Abstract. Keywords. Introduction

Transcription

1 Asia-Pacific Finance and Accounting Review Vol. 1, No. 3, April June 2013 pp , ISSN: Abstract Keywords Introduction Stock market efficiency is one the controversial topics of finance. Indian stock market has witnessed a series of reforms the last two decades in order to ensure operational efficiency, allocation efficiency and pricing efficiency. In case of operational efficiency and allocation efficiency are concerned, the regulatory authorities are vested with the responsibility of ensuring efficiency whereas in pricing efficiency, the investors are expected to earn risk adjusted return from their investment wherein M. S. Ramaratnam, Asst Professor (Senior Grade), Faculty of Management Studies, Sri Chandrasekharendra Saraswathi Viswa Maha Vidyalaya University, Enathur, Kanchipuram, Tamil Nadu. hellomsraman@gmail.com R. Jayaraman, Asst Professor, Faculty of Management Studies, Sri Chandrasekharendra Saraswathi Viswa Maha Vidyalaya University, Enathur, Kanchipuram, Tamil Nadu. jayarawman@gmail.com

2 the stock prices are assumed to be moving in an unbiased manner. The concept of efficiency has been developed by the various financial experts over the period of the year. Among them Fama (1965) defined an efficient market as a market where there are many numbers of rational investors actively competing with each other to predict market price of individual securities in which an important information is almost freely available to all participants. There are three forms of efficient market hypothesis (EMH). The forms of distinguished by the degree of information reflected in security prices. The weak form efficiency asserts that the past price will not influence the current price, so the investor cannot device a investment strategy to gain abnormal profit by way of analysing the past price pattern. Simply it denotes that technical analysis will be helpful to predict future stock price. The semi strong efficiency asserts that the current stock price reflects not only historical price information but also with publicly available information such as all types of corporate announcements will not be useful to investor to capitalize the information to yield surplus profit. The strong form of efficiency asserts that the stock prices reflect all relevant information including insider information. It implies that the insiders cannot make use of private information to design any form of superior strategy to earn abnormal profit. The present study is carried out to examine the weak form of efficiency on the Indian stock market. Since Indian stock market passes through the different stages over the last decade, the researcher takes the initiative in examining the efficiency of the stock market with respect to select indices (Market proxies) in the Bombay Stock Exchange. The time period of the study was taken from the year 2006 to The relevance of the period taken for the above study is asserted due to the recessionary trend started in the global context and Indian stock market too is not the exception in encountering the recessionary trend. About Bombay Stock Exchange Bombay Stock Exchange is regarded as the premier stock exchange and also one of the fastest growing stock markets among the merging market. The market capitalisation has witnessed a rapid growth over the last fifteen years. Four digits of index indicator have moved to seven digits in the last fifteen years. Further many structural changes took place in the Indian capital market to protect the interest of the common investor. SEBI, the apex body of regulation initiated many steps in the form of rules and regulation to bring consistency and uniformity in the trading activities over the stock market. In recent times Bombay Stock market has attracted foreign participants to increase the inflow of funds in the name of Foreign Institutional Investors (FIIs). The traditional system of trading has been changed in to automated system wherein rolling settlement has been introduced to curb the activities of speculators and arbitragers in a way to make abnormal profit. In addition to this the liquidity of stock exchange showed a drastic improvement over the period of the year to ensure in reducing the risk of the individual investors. All changes have taken place to indicate that Indian stock market is no way inferior to other stock market. Since Indian stock market has been under reform, Bombay Stock Exchange being a part of Indian stock market operations, BSE cannot be the exception in undergoing transitional phase and hence the present study is focused to examine the weak form efficiency of BSE during the last five years. Literature Review Kendall (1953) and Fama (1965) found existence of low degree of correlation in terms of price behavioural change with respect to weak form of efficiency. Granger and Morgenstern (1970) analysed the

