Capital Asset Pricing Model: The Indian Context

R Vaidyanathan

he Capital Asset Pricing model is based on two parameter portfolio analysis model developed by Markowitz (1952). This model was simultaneously and independently developed by John Lintner (1965), Jan Mossin (1966) and William Sharpe (1964). In equation form the model can be expressed as follows:

E (Ri) = Rf + i [E(rm) Rf] = Rf +im / m (E(Rm) Rf / m)

Where E(Ri) is expected return on asset i, Rf is the risk-free rate of return, E(Rm) is expected return on market proxy and i; is a measure of risk specific to asset i. This relationship between expected return on asset i and expected return on market portfolio is also called the security market line. If CAPM is valid, all securities will lie in a straight line called the security market line in the E(R), i frontier. The security market line implies that return is a linearly increasing function of risk. Moreover, only the market risk affects the return and the investor receive no extra return for bearing diversifiable (residual) risk. The set of assumptions employed in the development of the CAPM can be summarized as follows [Sears and Trennepohl (1993)]: 1. Investors are risk-averse and they have a preference for expected return and a dislike for risk. 2. Investors make investment decisions based on expected return and the variances of security returns, i.e. two-parameter utility function. 3. Investors behave in a normative sense and desire to hold a portfolio that lies along the efficient frontier. These three assumptions were also made in the development of the Markowitz and Sharpe single-index portfolio analysis models. In addition to these three, CAPM also makes the following assumptions. 4. There exists a riskless asset and investors can lend or invest at the riskless rate and also borrow at this rate in any moment. 5. All investments are perfectly divisible. This means that every security and portfolio is equivalent to a mutual fund and that fractional shares for any investment can be purchased in any amount. 6. All investors have homogenous expectations with regard to investment horizons or holding periods and to forecasted expected returns and risk levels on securities. This means that investors form their investment portfolios and revise them at the same interval of time

Research Papers in Applied Finance (e.g., every six months). Furthermore, there is complete agreement among investors as to the return distribution for each security or portfolio. 7. There are no imperfections or frictions in the market to impede investor buying and selling. Specifically, there are no taxes or commissions involved with security transactions. Thus there are no costs involved in diversification and there is no differential tax treatment of capital gains and ordinary income. 8. There is no uncertainty about expected inflation; or, alternatively all security prices fully reflect all changes in future inflation expectations. 9. Capital markets are in equilibrium. That is, all investment decisions have been made and there is no further trading without new information. Some of the above assumptions are clearly unrealistic. However, the assumptions are not as restrictive as it appears initially and some of them can be relaxed without altering the basic nature of the model as we explain below. [Sears and Trennepohl (1993)] Theoretical Implications of Relaxing the above-mentioned assumptions: 2. Inclusion of skewness (third moment) in the pricing model has led to the three moment CAPM. 4. a. Different borrowing and lending rates lead to different CAPM lines and no general equilibrium pricing model. b. No riskless asset exists, leading to the zero beta CAPM, which provides for a theoretical explanation of the basic CAPM empirical results. c. There is riskless lending but no riskless borrowing, leading to the zero beta CAPM 5. CAPM would be series of line segments, each representing portfolio positions with no fractional shares. 6. Different expectations lead to different CAPM lines and no general equilibrium pricing model. 7. a. Inclusion of transactions costs in the model would produce bands around the relationship, leading to fuzzy equilibrium. b. Consideration of taxes leads to an alternative CAPM model that incorporates the differential tax effects of dividends and capital gains.

Empirical Implications of Relaxing the above-mentioned Assumptions

2. The general conclusion of tests of this model indicate that skewness is important in the pricing of securities. In particular, whenever the market portfolio is positively (negatively) skewed, investors are willing to accept (require) a lower (higher) average return in exchange for positive skewness with the market portfolio. 4. a. Assumption cannot be tested empirically, band c. some empirical studies support the zero beta CAPM, but others do not. 5. Assumption has not been tested but is probably not a major empirical problem for CAPM.

