) and low book-to-market (glamour stock). Thus, we consider book-to-market ratio is another proxy for information asymmetry. Here, we hypothesized that lower the book-to-market ratio, more the undervaluation of the firm and higher the abnormal returns surround the stock splits announcement and executions. The magnitude of the split announcement reaction should be negatively correlated with the size and book-to-market ratio Ikenberry et al. (1996).

The next variable is promoters' holding. In India, most of business firms are promoted by some families. Large portion of shares are held by those families as promoters' holding. The announcement of the stock splits is known by the promoters beforehand. The higher the percentage of promoters' holding the lower would be the percentage of free floating shares available for regular trading. Hence, it is hypothesised that higher the promoters holding, the less is information asymmetry and less abnormal returns around the stock splits.

The fourth variable in the cross sectional regression is the trading volume. Generally the heavily traded stocks are more liquid and more information gets impounded in the price of the stock. This in turn, reduces the price fluctuation when management announced a stock split and lower would be the abnormal returns. There should be a negative relation between stock's trading volume and the abnormal returns. As a measure of trading volume, we calculate the average trading volume around the event period (-90 to -30) days Natural log is taken to reduce any scale difference with other variables.

The specification of regression equation including the dummy variable with other explanatory variables for optioned and non-optioned stocks is as follows:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

Where

M_[CAP.sub.n] = Average market capitalisation log(Market Capitalisation) of stock in the event days (-90, -30).

BV/[MV.sub.n] = Book-to-market ratio for the stocks in the quarter of the splits.

[PH.sub.n] = Promoters' holding proportion in the split company during the quarter of the splits. Here, we have taken the percentage of non-institutional promoters holding (percentage of shares held by families and friends)

[D_OPT.sub.n] = Dummy variable with a value of 1 if the stock was optioned at the time of split announcement; 0 otherwise.

Avg [Vol.sub.n] = Average trading volume log(Average Volume) of stock in the pre-announcement period (-90, -30).

7. Findings:

7.1 Abnormal Returns around the Announcement Day

The results of the tests are summarized in Table 2. It indicates that the stock returns of firms with stock splits were positively affected by the announcement of the stock splits. The average abnormal returns of t = 0 is 0.80% and statistically distinguishable from zero with a t-statistic of 4.56. Also, the average abnormal return of t = -1 is 0.63% with a t-statistic of 3.45. The two days cumulative average abnormal return is 1.43%.

The abnormal return is positive for each day of t = -10 to t = 0, except t = -8. Moreover, these positive abnormal returns are statistically significant for days from t = -10 to t = 0 except t = -8, 3, -2. This suggests that even before announcement of stock splits, there are some positive reactions in the stock market. However, from t = +3 to t = +10 each abnormal returns are negative except t = 7 and statistically significant for t = +4, +5, +6 and +8.

The cumulative average abnormal return at the starting event date of t = -10 to t = +10 is shown in Table 2. The cumulative average abnormal return from t = -10 to t = 0 is 6 percent and has increased another 0.50 percent following two days of the announcement date. These abnormal returns in case of small event window around the split announcement ranges in between 2 percent to 4 percent is similar as the same was found in US market Ikenberry et al. (1996) and Grinblatt et al. (1984). After t = +2, the cumulative average abnormal return has gone down consistently and reached to 3.36% on day t = +10. This particular result shows that, the most of the positive abnormal return are generated before the announcement and there is a negative post announcement drift. Hence, we can conclude that there is a leakage of announcement information before the actual announcement of the event and fund managers can make significant abnormal profit be gathering the announcement information.

7.2 Abnormal Returns around the Execution Day

Table 3 report the abnormal returns in the event window [-10, +10] around the ex-days of stock splits in the Indian stock market. On the day on execution there is negative abnormal returns of 0.9% which is highly statistically significant. On the contrary significant positive abnormal returns are observed on the previous day and the following day of execution of splits. There are positive abnormal returns during the window of -7 to -1, except day -5 and day -2. The cumulative average abnormal returns are 1.27 percent. The abnormal return turns to be negative from the t= +3 onward after the execution. In the interval from day t= +3 to day t= +10, the cumulative negative abnormal returns is 7.6 percent and throughout the period for each days negative abnormal returns were statistically significant except t= +3.