3 sample of five stocks at various interval of time to test random walk theory. In that short term stock prices witnessed a deviation from random walk model. found that the market series did not follow the random walk model and there was an evidence of auto correlation in both markets rejecting the weak form of efficiency. Cooper (1982) took a sample of 36 countries and Samantha (2004) studied BSE 100 for a time period employed monthly, weekly and daily data of stock of seven years from January 1993 to December market to validate the random walk model through Till July 1996 the market showed remarkable correlation analysis and run test. It was found that inefficiency and high level of efficiency was shown except UK & US other markets showed a result in during July 1996 to December consistent with random walk model. Verma and Rao (2007) examined the companies Harvey (1995) studied the return of market and the included in BSE 100 and applied serial correlation power of predictability with respect to six Latin and unit root for a period of three years from 1998 to American countries, Eight Asian countries and Two 2001 and found that mixed results have been African countries and it was found that strong emerged in the market concerned. correlation in the stock return series. Renuka Sharma and Ramesh Chander (2011) Dasgupta and Glen (1995) found that operational studied the market proxies of BSE such as Sensex, inefficiency reflected in price behavioural pattern so BSE 100 and BSE 500 to verify whether the market that the price inefficiency was caused in nineteen proxies have followed random walk model and it emerging market, made the equity return of the was found that the result of unit root test and auto market to be dependent. correlation has differed during the different phases With respect to the Indian market many number of of the study. researchers supported to witness the evidence of The above studies have been made to analyse the weak form efficiency, Barja (1981, 1987), Rao and weak form of efficiency at different levels over the Mukerjee (1971), Sharma and Kennedy (1977), period of the years. If we look at the result of the Gupta (1985), Ramachandran (1985), Dhankar study, the study revealed mixed opinion with respect (1991) found through their contribution that Indian to the existence of randomness in the stock return capital market obeyed the rule of weak form over various phases of the market. efficiency. Bhaumita (1997) took sensex as the barometer of the Indian Capital Market to prove stock prices closely represent a random variable. Mohanty (2001) documented the evidences in favour of random walk model and in his study he considered sensex, BSE 100 and BSE 200 for the period of five years from 1996 to The result of the study supported the null hypothesis of Unit root test but the variance ratio test was rejected. Gupta and Basu (2007) tested two major equity market in India for the period of 1991 to 2006 and Objectives of the Study To know whether the returns of the index are normally distributed. To determine whether Indian stock market is weak form efficient. Hypothesis To meet the above objectives the following hypothesis are tested: H1: The stock returns are normally distributed H2: The series of stock return are independent to

4 each other and thereby investors cannot predict stock price to yield abnormal profit. Research Methodology In order to verify the above hypothesis the study has obtained 1,473 observations on daily basis from six market proxies viz BSE Sensex, BSE Auto, BSE Bank, BSE FMCG, BSE Health care and BSE Metal for the period of five years from 03/04/2006 to 31/03/2012. The required data for the present study was drawn from database provided by Yahoo finance. The daily closing prices are transformed in stock return series by using the formula. Unit Root Test The unit root test checks whether a series is stationary or not. Stationary condition is tested by using Augmented Dickey Fuller test and Phillipperron test. The ADF Test makes a parametric correction for higher order correlation. The ADF approach controls higher order correlation by adding lagged differences in terms of dependent variable. The formula for ADF test is as follows To test for stationary, the null hypothesis is: H0: ë = 0 and the alternative hypothesis is H1: ë < 0 Philip perron proposed a non parametric method for controlling higher order serial correlation in a series. The PP test makes a correction to the t- statistic of the co-efficient from the auto regression. The advantage of PP test is that it is free from parametric error. The objective of the study is to examine the random character of stock returns and to verify the level of dependence in the sequence of stock returns over the study period. The following methods are used to verify the above said objective: Descriptive statistics Auto correlation Unit root test ( ADF Test & PP Test) Descriptive Statistics The computation of descriptive statistics such as Skewness, Kurtosis and Jarque-bera test provides elimentary evidence about changes in the time series behaviour and explains the fact whether the return distribution of indices are normally distributed or not. Owing to the afore said fact, it is imperative to study further about time series properties such as the presence of Unit root in all the return series. Decision rule If t > critical value Accepting null hypothesis, i.e., unit root exist If t < critical value reject null hypothesis, i.e., unit root does exist Auto Correlation The auto correlation tests are used to notice any predictable trend in the stock price. The correlation of zero indicates that the price changes in consecutive time periods are uncorrelated with each other and it can be viewed as the rejection of hypothesis is that the investor can predict future price from the past one. A positive and statistically significant correlation can be viewed as evidence of price momentum in the markets and would suggest