Capital Asset Pricing Model: The Indian Context 6. This is a major empirical problem for the CAPM. 7. a. Because of relatively low transactions costs, this does not seem to be a major problem. b. Most studies have shown that tax effects via the dividend yield is important in the pricing process. In particular, there is a positive relationship between dividend yields and average returns. One of the important outcome of the CAPM assumptions is that all investors hold a portfolio which is a combination between riskless portfolio and market portfolio. This is because all investors will have identical efficient frontiers due to the assumption of homogeneous expectations. They can however have different utility functions, which will decide what combination of riskless portfolio and market portfolio the investor will choose. This implies that all investors hold the same combination of risky securities namely, the market portfolio. This is also known as the separation theorem. The market portfolio in CAPM is the unanimously desirable risky portfolio which contains all risky assets. Thus return on market portfolio is weighted average of return of all risky assets in the market and in theory it should contain, besides ordinary shares, all assets, like art objects, commodities, real estates and so on. However in practice it is impossible to construct a market proxy which contains all assets and thus, all the commonly used market indices roughly replicate the market. The total risk of a portfolio can be measured by the variance of its return. In a more general situation of a portfolio p consists of n shares and any individual share i has a weightage of Xi in the portfolio, then the total risk can be expressed as follows:

2p = n t=1 Xi2 + (n t=i Xi i)2 2m

2 2 i-e p = ep

+ p2 2m

Total Risk = Unsystematic Risk + Systematic Risk If CAPM holds, then investors should hold diversified portfolios and the systematic risk or non-diversifiable risk will be the only risk which will be of importance to the investors. The other part of the risk, known as the diversifiable risk or unsystematic risk, will be reduced to nil by holding a diversified portfolio. Thus beta, which is a measure of the non-diversifiable risk in a portfolio, is most important for investors, from the point of view CAPM theory. In case the CAPM holds in the market, an investor will no longer require any sophisticated portfolio selection technique to select his portfolio. He will choose a mix between risk-free rate and the market portfolio based on his utility function. In other words optimal investment decision will be simply to buy the market portfolio. This investment decision is independent from the decision about how to finance the investment i.e. whether to lend or borrow at the risk-free rate. Ideally, if CAPM holds, there will not be any identifiable inefficiency in the market and all securities will lie on the

Research Papers in Applied Finance security market line (no security can be found which is wrongly priced). However, such a situation is not realistic even in a highly developed and efficient capital market as in the United States. But on an average, if the inefficiencies in the market are not extreme, the assumptions of the CAPM can be approximately valid even in a realistic situation. In such a situation majority of the securities (assets) in the market will be efficiently priced. Thus even though it is known that no market in the world is efficient in a perfect sense, empirical tests of CAPM can still give meaningful results. In testing CAPM, an equation which is similar in nature to the one factor market model of Sharpe (1964) is used. The one factor market model, which is based on much simpler assumptions, can be compared with the CAPM. In equation form the market model can be expressed as follows:

Rit = ttP +it

Where i =COV (Ri, Rm) / m The market model assumes that the joint probability distribution between R it and Rmt is stationary and bivariate normal, where Rmt is a factor which describes security returns. The CAPM as described earlier is a one period model describing expected return. The market model can be expressed as (given expected market return):
2

Putting this value of i in the original market model one can get the following equation:

E (Rit) = i + iE(Rmt) If CAPM holds then i =Rf (1-i)

Rit = Rft + i(Rmt - Rft) + it

This is the ex-post version of CAPM. If Rft and Rmt are correlated over time then estimates of i and i will be biased in opposite directions. However, even if Rft and Rmt are correlated, the standard deviation of Rft is close to zero. Therefore the covariance between Rmt and Rft is practically nil and market model beta and CAPM beta are close in magnitude (Alexander & Francis, 1986).