Like announcement effect of stock splits, the ex-day effect is also pronounced in India as compared to the US. The effects of the positive abnormal returns is observed before the split execution but quickly get revered after the execution of splits. The market inefficiency could be a plausible explanation of the ex-days effects and might be an area of future research.

7.3 Evidence from Cross-Sectional Regression:

The dependent variable in the cross-sectional regression model is the abnormal returns and cumulative average abnormal returns surrounding the days of split announcement and execution. The explanatory variables are the logarithms of average size of the splitting firms, book-to market ratio during the split quarter, Promoters' holding during the split quarter, the logarithms of average trading volume. A dummy variable representing whether the stock is optioned or not.

Regression results are shown in table 4 and table 5. We do not find any evidence of significant difference between the announcement effects of optioned stocks and non-optioned stocks. This results does not consistent with our hypothesis that the announcement of stock splits conveys less new information to the market for optioned than non-optioned stocks.

In case of other variables, we find that for announcement day -1, the coefficient of average volume traded is negative and significant. It implies that heavily traded stocks are more liquid and more information get impounded in the price of the stocks therefore lower would be the abnormal returns. The coefficients of other variables are not significant for announcement day 0, +1 and -1.

Furthermore, we take cumulative average abnormal returns (-5 days to +5 days) surrounding the announcement days, as dependent variable and regress the above mentioned independent variables on it. Table 4 confirms the results as there is negative and highly significant relation for the level of market capitalization and book-to-market ratio. This results suggest that smaller the size, more the information asymmetry and higher would be the abnormal returns around the split announcement. Also, lower the book-to-market ratio, more the undervaluation of the firm and higher the abnormal returns surrounding the announcement. The coefficient of promoters holding is negative and significant only at the 10 percent level.

For execution day -1, the coefficient of book-to-market value is positive and significant, similarly, for execution day 0, the coefficient of volume traded is positive and significant. Although, the coefficient value is significant but these are contrary to the previous research that there are negative relationship between the book-to-market value and volume traded with the abnormal returns. For execution day +1, the coefficient of market capitalization is negative and significant.

Similarly, we take cumulative average abnormal returns (-5 days to +5 days) surrounding the execution days, as dependent variable and regress the above mentioned independent variables on them. Table 5 confirms the results as there is positive and highly significant relation with the average volume traded. The coefficient of market capitalization is negative and significant only at the 10 percent level.

8. Conclusion:

Apparently stocks splits look to be a purely cosmetic event but the existence of ample empirical evidence from US and European stock market show that stock splits are associated with abnormal returns on both the announcement and the execution day. There have been very few studies documenting the stock return behavior around stock splits in the Indian context. The study uses a large sample of stock splits data from 2000 to 2008 drawn from CMIE Prowess.

We find a significant positive abnormal return on and before the announcement and execution days of stock splits respectively. But those positive abnormal returns did not sustain and get reversed in few days after the split announcement and effective split days. Hence, it can be concluded that there is no clear evidence about short term positive wealth effect associated with stock splits in Indian market.

The cumulative average abnormal return from t = -10 to t = +2 is 6.5 percent and this cumulative average abnormal return has gone down consistently and reached to 3.36% on day t = +10. This particular result indicating that, the most of the positive abnormal return are generated before the announcement and having a positive pre announcement drift and there is a negative post announcement drift. It is apparently evident that there is significant price rise before the split announcement which could be interpreted as a leakage of announcement information and might lead to a strong evidence of insider trading before the actual announcement of the event. This may further call for an investigation of market efficiency of Indian stock market with respect the split event.

In the second part of the study we try to see whether and how option affects the effect of stock splits announcements on stock returns. We also measure the announcement effects of stock splits on variety of factors like book-to-market, average daily market capitalisation, holding pattern of promoters and average daily volume of trading. We find that there is no evidence of a significant difference between the announcement effect of optioned stocks and non-optioned stocks. In case of other variables, we find that the abnormal returns are mostly attributed and statistically significant for average daily trading volume. The other variables like book-to-market and average daily market capitalization is showing some significant results on some days surrounding the split announcement and execution.