5 that returns in a period are more likely to be positive. The above statement holds true for negative correlation also. The auto correlation matrix is estimated by: N Number of observation Statistical testing of auto correlation matrices requires standard error of estimated matrices (S.E. (k)), which is obtained as When n is sufficiently large (n > 50), approximation value of standard error of estimated (S.E. (k)) matrices is given by: Instead of testing significance of any individual autocorrelation matrix, the researchers tested joint hypothesis that all to ascertain lags are simultaneously equal to zero Empirical Evidences Descriptive Statistics Chart 1.1 Showing the Histogram and Descriptive Statistics of BSE Auto Chart 1.2 Showing the Histogram and Descriptive Sstatistics of BSE Bank

6 Chart 1.3 Showing the Histogram and Descriptive Statistics of BSE FMCG Chart 1.4 Showing the Histogram and Descriptive Statistics of BSE Health Care Chart 1.5 Showing the Histogram and Descriptive Statistics of BSE Metal

7 Chart 1.6 Showing the Histogram and Descriptive Statistics of BSE Sensex Auto Correlation Table 1.1 Showing the Auto Correlation Results of the Select Indices Lag BSE-METAL BSE-BANK BSE-AUTO BSE-SENSEX BSE-HC BSE-FMCG Source: Computed Data. The descriptive statistics in the above charts indicate the typical characteristics of time series data with respect to all the return series representing the sample market index. The Kurtosis values are 7.97, 7.28, 9.30, 9.70, 6.96 and 8.99 for BSE Auto, BSE Bank, BSE FMCG, BSE Health Care, BSE Metal and BSE Sensex respectively. The values for all market proxies are well above 3, indicating the return of the series are not normally distributed. In addition to that the jarque- bera test statistics for BSE Auto, BSE Bank, BSE FMCG, BSE Health Care, BSE Metal and BSE Sensex as shown are , , , , and respectively and statistically significant. With the help of the values of kurtosis and jarquebera test, the study explained the fact that return distribution of indices are not normally distributed.

8 1.0 Chart 1.7 Showing the Autocorrelation of BSE METAL BSE Metal Coefficient Upper Confidence Limit Lower Confidence Limit 0.5 ACF Lag Number As far as the BSE Metal index is concerned, lags 4,5,10 and 11 shows the negative correlation and the other lags except lag 12 show the positive correlation. All the sixteen lags are indifferent to zero at 5 per cent level of significance showing independent of stock price return: Chart 1.8 Showing the Autocorrelation of BSE BANK BSE Bank Coefficient Upper Confidence Limit Lower Confidence Limit ACF Lag Number

9 As far as the BSE Bank index is concerned, Out of sixteen lags, lag 2, 4, 5, 6 and 7 show the negative correlation and the the other lags show the positive correlation. All the sixteen lags are indifferent to zero at 5 per cent level of significance showing independent of stock price return: Chart 1.9 Showing the Autocorrelation of BSE AUTO BSE Auto Coefficient Upper Confidence Limit Lower Confidence Limit ACF Lag Number As far as the BSE Auto index is concerned, lags 4, 5, 6 and 14 show the negative correlation and the other lags show positive correlation. All the sixteen lags are significantly indifferent from zero at 5 per cent level of significance and the lags strongly disproved the statement of null hypothesis of independence in the successive movement of the stock index return series from the previous one: 1.0 Chart 1.10 Showing the Autocorrelation of BSE SENSEX BSE Auto Coefficient Upper Confidence Limit Lower Confidence Limit 0.5 ACF Lag Number

10 As far as the BSE Sensex is concerned, Out of 16 lags, 6 lags have negative correlation co-efficient indicating the irrationality of the behaviour of 1.0 Chart 1.11 showing the Autocorrelation of BSE HC Bse Health Care investors due to the over expectation but still all the 16 lags were indifferent to zero at 5 per cent level of significance showing higher level of independence of stock price return found to be existing: Coefficient Upper Confidence Limit Lower Confidence Limit 0.5 ACF Lag Number As far as the BSE Health care is concerned, Out of 16 lags, 6 lags showing negative correlation and all 16 lags are insignificantly different to zero at 5 per cent level of significance indicating validation of statement in the form of random walk model in the daily return series of the stock: 1.0 Chart 1.12 Showing the Autocorrelation of BSE FMCG BSE FMCG Coefficient Upper Confidence Limit Lower Confidence Limit 0.5 ACF Lag Number