Survey of Literature
Tests of CAPM
In the following some direct tests of CAPM, which were conducted in different periods have been listed. Among the tests listed the two most important and widely followed test are that by Black, Jensen, Scholes (1972) and Fama-MacBeth (1973). all the tests are listed in chronological order. Fischer Black, Michael C. Jensen and Myron Scholes (1972) studied the following equation using 60 months data from NYSE:

Capital Asset Pricing Model: The Indian Context

rit - rn = i + i (rmt rn) + it

Between 1926-64 thirty five overlapping five year period was used to estimate the constant term and beta in the above equation. The estimates of beta were used to rank the stocks into 10 portfolios. For each 5 year period subsequent 12 months were used to calculate the rates of return for the portfolios, thus obtaining a set of monthly rates of return for the period 1931 to 1965 for each of the 10 portfolio. Time series OLS estimate of and was obtained for 420 months data and 4 sub-periods data. Except for the first subperiod the and are found to be inversely related. Cross sectional test of CAPM was done using the following equation:

ri - rf = to + ti i + i, for i=110.
This was estimated for the entire 35 years period as well 4 sub-periods. The estimate of intercept term was significantly different from 0 and estimate of slope significantly less than average excess return on market portfolio for all the sub-periods as well as for the 35 year period. Thus the intercept term is too large and slope too small for CAPM. The authors argued that this could happen if zero beta CAPM is valid. Then the following equation should hold:

rit - rzt = i + i (rmt rzt) + it

Where rz is the return on zero beta portfolio and expected value of i is zero. The results are consistent with this version of the CAPM. However, the authors do not perform a separate test to examine this. The most important empirical investigation of CAPM was done by Eugene F. Fama and James D. MacBeth (1973). In this study they divided the sample stocks in to 20 portfolios on the basis of estimated beta rankings. The data used for this study are monthly percentage returns including dividend and capital gain for all stocks in the NYSE for the period between January '26 to June '68. They used three separate periods for portfolio formation, beta estimation and final testing. Beta estimated from the data in each period of portfolio formation were used to rank the shares and form 20 equal sized portfolios. The following 5 years data were used to recompute beta, and these were averaged across securities within portfolios to obtain 20 portfolio betas for the risk-return tests. These portfolio betas were computed month by month in order to account for any delisting of securities. Measure of non-beta risk was obtained as the standard deviation of the residual risk in the market model for each securities. For each month in the test period they regressed portfolio return with the portfolio beta, portfolio beta square and portfolio residual error each estimated from the previous estimation period. The following equation was used:

Rpt = ot + it pt-1 + 2t 2pt-l + 3t S (ept-l) + pt for p = 120

Research Papers in Applied Finance Fama and Macbeth (1973) had estimated the above full equation as well as forms of equation where values of the coefficients of both beta square and residual error are both separately and simultaneously forced to zero. If both theory and empirical evidence indicate that one or more variable has no influence, better estimates can be made when those variables are excluded. The results indicate that neither the beta squared term nor the residual risk has any influence on stock returns. The performance of beta coefficient over the entire period indicates the relationship between expected return and beta is linear and positive. Fama and Macbeth finds that the intercept term is generally substantially greater than riskfree rate of return and beta coefficient is substantially less than premium on market portfolio. This indicates that zero beta model is more consistent with equilibrium condition than the simple CAPM.

Criticism
Richard Roll (1977) in his paper asserted that the asset pricing theory is not testable unless the exact composition of the true market portfolio is known. Using a market proxy is subject to two problems, firstly the proxy may be mean variance efficient even when true market portfolio is not mean variance efficient and secondly proxy may be inefficient when market can be either efficient or truly inefficient. The exact composition of the market portfolio is practically not possible to determine. Thus according to the author it is not possible to test CAPM empirically. Literature on Anomalies in CAPM Jack Clark Francis and Frank 1. Fabozzi (1979) conducted a study over a period of 73 months between December 1965 and December 1971 on 694 stocks listed in NYSE. The study looked into the stability of the single index market model (SIMM). The result of the study supports the hypothesis that SIMM is affected by macroeconomic conditions. The intertemporal instability in the betas frequently observed could be due to this business cycle economics. Elroy Dimson (1979) proposed that the intervaling effect (tendency of explanatory power of market model regression equation and mean value of beta estimated from value weighted index to rise as differencing interval is increased) found to exist in the testing of market models is indicative of a possible non-trading problem. The author suggests a method to estimate beta when the data suffers from this problem. The method, called aggregated coefficient method (AC) were applied on listed stocks of London Stock Exchange between January 1955 and December 1974. The author argued that under nontrading problem it is not possible to determine the most efficient method of estimating betas and thereforeAC method is attractive when transaction times are not known. Robert H. Litzenberger and Krishna Ramaswamy (1979) derived an after tax version of CAPM. Share price data from January '36 to December '77 was used in the study (504 periods). The results of the study indicate that there is a strong positive relationship between before-tax expected returns and dividend yields of common stocks. Richard Roll (1981) found the trading infrequency to be an important cause of bias in short interval data. As the small firms are traded less frequently the risk measures for these