In our study we did not categories the sample into dividend paying and non-dividend paying stocks. It may be interesting to investigate whether any differential price effect would appear under these two different sets of firms. This will lead us to further insight about the effect of stock splits.

Reference:

Angel, J. (1997) 'Tick Size, Share Price and Stock Splits', Journal of Finance, Vol.52, pp.655-681.

Asquith, P. Healy, P. and Palepu, K. (1989) 'Earnings and Stock Splits', Accounting Review, Vol. 64, pp.387-403.

Baker, H. K. and Gallagher, P. L. (1980) 'Management's View of Stock Splits'. Financial Management, Vol. 9, No. 2, pp.73-77.

Basu, S. (1977) 'Investment Performance of Common Stocks in Relation to Their Price Earning Ratios: A Test of the Efficient Market Hypothesis', Journal of Finance, Vol. 32, pp.663-682.

Bechmann, K. and Raaballe. J. (2004) 'The difference between stock splits and stock dividends: Evidence from Denmark', Working Paper Copenhagen Business School.

Black, Fisher. (1986) 'Noise', Journal of Finance, Vol. 40, pp.529-543.

Brennan, M and T, Copeland. (1988) 'Beta changes around stock splits: A Note'. Journal of Finance, Vol. 43, pp.1009-1013.

Brown, S. and Warner, J. (1980) 'Measuring Security Price Performance', Journal of Financial Economics, Vol.8, pp.205-258.

Byun, J. and Rozeff, M. (2003) 'Long-run Performance after Stock Splits: 1927 to 1996', Journal of Finance, Vol. 58, pp.1063-1085.

Chan, I. Hamao, Y. and Lakonishok, J. (1991) 'Fundamentals and Stocks Returns in Japan', Journal of Finance, Vol. 46, pp.1739-1764.

Chakravarty, S. Gulen. H and Mayhew, S. (2004) 'Informed Trading in Stock and Option Markets', Journal of Finance, Vol. 58, No. 3, pp.1235-1257.

Chern, KY. Tandon, K. Yu, S. and Webb, G. (2008) 'The Information Content of Stock Splits Announcements: Do Option Matter?', Journal of Banking and Finance, Vol. 32, No. 6, pp.930946.

Conrad, J. (1989) 'The Price Effect of Option Introduction', Journal of Finance, Vol. 44, No. 2, pp.487-498.

Conroy, R, M. Harris, R. and Benet, B. (1990) 'The effects of Stock Splits on Bid-Ask Spreads', Journal of Finance, Vol.45, pp.1285-1295.

Copeland, T. (1979) 'Liquidity Changes Following Stock Splits', Journal of Finance, Vol. 34 pp.115-141.

Damodaran, A. and Lim, J. (1991) 'The Effect of Options Listing on the Underlying Stocks Return Processes', Journal of Banking and Finance, Vol. 15, No. 3, pp.647-664.

De Bondt, and Thaler, R. (1985) 'Does the Stock Market Overreacts', Journal of Finance, Vol. 40, pp.793-805.

Desai A and Jain P C. "Long-run Common Stock Returns Following Stock Splits and Reverse Splits" Journal of Business. 1997, Vol. 46 (421-431).

Desai, A Nimalendran, M. and Venkataraman, S. (1998) 'Changes in Trading Activity Following Stock Splits and their Effect on Volatility and the Adverse Information Component of Bid-Ask Spread', Journal of Financial Research, Vol. 21, pp.159-184.

Dhar, R. and Goetzmann, W, N. (2005) 'The Impact of Clientele Changes: Evidence From Stock Splits', Yale ICF Working Paper No. 03-14. EFA 2005 Moscow Meetings Paper, AFA 2005 Philadelphia Meetings.

Dravid, A, R. (1987) 'A Note on the Behavour of Stock Returns around Ex-dates of Stock Distributions', Journal of Finance, Vol. 42, pp.163-168.

