ICOQM-10

June 28-30, 2011

Testing Weak Form Stock Market Efficiency on Bombay Stock Exchange of India
Uttam Sapate [email protected] Indira College of Engineering & Management, Pune Valeed Ansari [email protected] Aligarh Muslim University, Aligarh
The weak form of market efficiency hypothesis (EMH) states that the stock prices reflect all the past information making technical analysis futile and abnormal profits cannot be made by studying the past & present stock prices to predict future stock prices. On 200 secondary shares of BSE forming part of BSE 200 index using daily closing prices for ten years, weak form market efficiency is tested utilizing Autocorrelation analysis, Ljung Box Q (LBQ) statistics and Runs test. The results signify that trading strategies based on historic prices cannot be used to gain abnormal profits consistently. Keywords: Indian Stock Markets, Weak Form Efficiency, Random Walk Theory

1. Introduction
The Efficient Market hypothesis (EMH) assumes that stock prices adjust rapidly to the arrival of new information, and thus, current prices fully reflect all available information. Fama (1970) formalized the theory, organized the empirical evidence, and divided the EMH into three sub-hypotheses. The weak-form EMH states that current stock prices fully reflect all historical market information such as: prices, trading volumes, and any market oriented information. The semi strong-form EMH asserts that prices fully reflect not only the historical information but also all public information including non-market information, such as earning and dividend announcements, economic and political news. Finally the strong-form EMH contends that stock prices reflects all information from historical, public, and private sources, so that no one investor can realize abnormal rate of return. Stock market efficiency is an important concept, both in terms of an understanding of the working of stock markets and in their performance and contribution of the development of a countrys economy. If the stock market is efficient, the prices will represent the intrinsic values of the stocks and in turn, the scarce savings will be optimally allocated to productive investments in a way that benefits both individual investors and the country economy. For many years, economists, statisticians and teachers of finance have been interested in developing and testing the existence of weak form of market efficiency. In Fama survey the vast majority of those studies were unable to reject the efficient markets hypothesis for stocks. On the other hand, there are several anomalous departures from market efficiency in the literature. Although a precise formulation of an empirically refutable efficient market hypothesis must obviously be model specific, historically the majority of such tests have focused on the ability to forecast of common stock returns. Within this framework, which is called random walk of stock prices, few studies have been able to reject the random walk model statistically. The investigation of the validity of the EMH has been and still is one of the favorite topics of the financial literature. A significant number of studies have examined the efficiency of stock markets using different methodological techniques. The parametric tests include the autocorrelation test, LBQ test. The non parametric tests include the runs test. The notion of an efficient capital market is based on the following set of assumptions. A large number of independent competing profit-maximizing participants to analyze and value securities; New information regarding securities comes to market in a random fashion; The market mechanism tends to adjust security prices rapidly to reflect the effect of new information. The term random does not imply that price movements are erratic or chaotic, just that prices respond only to new information. This information may be randomly good or bad and prices will therefore move in an unpredictable manner. The movements themselves are a perfectly rational response to the available information. The EMH has very important implications for both investors and authorities. This study aims at testing the weak-form EMH in the Indian Stock Market. The study is organized in four sections as follows: Section 2 229

ICOQM-10

June 28-30, 2011

provides brief literature review; Section 3 describes the Methodology adopted; Section 4 presents the analysis & results; Section 5 presents conclusions & remarks.

2. Review of Literature
The empirical testing of this random walk hypothesis (RWH) was initially mixed. Early studies by (Samuelson, 1965 and Fama, 1965) examined it and could not refute a random walk. RWH were tested heavily in both developed and developing markets. The developed markets are found to be weak-form efficient. That means that successive returns are independent and follow random walk (Fama 1965, 1970). These results of weak-form efficiency are confirmed considering a low degree of serial correlation and transaction cost. All these research works support the proposition that price changes were random and past changes were not useful in forecasting future price changes particularly after transactions costs were taken into account. Though it is generally believed that emerging markets are less efficient, the empirical evidence does not always support this thought. In India, weak form market efficiency has been widely researched and studies have generated a large amount of evidence of favor of weak form efficiency, e.g. Baruna (1981), Barua and Raghunathan (1987), Rao and Mukherjee (1971), Sharma and Kennedy (1977), Sharma (1983), Gupta (1985), Ramachandran (1985), Dhankar (1991), Saxena (1992), Belgaumi (1995), Mittal (1995), Yalawar (1988), Mishra (2000), and Gupta (2001). There have been only a few studies (e.g. Chaudhaury, 1991; Kulkarni, 1978) which did not support the weak efficiency hypothesis. Subramanian (1993) studied weak from of market efficiency and noted prominent spikes at lower frequency range through spectral analysis. It noted the presence of periodic cycles in the movement of share price which are against the assertion of the weak form of EMH. Gupta and Gupta (1997) opined that phenomenon of large departures from random price behavior might have been due to structural transformation taking place in the Indian Capital Market, Deb (2003) tested weak from efficiency using both parametric and non-parametric tests across five major market indices of the Indian stock market. It depicted that prices in the Indian stock market do not follow the random walk model except for BSE 100 indices, which endorses efficiency for the variance ratio test.