Capital Asset Pricing Model: The Indian Context firms are downward biased. The bias is very large in daily data and is also present in returns from monthly data. According to the author this bias can possibly explain effects like the small firm effect, low P/E ratio effect and high dividend yield firm effects, present in the market. Jeffrey F. Jaffe, Randolph Westerfield, M. A Christopher (1989) studied two sets of indices from the US and one set each from Canada, Australia, England and Japan. The authors find that Monday return for common stocks is negative only when market has declined in the previous week. The findings are inconsistent with market efficiency. The inconsistency cannot be explained by serial correlation arising from infrequent trading or higher risk on those particular Mondays. Hee-Kyung K. bark (1991) in his study used the Fama-MacBeth methodology to test the CAPM in the Korean market. The data was collected from the Daewoo- Yonsei database on monthly stock returns between the period January 1980 to December 1987. The period was subdivided into five overlapping periods of 4 years. The study tests the positive risk return trade off of CAPM. For the entire period there was a negative sign in the market premium. The residual risk was also found to be a significant factor. Thus the results indicate CAPM cannot be a predictive model in the Korean market. Eugene F. Fama and Kenneth R. French (1992) tested CAPM using stock return data between 196390 from NYSE, AMEXA and NASDAQ. The results do not support the Sharpe-LintnerBlack CAPM model's positive relation between average stock return and beta. The results indicate that size and book-to-market equity capture cross sectional variation in average stock returns associated with size, E/P, book-to-market equity and leverage. The authors concluded that if asset pricing is rational than size and book-to-market equity must be a proxy for risk. However, if asset pricing is irrational then these two variables may not be proxy for risk and the persistence of these results are doubtful. As indicated in the literature survey, most of the empirical tests of CAPM have been conducted on NYSE and based on the basic methodology adopted by B lack, Jensen, Scholes (1972) and Fama-MacBeth (1973). Besides testing for CAPM, many of the studies have concentrated on anomalies in the CAPM model, namely, size effect, seasonal effect, tax effect, dividend effect and problems due to misspecification in the CAPM model. In spite of the criticism of Roll (1977) on the validity of tests of CAPM, it is clear that the studies on CAPM have provided valuable insights to the stock returns behavior in various capital markets across the world. If systematic risk and return are linearly related and residual risk is unrelated to return, it will have important implication for investors. The tests of CAPM, though not entirely satisfactory and suffers from the market index identification problem, often produces results that can be expected from tests of CAPM. While one should continue the search for true tests of CAPM, one can perhaps proceed on the basis of the results produced by tests of observable, but not optimum, phenomena. Tests in the Indian Context N. Krishna Rao (1971) tested the Random Walk hypothesis on Indian Aluminium weekly average share price data for a period of 16 years (1955-70) collected from the Calcutta Stock Exchange. Spectral analysis of the data indicated that Random Walk hypothesis holds for Indian Aluminium.