Dubofsky, D, A. (1991) 'Volatility Increases subsequent to NYSE and AMEX Stock Splits', Journal of Finance, Vol. 46, pp.421-431.

Easley, D. O'Hara, M. and Srinivas, P. (1998) 'Option Volume and Stock Prices: Evidence on Where Informed Traders Trade', Journal of Finance, Vol. 53, No. 2, pp.431-465.

Elfakhani, S. and Trevor, L. (2003) 'The Effect of Split Announcements on Canadian Stocks'. Global Finance Journal, Vol.14, pp.197-216.

Fama, E. Fischer, L. Jensen, M. and Roll, R. (1969) 'The Adjustment of Stock Prices to New Information', International Economic Review, Vol. 10, pp. 1-22.

Fama, E. and French, K. (1992) ' The Cross-section of Expected Stock Returns', Journal of Finance, Vol 46, pp.427-466.

Figlewski, S. and Webb, G. (1993) 'Options, Shot Sales and Market Completeness', Journal of Finance, Vol. 48, No. 2, pp.761-777.

Grinblatt, M, S. Masulis, R, W, and Titnam, S. (1984) 'The Valuation Effects of Stock Splits and Stock Dividends', Journal of Financial Economics, Vol.13, No. 4, pp.461-490.

Gupta, A. and Gupta, O, P. (2007) 'Market Reaction to Stock Market Splits: Evidence from India'. The Icfai Journal of Applied Finance, Vol. 13, No. 1, pp.5-22.

Ikenbery, David L. Rankinge, G. Stice, E, K. (1996) 'What Do Stock Splits Really Signal?', Journal of Financial and Quantitative Analysis, Vol. 3, pp.357-375.

Jaffe, J. Keim, B. and Westerfield, R. (1989) 'Earning Yield, Market Values and Stock Returns'. Journal of Finance, Vol.44, pp.135-148.

Jennings, R. and Starks, L. (1986) 'Earnings Announcements, Stock Price Adjustment, and the Existence of Options Markets', Journal of Finance, Vol. 41, No. 1, pp.107-125.

Jog, V. and Zhu, P. (2004) 'Thirty Years of Stock Splits, Reverse Stock Splits, and Stock Dividends in Canada', SSRN Working Paper Series.

John, K. Koticha, A. and Subrahmanyam, M. (1993) 'The micro-structure of options markets: Informed trading, liquidity, volatility and efficiency' SSRN Working Paper.

Joshipura, M. (2008) 'Price and Liquidity Effects of Stock Splits: An Empirical Evidence from Indian Stock Market', National Stock Exchange of India Publication Series.

Klein, L, S. and Peterson, D, R. (1989) 'Earnings Forecasts Revisions Associated with Stock Splits Announcements', Journal of Financial Research, Vol.12, pp.319-328.

Koski, J, L. (1998) 'Measurement Effects and the Variance of Returns after Stock Splits and Stock Dividends', The Review of Financial Studies, Vol. 11, pp.143-162.

Kumar, R, Sarin, A. and Shastri, K. (1998) 'The Impact of Options Trading on the Market Quality of the Underlying Security: An Empirical Analysis', Journal of Finance, Vol. 83, No. 2 pp.717-732.

Lakonishok, J. Shleifer, A. and Vishny, R. (1994) 'Contrarian Investment, Extrapolation and Risk', Journal of Finance, Vol. 49, No. 5, pp.1541-1578.

Lakonishok, J. and Lev, B. (1 987) 'Stock Splits and Stock Dividend: Why, Who and When'. Journal of Finance, Vol.42, pp.913-932.

Lamoureux, C. and Poon, P. (1987) 'The Market Reaction to Stock Splits', Journal of Finance, Vol.42, pp.1347-1370.

Leland, H, E. and Pyle, D, H. (1977) 'Informational Asymmetries, Financial Structure and Financial Intermediation'. Journal of Finance, Vol. 32, No. 2, pp.371-387.

McNichols, M. and Dravid, A. (1990) 'Stock Dividend, Stock Splits and Signaling', Journal of Finance, Vol. 45, No. 3 pp.857-879.