3. Methodology
Objectives of the Study Thus, a large majority of studies favor prevalence of weak form stock market efficiency on the Indian bourses. Some studies deny its existence to keep the academic debate alive on the subject. Under this backdrop, the present study was conceptualized to explore plausibility of profitable trading strategies based on past prices. In this regard, it may be construed as yet another attempt to re-examine the weak form efficiency proposition and to verify results reported in earlier studies for strategizing investment decisions profitably. The basic objective of this study is to assess whether stock prices in the Indian stock markets move as the random walks theory suggests. In other words, is the Indian stock market mechanism efficient in the manner stipulated in its weak form preposition? The objectives conceptualized in the study under consideration is to empirically test whether the weak form of efficient market hypothesis holds well in Indian stock markets encompassing observable interdependence and non-randomness. Hypotheses For the present study the hypothesis formulated (Ho) examines whether the stock returns follow a random walk (weak - form efficiency) during the study period. Null Hypothesis (Ho): The Indian stock market returns are random during the study period. Alternate Hypothesis (Ha): The Indian stock market returns are not random during the study period. Data Description In order to examine the validity of the above said null hypotheses, the daily closing share price of 200 stocks listed as part of BSE benchmark index BSE-200 as on 31st March 2010 are used. These stock prices then were converted into return series using MATLAB software function price2ret. The stock prices have been obtained from BSEs website and have been refined to remove duplicate data entries. There is collection of closing price for each stock for the period from 01/04/2000 to 31/03/2010. The computation of return series can alternatively be done using the formula given below:

Where; Rt = Daily stock return for period t Pt = Daily stock price for period t Pt-1 = Daily stock price for period t-1 Ln = Natural log 230

ICOQM-10

June 28-30, 2011

Autocorrelation Test & Ljung Box Q (LBQ) Statistics Autocorrelation is a reliable measure for testing of independence of random variables in return series. This test statistic is widely used to notice any perceptible trend in stock prices. The serial correlation matrices that measures correlation between price changes inconsecutive time periods and is a measure of how much price change in any period depends upon price change over the previous time period. A serial correlation of zero would imply that price changes in consecutive time periods are uncorrelated with each other, and can thus be viewed as a rejection of the hypothesis that investors can learn about future price changes from the past ones. A positive and statistically significant serial correlation could be viewed as evidence of price momentum in markets, and would suggest that returns in a period are more likely to be positive (negative) if the prior period returns were positive (negative). In a more precise way, serial correlation coefficients provide a measure of relationship between value of random variable (X), in time t and its value k-periods earlier. It indicates whether daily price changes in the period t are influenced by price changes occurring k-days earlier, where k=1,2,3. n. In the present study we have considered time lags of 1,2,3, .20 days. The LBQ statistic is often used as a test of whether a tine series is white noise. In large samples, it is approximately distributed as a chi-square distribution with m degree of freedom. In case the computed Q exceeds its critical value on the chi-square distribution at a given significance level, null hypothesis is rejected. The Runs Test for Randomness A In order to test for weak-form efficiency runs test is widely used as it does not require returns to be normally distributed. A run is defined as a succession of identical symbols which are followed or preceded by different symbols or no symbol at all. The number of runs is computed as a sequence of the price changes of the same sign (such as; ++, --, 00). When the expected number of run is significantly different from the observed number of runs, the test reject the null hypothesis that the daily returns are random. The expected number of runs is represented by:

Where n represents the number of observations, na and nb respectively represent observations above and below the sample mean (or median), and r represents the observed number of runs. The standard error can therefore be written as:

The asymptotic (and approximately normal) Z-statistic can be written as follows:

The null hypothesis for this test is for temporal independence in the series (or weak-form efficiency).

4. Results and Discussion

The independence hypothesis can be investigated by examining the autocorrelation function (ACF). The ACF shows the pattern of autocorrelations present in the time-series as well as the extent to which current values of the series are related to various lags of the past data. Autocorrelation tests show whether the serial correlation coefficients are significantly different from zero. In an efficient market, the null hypothesis of zero autocorrelation will prevail. The coefficients were obtained for 1 to 20 time lags to examine results for varying periodicity. Needless to add, such coefficients suffice the purpose for efficiency of market mechanism in its weak form. A significant interdependence in stock returns invalidates weak form market efficiency, while its absence endorses the weak form efficiency for absence of any kind of a trend in stock prices. Autocorrelation coefficients obtained for daily stock returns for 200 stocks forming part of BSE-200 were calculated using MATLAB software for 1 to 20 lags. Based on autocorrelation results on return series it is observed that there is very low degree of correlation at all the lags from 1 to 20 (Autocorrelation value less than 0.1 in all cases except for two stocks of Sr. No. 99 & 194). Sample stock returns consistently displays insignificant autocorrelation pattern at various lags. Seventy stocks out of 200 stocks do not show any serial correlation at 1 20 lags. The independence of stock returns as document through autocorrelation matrices is revalidated for significance through rigorous statistical analysis in order to conclusively comment on the status of market 231