Research Papers in Applied Finance J.L. Sharma and Robert. E. Kennedy (1977) tested the applicability of Random Walk hypothesis in India, London and US. Data for last Friday in every month was collected for a period of 132 months. Spectral density estimated fot first difference series for each index confirmed randomness of the series. No systematic cyclical component was found in the indexes. S.K. Barua (1981) in his paper examined the efficiency in the Indian capital market using Runs test and Serial Correlation test with lags up to period 8 on 20 securities and the market index. Daily data over the period of July 1977 to June 1979 was collected. In summary the preliminary evidence suggests towards market efficiency in the Indian capital market. L. C. Gupta (1981) studied share price data between the period of 1960-76. Each year's high and low price for the sample shares were considered. A total of 606 equity shares for one or more holding periods were considered in the study. The data was collected from Bombay, Calcutta and Madras stock exchange. The long-term rates on equities were less than that on debentures, preference shares, company deposits and long-term bank deposits most of the time. The belief that equities provided hedge against inflation was found to be unfounded. The author doubted the applicability of CAPM in the Indian capital market. J.L. Sharma (1983) tested the market efficiency in the Indian capital market. The data consisted of 23 stocks listed in the Bombay Stock Exchange between the period 1973-78. Thus the results indicate that Random Walk hypothesis holds for the Bombay Stock Exchange. Y.B. Yalwar (1985) took a sample of 122 actively traded shares between the period 1963-82. He calculated rates of return for each sample stock using geometric mean monthly return method for holding periods of 1 year, 5 year, 10 year and 15 year. The study showed that the equity returns are high and consistent with the market risk premium. Beta estimates for all securities were positive and significant except for two samples. Second pass regression was tested on mean return. The results indicated CAPM is a good descriptor of the Indian capital market. The BSE is efficient at least in the weak form as far as pricing of actively traded shares are concerned. S:K. Barua and V. Raghunathan (1986) studied the efficiency in the Indian Capital Market with a special reference to Reliance issue of convertible debenture. They showed that an investor operating in the forward market can earn abnormal return compared to an investor operating in the cash market. S.K. Barua and V. Raghunathan (1987) tested the market efficiency in India based on actual returns. The results indicate that the Indian capital market in inefficient in pricing its securities. N.P. Srinivasan and M.S. Narasimhan (1988) in their note criticise the papers ofBarua and Raghunathan (1986 & 1987) since it considered only ex-post return. Ex-post violation of risk-return parity need not suggest market inefficiency. V.K. Vasal (1988) in his study looked into the effect of corporate financial decisions and share price behavior in the Indian capital market. The sample data for the study was taken from Indian Cotton Textile Industry. The test period was three cross sectional periods 1979,80 and 81. The empirical results indicate that the Indian capital market is reasonably efficient in valuing a firm. N. Krishna Rao (1988) tested the efficiency level in the Indian capital market. The sample consists of week-end prices of 10 blue chip companies in the Bombay Stock Exchange adjusted for bonus and rights issue. The period is July 1982 to June 1987. The

Capital Asset Pricing Model: The Indian Context result of the above studies supports the hypothesis that Indian capital market is at least weakly efficient. Uma Shashikant (1988) conducted the study on a sample of 100 companies, selected from actively traded shares in 1964. The study period was 1965-87. contrary to the findings of L.c. Gupta the rate of return was found to be increasing withholding period. Except for war years of 1965 and 1971 other returns were positive, confirming sub Martingale model of share price. A study on asset pricing in Indian markets has been conducted by J.R. Verma (1988). The study does not reject the CAPM. However, as the author notes, replication on larger samples of securities is desirable to provide conclusive evidence in favor of the theory. G. C. Maheshwari and K.R. Vanjara (1989) conducted the study on a sample of 142 securities. The scrips were selected on the basis of their performance and assets during the year 1986. The study spanned from January 1, 1980 to December 31, 1986. The study indicates that the Indian capital market is probably not very efficient. LM. Pande and Ramesh Bhat (1989) mailed questionnaires to 600 prepares and users of accounting information to study their perception about the Indian capital market. The result of the study indicates that the Indian capital market is perceived to be inefficient. S.K. Barna and V. Raghunathan (1990) in their paper studied 23 leading company stock prices. They calculated PIE ratio based on fundamental analysis and compared them with actual PIE data. The results indicated on an average shares are over valued in the Bombay Stock Exchange. Obaidullah (1991) tested for the normality of stock market returns in India. He used sensex data from April 1979 to August 1991 and Natex data from April 1984 to November 1991. He found that daily returns differed significantly from normality whereas monthly sensex returns differed significantly from normality whereas monthly sensex returns were not significantly different from a normal distribution. The monthly returns were positively skewed and leptokurtic but not statisticaIly significant. Further the deviations were lesser where logarithmic price changes were used as against the percentage price changes. A study by Obaidullah (1991 a) reported that the stock price adjustment to release of value-relevant information is inaccurate. This implies that at any given point in time there are undervalued and overvalued stocks in the market. Prices are not equal to their fundamental intrinsic values. Hence, a risk-return parity cannot be expected to hold good. Another study by Obaidullah (1991 b) attributes abnormal returns to price-earnings ratios. The abnormal returns are also observed to persist. The so-called CAPM equilibrium is never reached. R.N. Agarwal (1991) looked into dividends and stock prices in the commercial vehicle sector in India. He covered the period between 1966-67 to 1986-87. The regression result suggests that current dividend behavior is explained by the current level of net profits and the two past dividends. The adaptive expectation hypothesis (past share prices are expected to be adapted) are supported by the result. Satinder Palaha (1991) conducted the study on a sample of 419 stocks divided into 11 industry groupings. Individual security betas of forty stocks as well as portfolio beta were estimated for the entire period of 1976-85 as well as its 5-year sub-periods. OLS was used in the estimation procedure.