Mishra, A, K. (2007) 'The Market Reaction to Stock Splits--Evidence from India', International Journal of Theoretical and Applied Finance, Vol.10, pp.251-271.

Ohlson, J, A. and Penman, S, H. (1985) 'Volatility Increases Subsequent to Stock Splits: An Empirical Aberration', Journal of Financial Economics, Vol. 14, pp.251-266.

Rosenberg, B. Reid, K. and Lanstein, R. (1984) 'Persuasive Evidence of Market Efficiency', Journal of Portfolio Management, Vol. 11, pp.9-17.

Ross, S, A. (1976) 'Options and Efficiency', The Quarterly Journal of Economics, Vol. 90, pp.75-89.

Ross, S, A. (1977) 'The Determination of Financial Structure: The Incentive Signaling Approach', Bell Journal of Economics, Vol. 8, No. 1, pp.23-40.

Schultz, P. (2000) ' Stock Splits, Tick Size and Sponsorship' Journal of Finance, Vol. 55, pp.429-450.

Sharpe, F, W. (1964) 'Capital Asset Pricing: A theory of Market Equilibrium Under Conditions of Risk', Journal of Finance, Vol. 19, pp.425-442.

Sheikh, A. (1989) 'Stock Splits, Volatility Increases, and Implied Volatilities', Journal of Finance, Vol. 44, pp.1361-1372.

Woolridge J R and D P Chambers. "Reverse Splits and Shareholder Wealth". Financial Management Autumn. 1983, (5-15).

Wu, L and Chan B. "On the Existence of an Optimal Stock Price; Evidence from Stock Splits and Reverse Stock Splits in Hong Kong" Internaytional Journal of Business, 1997, Vol. 2, 45-67.

Wulff, C. (2002) 'The Marker Reaction to Stock Splits: Evidence from Germany', Schmalenbach Business Review, Vol. 54, pp.280-297.

Dr. Pradip Banerjee

Assistant Professor

Institute of Management Technology

Raj Nagar 

Ghaziabad--201 001

U.P.

Phone: [phone] (Extn. 299)

Email: [e-mail]

Dr. Prithviraj S. Banerjee

Assistant Professor, Globsyn Business School

XI-11&12, Block EP, Sector-V

Salt Lake Electronics Complex, Kolkata-91

Email: [e-mail]

Table 1: Description of the Sample Stock Splits.

Year      All       Number     Percentage   Number of
        Splits    of Splits    of Sample     Optioned
        Stocks      sample                    Stocks

2000      39          17          5.6           7
2001      33          7           2.3           2
2002      38          7           2.3           2
2003      32          10          3.3           5
2004      39          23          7.6           8
2005      159         82          27.2          18
2006      107         56          18.5          13
2007      117         49          16.2          11
2008      100         51          16.9          7

Total     664        302         100.0          73

Table 2; Panel A: Abnormal Returns Associated with the
Announcement of Stock Splits.

This table presents the daily average abnormal returns (AARs) and
the corresponding t-statistic for the AAR over the period
[t.sub.-10] to [t.sub.+10] for the full sample of 302 stock
splits over the period of 2000 to 2008. Market model is used to
compute the abnormal returns: [AR.sub.jt] = [R.sub.jt]--
([[alpha].sub.j] + [[beta].sub.j] [R.sub.mt]). The table also
presents the cumulative average abnormal returns (CAAR). The t-
value indicates whether the ARs and CAARs are significantly
greater than zero.