ICOQM-10

June 28-30, 2011

efficiency in its weak form. LBQ-statistics is primarily developed to test joint hypothesis about the significance of all autocorrelation matrices at a given lag. It is used to derive more generalized conclusions and evaluate the validity of overall autocorrelation matrices. Using MATLAB the LBQ test was conducted. Results of LBQ statistics on return series exhibit mixed evidence for serial correlation at the corresponding element of lags. For 101 out of 200 stocks returns series indicate existence of serial correlation. The findings of serial correlation and LBQ statistic are inconclusive on acceptance of weak form of efficiency in Indian stock market. Stock market efficiency in its weak form can alternatively be studied by testing the stock prices for randomness. A random stock price behavior supports weak form market efficiency. For this runs test can be used which has its own merit over the serial correlation matrices and it considers only the directional shift in stock price movements and thus ignores the magnitude of price change. It is a nonparametric or distribution free test and makes no presupposition on the distribution of stock prices. The runs test was conducted using MATLAB and results provided in the Table 1 which can be used to examine the serial independence in share return movements.
Table 1 Summary of Runs test as applied on BSE 200 Stocks for Period from 2000 - 2010 Runs test Failed ( H = 1) Number of Stocks % Stocks 57 28.5% Runs test Not Failed ( H = 0) 143 71.5% Total Stocks of BSE 200 200 100%

Runs test reveal that the returns of 143 stocks out of the 200 are not significant at the five percent level with Zvalues within (1.96) and thus the hypothesis of randomness cannot be rejected. Therefore, the null hypothesis of randomness cannot be rejected and hence the results support the weak-form EMH of Indian Stock Market. This indicates that the stock prices traded on the Indian stock market follow the random walk model.

5. Conclusions
Based on Autocorrelation test, LBQ test and runs test carried out on the sample drawn from BSE it is concluded that the stock market returns follow random walk and they support the weak form of market efficiency. Hence, the empirical study supports the weak-form EMH of Bombay Stock Exchange (BSE) of India based on Autocorrelation test, LBQ test and runs test. The results obtained for BSE are considered to be applicable to Indian Stock Markets in general. It means that the Indian stock markets are weak form efficient and abnormal returns cannot be generated based on past price trends / information.

6. Scope for Further Research

The study can further be extended using other advanced unit root tests, Variance ratio tests and model comparison test.

7. References
1. 2. Anand Pandey, Efficiency of Indian Stock Market, SSRN id 474921, Oct 2003. Azizjon A Alimov, Debasish Chakraborty, Raymond A K Cox, Adishwar K Jain, "The Random Walk Hypothesis on the Bombay Stock Exchange," Finance India, Sept 2004. 3. Burton G. Malkiel, A Random Walk Down Wall Street, book by W. W. Norton & Company. 4. Deepak Khatri, Existence of Unit Root in Stock Index: Evidence from Indian Stock Market, SSRN id 643262. 5. Eugene F. Fama, Random Walks in Stock Market Prices, Financial Analyst Journal, Oct 1965. 6. Graham Smith, Random walks in Middle Eastern stock markets, Applied Financial Economics, Apr 2007. 7. Kapil Gupta, Dr. Balwinder Singh, Random Walk and Indian Equity Futures Market, SSRN id 874913, 2006. 8. Kausik Chaudhuri, Yangru Wu, Mean Reversion in Stock Prices: Evidence from Emerging Markets, Managerial Finance 2003. 9. Natalia Abrosimova, Gishan Dissanaike, Dirk Linowski, Testing the Weak-Form Efficiency of the Russian Stock Market, SSRN id 302287, April 2005. 10. Nauzer J Balsara, Gary Chen, Lin Zheng, The Chinese Stock Market: An Examination of the Random Walk Model and Technical Trading Rules, Journal of Business & Economics, Spring 2007. 232

ICOQM-10

June 28-30, 2011

11. Poshakwale S., The Random Walk Hypothesis in the Emerging Indian Stock Market, Journal of Business Finance & Accounting [serial online], November 2002. 12. R. Vaidyanathan, Kanti Kumar Gali, Efficiency of Indian Stock Markets, Indian Journal of Finance & Research, July 1994. 13. Ramesh Chander, Kiran Mehta, Renuka Sharma, Empirical Evidences on Weak Form Stock Market Efficiency: The Indian Experience. 14. Samir K. Barua, V. Raghunathan, Jayanth R. Varma, Research on the Indian Capital Market: A Review, Vikalpa 1994. 15. Silky Vigg, Simranjeet Kaur, Navita Nathani, Prof. Umesh Holani, Efficient Market Hypothesis: A Case Study on Bombay Stock Exchange, Indian Journal of finance, Oct 2008.

233