236

Research Papers in Applied Finance

Over two sub periods the betas showed a high degree of stability. However, the individual security betas showed high instability. Out of 40 individual securities only in 5 cases the model had some degree of relevance. Vaidyanathan & Ray (1992) found that for companies belonging to chemical industries, market risk is less than 40%, as a percent of total risk. In the case of other types of industries market risk was less than 50% of total risk. When individual investors hold small number of stocks and in the context of large proportion of firm specific risks, efficient portfolios do not get formed. Vaidyanathan & Gali (1993) found a settlement period effect in the Bombay Stock Exchange scrips during 1989 and 1990. The average return on the first trading day of the settlement period is usually higher than that on the last trading day and the intermediate days. In fact it is higher than the overall daily average returns. Ray (1994) conducted a test of CAPM using 170 actively traded scrips on the Bombay Stock Exchange. He used monthly data over the period 1980-91. He used three market indices, the RBI index, ET index and the BSE Sensitive Index. He used the Fama-MacBeth methodology and found that CAPM does not seem to hold for the Indian capital market. A study conducted by Obaidullah (1994) used monthly stock price data for a period of sixteen years (1976-91) for a sample of thirty stocks. The results from the exercise, however, do not lend themselves to any supportive or contradictory interpretation. The coefficients of 2p are, in general, not statistically significant. This is in conformity with the CAPM. However, in the multiple regression model, the coefficients of p also in most cases become statistically insignificant which is contrary to what the CAPM predicts. Hence he suggests that CAPM as a description of asset pricing in Indian markets does not seem to rest on solid grounds. Vaidyanathan & Gali (1994 a) studied the variation in various indices (Sensex, ET index and Natex) and found that one scrip (Reliance) explained more than half of the variation in the indices during 1989 and 1990. In case Hindustan Lever is also considered then the two scrips explain around 70% of the variation in Sensex and Natex. Vaidyanathan & Gali (1994 b) studied the efficiency of the stock market (weak form) using runs, serial correlation and filter tests at four different points for the period 1980 to 1990 for ten scrips. The evidence from all the three tests support the weak form of efficiency. Sehgal (1994) used data of the Natex and 80 individual securities over the period April 1984 to March 1993 and used logarithmic price changes. Testing for the significance of skewness and kurtosis we found that for Natex skewness is not significant but kurtosis is significant. For individual securities a vast majority had significantly positive kurtosis. Further, each of the randomly formed portfolios of eight securities were also found to significantly deviate from normality. However, the sample period includes the security scam period of February, 1992 to May, 1992 during which period there were extreme variations in the indices and stock prices. The effect of these could affect the outcome of the test. Gali (1995) has tested for the normality of the returns of Sensex, ET index and Natex during May, 1987 to June, 1994. He constructed daily, weekly, settlement periodwise and monthly returns. Monthly and settlement period-wise returns were normal for all the indices.