Trading    Daily AAR     t-statistic      CAAR       t-statistic
Days

-10         0.334097    2.385158 ***    0.334097    2.385156 ***
-9          0.529435    2.661374 ***    0.863531    3.568434 ***
-8          -0.00074      0.425927      0.862795    3.159523 ***
-7          0.511164    3.178794 ***    1.373959    4.325623 ***
-6          0.73292     4.048576 ***    2.106879    5.679531 ***
-5          0.975159    6.131698 ***    3.082038    7.687931 ***
-4          0.873024    4.947369 ***    3.955062    8.987569 ***
-3          0.338955       1.34217      4.294018    8.881629 ***
-2          0.267041      1.223323      4.561059    8.781455 ***
-1          0.631868    3.450869 ***    5.192926    9.422079 ***
0           0.799792    4.563197 ***    5.992719    10.35945 ***
1           0.446556     1.921392 *     6.439274    10.47308 ***
2           0.106593       -0.0114      6.545867    10.05905 ***
3           -0.11046      -1.52389      6.435411    9.285865 ***
4           -1.18972     -2.31859 **    5.245695     8.37234 ***
5           -0.68774    -4.18391 ***    4.557958    7.060506 ***
6           -0.39887    -2.86371 ***    4.159086    6.155146 ***
7           0.080234      0.454657      4.239319     6.08889 ***
8           -0.47314    -2.84286 ***    3.766183    5.274294 ***
9           -0.28085      -1.60142       3.48533    4.782658 ***
10          -0.12385      -0.66036      3.361476    4.523294 ***

 *, ** and *** indicate significant at 10%, 5% and 1% respectively.

Table 2; Panel B: CAARs and t-values of CAAR

Duration                             CAAR       t-statistic

Pre-event (t = -10 to -1)            5.1929     8.330818 ***

On-event (t = 0 to +1)               1.2462     4.585299 ***

Post-event (t = +2 to +10)           -3.0778    -5.18383 ***

Full test period (t = -10 to +10)    3.3615     4.523297 ***

*, ** and *** indicate significant
at 10%, 5% and 1% respectively.

Table 3 Panel A: Abnormal Returns Associated with the
Execution of Stock Splits.

This table presents the daily average abnormal returns (AARs) and
the corresponding t-statistic for the AAR over the period
[t.sub.-10] to [t.sub.+10] for the full sample of 295 stock
splits over the period of 2000 to 2008. Market model is used to
compute the abnormal returns: [AR.sub.jt] = [R.sub.jt]--
([[alpha.sub.j] + [[beta].sub.j] [R.sub.mt]). The table also
presents the cumulative average abnormal returns (CAAR).
The t-value indicate whether the ARs and CAARs are significantly
greater than zero.

Trading    Daily AAR     t-statistic       CAR       t-statistic
Days

-10         0.094492      0.599407      0.094492      0.599405
-9          0.059229      0.857223      0.153721       1.02999
-8          -0.14901      -0.25469      0.004712       0.69394
-7          0.286406     2.014145 **    0.291118      1.608039
-6          0.341951     2.012063 **    0.633069      2.338093
-5          -0.20792      -0.57232      0.425154     1.900728 *
-4          0.127755      1.018026      0.552909     2.144509 **
-3           0.0171       0.184294      0.570009     2.071162 **
-2          -0.05581      0.432249      0.514197     2.096793 **
-1          0.756445    5.282251 ***    1.270643    3.659582 ***
0           -0.90209    14.69502 ***    0.368552    7.919975 ***
1           0.69799     4.015872 ***    1.066543    8.742078 ***
2           0.266975      0.732528      1.333518    8.602283 ***
3           -0.2331       -1.75508      1.100414    7.820303 ***
4           -0.42593    -3.26473 ***    0.674488    6.712182 ***
5           -0.29508    -3.13881 ***    0.379404    5.714344 ***
6           -1.53403    -9.25818 ***    -1.15462    3.298295 ***
7           -1.86633    -9.73568 ***    -3.02096       0.91065
8           -1.66426    -9.43615 ***    -4.68521      -1.27843
9           -1.0337     -5.06686 ***    -5.71891    -2.37904 ***
10          -0.5426     -2.45514 ***    -6.26151    -2.85746 ***

*, ** and *** indicate significant at 10%, 5% and 1% respectively

Table 3; Panel B: CAARs and t-values of CAAR

Duration                              CAR          t-statistic

Pre-event (t = -10 to -1)             1.270643     3.65958 ***
On-event (t = 0 to +1)                -0.2041      13.23058 ***
Post-event (t = +2 to +10)            -7.32805     -14.4593 ***
Full test period (t = -10 to +10)     -6.26151     -2.85746 ***

 *, ** and *** indicate significant at 10%, 5% and 1%
respectively.