11 As far as the BSE FMCG is concerned, though half of the auto correlation co-efficient had with negative side showing over expectation of the investor but the signs were statistically significant. All 16 lags are insignificantly different to zero at 5 per cent level of significance indicating validation of statement in the form of random walk model in the daily return series of the stock. ADF Test Unit Root Test The result of the unit root test at both ADF and PP test are not in tandem with the result obtained by the auto correlation test. The result obtained by the auto correlation vindicated that the market proxies have the independent return series. Whereas DF and PP test showed that the market proxies were stationary and further indicated that the absence of unit root in daily return of different indices representing sample proxies. Table 1.2 Showing the Results of ADF Test of the Select Market Proxies Market Proxy Value of t-statistic Critical value at 5% LoS Critical value at 1% LoS Status of Hypothesis BSE AUTO BSE BANK BSE FMCG BSE HEALTH CARE BSE METAL BSE SENSEX Source: Computed data. PP Test Market Proxy Table 1.3 Showing the Results of PP Test of the Select Market Proxies Value of t-statistic Critical value at 5% LoS Critical value at 1% LoS Status of Hypothesis BSE AUTO BSE BANK BSE FMCG BSE HEALTH CARE BSE METAL BSE SENSEX Source: Computed data.

12 References The value of t-statistics is less than the critical value Cooper, J.C.B (1982), World stock market: Some at 1 per cent and 5 per cent level of significance in random walk tests, Applied Economics, both ADF Test and PP Test. As a result of that the null Vol.14, pp., hypothesis is rejected showing that the there is no unit root in the market returns of the select proxies Fama and Eugene (1965), The behaviour of stock such as BSE AUTO, BSE BANK, BSE FMCG, BSE market prices, Journal of Business, 38, pp, 34 HEALTH CARE, BSE METAL and BSE SENSEX 105. during the study period and it is further implied that Garnger, C.W.I. and Oskar Morgenstern (1970), the returns of the market proxies are stationary. What the random walk model does not say, Conclusion Financial Analysis Journal, Vol.26, No.3, pp., As stated earlier in this paper, an attempt has been made to examine the weak form of efficiency of Harvey, C.R. (1995), Predictable risk and return in Indian stock market with special reference to select emerging markets, Review of Financial market proxies of Bombay Stock Exchange (BSE). Studies, Vol.08, No. 3, pp., Since Indian Capital Market has received a bolt due to the worldwide recession, the researcher was Kendall.M (1953), The analysis of economic time interested in verifying the movement of market series, Journal of the Royal Statistical Society, indices in terms of price efficiency during the period Series A, 96, pp., By analysing the facts and figures over the Mohanty, Pitabas (2001), Efficiency of the market period of the years, the market has showed evidence for small stocks, NSE Research Paper No.1 in terms of independence of successive stock return series in the select market proxies indicating Indian Ramachandran. J. (1985), Behaviour of stock Stock Market is moving towards efficiency at least market prices, Trading rules, Information & in the weak form. But at the same time it is to be market efficiency, Doctoral dissertation, noted that series of return are stationary with respect Indian Institute of Management, Ahmadabad. to ADF and PP test. Rao. K. N. and Mukherjee. K. (1971), Randomwalk Suggestion hypothesis: An empirical study, Though the efficiency of stock market in weak form Arthaniti, Vol.14, No.142. is proved through the time series analysis, the fact is Renuka Sharma and Ramesh Chander (2011), that the existence of irrationality of the market is Market proxies at BSE and weak form witnessed due to the prevailing of fundamental efficiency, Indian Journal of Finance, Vol.V. factors used by the market participants. Hence it is No.3, pp., suggested to apply more models in future to sort out the anomalies pattern of the stock return and thereby Sharma, J.L. and Kennedy, Robert E., (1977), A better and healthier remarks could be incorporated Historical analysis of market efficiency: Do on the weak form of efficiency of the Indian stock historical returns follow a random walk, Journal market. of Financial and Strategic Decisions, 10.