Capital Asset Pricing Model: The Indian Context Imperfections in the Indian Capital Market The results from some of these studies indicate that the Capital Asset Pricing Model probably cannot explain the risk-return relations in the Indian capital market. There could be several reasons for the empirical data not supporting the model. As pointed out in the literature survey, one of the shortcomings of any ex-post test of CAPM is the difficulty in defining the market portfolio. The assumptions of CAPM imply that the market portfolio reflects the universally preferred combination of risky assets. The market portfolio in CAPM should ideally include all assets. Naturally, for testing purposes only a reasonable proxy for the market portfolio has to be used. Thus, if the market proxy is not properly defined tests of CAPM may give misleading results. However many studies have used different indices, namely, RBI Index, Economic Times Index and the BSE Sensitive Index, as market proxy. Thus, the possibility that the results are distorted due to problems in the construction of the market index appears to be quite low and other more probable causes need to be explored. As indicated earlier, the results of a Korean study, Bark (1991), do not support CAPM. For the entire period (1980-87) of the study, there was a negative sign in the beta coefficient and the residual risk was found to be statistically significant. The South Korean market in many ways is similar to the Indian market. Unlike stock markets in the developed countries the markets in both South Korea and India are relatively new and growing. As the author of the paper noted, the Korean market suffers from inefficiency due to information barrier and inadequacies in the infrastructure. The influence of large shareholders, who trade on monopolistic information, also impair the efficiency of the Korean market. The Korean investors also hold non-diversified portfolio, which contradicts one of the assumptions of CAPM. The Fama & French (1992) study was conducted on return data between 1963-90. The results of this study do not support the positive risk-return hypothesis of CAPM. The authors find that the size and book to market value of equity explain the crosssectional variation in average returns during the period. However, the authors are not sure whether asset pricing is rational and whether the two factors, size and book to market of equity, can be regarded as a proxy of risk. Thus, it can be seen that there have been less empirical support to CAPM in the very recent studies abroad and the ability of beta to reflect the risk of a security is doubtful. In this background the results of various studies in India do not come as a major surprise. Moreover the efficient market assumptions behind CAPM is likely to be less valid in India compared to the developed country markets, where the securities trading is much more efficient in terms of greater transparency in transactions, faster and easier availability of information related to the market, shorter settlement periods, less transaction cost, greater liquidity and depth of the market, etc. Some of the more important factors which may cause CAPM to be ineffective in the Indian context and has the potential to reduce the efficiency level of the Indian Capital Market are described in the following page.

Research Papers in Applied Finance 1. Non-Diversified Portfolio Holding Like the Korean investors mentioned in the study of Hee-Kyong Bark (1991), the Indian individual investors also hold undiversified portfolio. The median size of share portfolios for Indian households is only 4.7 (L.C. Gupta, 1991). The percentage distribution of share portfolios as found by L.C. Gupta from a survey is reproduced below.

Table 1: Portfolio Holding Pattern among Investors

No. of Companies 1 2 3 4-5 6-10 11-20 20 Distribution of Share Portfolio % Distribution 13.1 11.9 9.9 17.4 18.1 13.9 15.7 Cumulative % 13.1 25.0 34.9 52.3 70.4 84.3 100.0

Source: Indian Share Owners -A Survey, L.c. Gupta, 1991, Page 55. Table 1 clearly indicates that the average investor in India holds very few scrips in their portfolio. This goes directly against the expectations of CAPM where the investors are expected to hold a combination of risk-free asset (or zero beta asset) and market portfolio. The investors are not expected to hold an undiversified portfolio as they are not rewarded for bearing unsystematic risk according to CAPM. It is also to be noted that as the study by Vaidyanathan and Ray (1992) indicates, unsystematic risk constitutes more than 60 percent of total risk for many companies. Hence, holding small number of securities or undiversified portfolios can add to market inefficiency.