Table 4: Cross-Sectional Regression with announcement of stock
splits

These table summaries results of regression on abnormal returns
on days -1, 0 and +1 around the announcement of stock splits.
The daily abnormal returns on day t is the difference between the
actual returns and the estimated returns from the market model.
The coefficients of market model are estimated from pre-split
period t = -210 to -11 days. The free float value weighted index
Sensex (Bombay Stock Exchange) is taken as the proxy for the
market return index. The variables included in the model are
market capitalisation, book-to-market ratio, Promoters' holding
and average trading volume. A dummy variable representing whether
the stock is optioned or not. The regression coefficient are
shown and p-values are in parentheses.

Dependent Variable: Abnormal Return

                      Announcement Day -1     Announcement Day 0

Intercept               2.2019 (-0.0284)       -0.7835 (-0.433)
Market Cap              -0.2944 (-0.768)       0.8137 (-0.416)
Volume                -2.1079 (0.0358) **      0.7338 (-0.4636)
B/M                     0.2233 (-0.823)        0.5608 (-0.575)
Promoters (%)           0.2372 (-0.812)        0.6736 (-0.501)
Option                  -0.1283 (-0.897)       -0.6428 (-0.52)
R Square                    0.02445                 0.0076
Standard Error              0.034975               0.043241
F-statistic                 1.438799               0.441668
Prob (F-statistic)          0.210271               0.819194

                      Announcement Day +1      Announcement Day
                                                  (-5 to +5)

Intercept               0.1402 (-0.888)        3.1759 (-0.0016)
Market Cap              0.6308 (-0.528)      -2.8414 (-0.0048) ***
Volume                  0.1157 (-0.907)        -0.3223 (-0.7474)
B/M                     0.7873 (-0.4317)     -8.5012 (-0.000) ***
Promoters (%)           -0.7254 (-0.468)        -1.8186 (-0.07)
Option                  -0.5627 (-0.574)       0.5216 (-0.6022)
R Square                     0.0050                 0.2281
Standard Error              0.046662               14.80122
F-statistic                 0.291644               16.96311
Prob (F-statistic)           0.9174                1.08E-14

 *, ** and *** indicate significant at 10%, 5% and 1% respectively.

Table 5: Cross-Sectional Regression with execution of stock splits

These table summaries results of regression on abnormal returns
on days -1, 0 and +1 around the execution days of stock splits.
The variables included in the model are market capitalisation,
book-to-market ratio, Promoters' holding and average trading
volume. A dummy variable representing whether the stock is
optioned or not. The regression coefficient are shown and
p-values are in parentheses.

Dependent Variable: Abnormal Return

                             Event                  Event
                             Day -1                 Day 0

Intercept               0.2809 (-0.7789)      -1.5701 (-0.1175)
Market Cap             -0.0402 (-0.9679)      -0.8600 (-0.3904)
Volume                  0.6647 (-0.5067)      3.186 (-0.001) ***
B/M                   2.0018 (-0.0462) **      0.9544 (-0.3407)
Promoters (%)           -0.5853 (-0.558)       -0.9275 (-0.354)
Option                 -1.4865 (-0.1382)      -1.5888 (-0.1132)
R Square                    0.028072               0.058618
Standard Error              0.037651               0.170305
F-statistic                 1.61167                3.474557
Prob (F-statistic)          0.156971               0.004609

                             Event                Event Day
                             Day +1               (-5 to +5)

Intercept               0.2296 (-0.8185)      -0.6526 (-0.5145)
Market Cap            -2.2142 (-0.027) **     -1.6420 (-0.1017)
Volume                  0.6956 (-0.4872)     2.6428 (-0.0086) ***
B/M                     1.0428 (-0.2979)       0.8742 (-0.3827)
Promoters (%)           1.6026 (-0.1101)      -1.0767 (-0.2825)
Option                  1.4899 (-0.1373)      -0.2885 (-0.7731)
R Square                    0.026011               0.050195
Standard Error              0.042783               0.213097
F-statistic                 1.490161               2.948874
Prob (F-statistic)          0.193053               0.013018

*, ** and *** indicate significant at 10%, 5% and 1% respectively.