2. Liquidity Liquidity is possibly the most serious problem faced by the Indian investors. A consultative paper by SEBI indicated a poor liquidity situation at the stock exchanges in India. Based on 1984-85 data this paper indicated that only 6% of shares are traded daily on all the stock exchanges, 12% are traded one in a fortnight, 28% are traded once in a month and another 28% are traded once in a year. A more recent data on the frequency of trading at BSE is provided in Table 2. From the table one can see that at Bombay Stock Exchange only 20% of the shares were traded frequently. More than 50% of the shares were traded on less than 10% occasions. Overall, the liquidity position in the exchange is highly unsatisfactory.

Capital Asset Pricing Model: The Indian Context Table 2: Frequency Distribution of Trading at BSE (April, '88-March, '89) % of Trading Days No. of Companies % of Total

20.0 > 90% 456 6.7 80-90% 152 4.1 70-80% 94 2.5 60-70% 57 3.2 50-60% 72 2.9 40-50% 66 2.2 30-40% 51 3.5 20-30% 79 4.1 10-20% 93 50.8 Up to 10% 1155 Source: Stock Exchange Trading in India, L C. Gupta, 1992, Page 56. The trading in the exchanges in India is highly concentrated on a few scrips. The trading velocity (total trading volume in the year divided by market capitalization) is a good indicator of the level of activity in a scrip. Table 3 gives a comparison of share trading velocity at various exchanges. Table 3: Share Trading Velocity in Major Stock Exchanges Japan USA UK Germany France Netherlands BSE Top 5 Cos Top 25 Cos Top 50 Cos Other Cos All Shares at BSE 0.64 0.57 0.39 1.72 0.29 0.57 1.81 1.45 1.04 0.19 0.57

Source: Stock Exchange Trading in India, L. C. Gupta, 1992, Pages 8 and 9. As can be seen the trading velocity at BSE, if the top 50 shares (by trading volume) are excluded, is far below the figures of the developed countries. As the above discussion suggests, the stock markets in India are lacking in both breadth and depth. Much of the trading in the exchanges are concentrated in a few scrips. Most of the scrips outside the specified list suffer from infrequent trading problem. Large institutional holdings (almost 40%) have further contributed to the problem. The absence of a large and well spread out investor base has reduced the amount of floating stocks available in the market, and thus seriously impaired the liquidity. Lack of liquidity can violate the assumptions of CAPM in two ways. Firstly it results in a transaction cost for the investors. If the transaction cost is added to the CAPM model, there will be a price band around the SML in which the scrips can lie. Within this band, it will not be profitable for investors to buy or sell shares. Secondly, CAPM assumes that all assets are infinitely divisible and readily marketable. This assumption is

Research Papers in Applied Finance also violated in India, due to the low liquidity observed in majority of the shares, as discussed earlier. Low liquidity can also result in inefficient pricing of scrips and price setting behavior by investors (non-price taker).

3. Insider Trading
Insider trading is believed to be rampant in the Indian market. The lack of transparency in the trading system facilitates insider trading. Earlier there was virtually no law against insider trading. After SEBI was formed, it has taken several steps to protect the small investors and prevent insider trading. In specific cases it can carry out investigations on alleged insider trading. Greater transparency in transactions will make insider trading more difficult to hide. However, the task of detecting insider trading is a difficult one. Even in developed countries, where there are elaborate systems to prevent insider trading in existence, insider trading allegedly takes place. The best way to reduce the possibility of insider trading is to reduce the scope of making profit through it. This can be achieved by ensuring speedy availability of price sensitive information to the public. Insider trading in a way can improve the efficiency in the market, as it can correct prices of scrips for information, which is not even publicly available. However, in a market dominated by insider trading, the investors cannot have homogeneous expectations as assumed in CAPM. Moreover, the very presence of insider trading implies that market price do not reflect all information (otherwise insider trading will not be profitable), i.e. market is not perfectly efficient. 4. Lack of Transparency Indian stock markets suffer from lack of transparency between members and constituents. Members perceive that the prices of transactions are not properly reflected in their gains. All intra-day quotations are not readily available. Since exact time of the transaction is not known, disputes persist. Also it is felt that some transactions are not reported. In such a context any analysis has to consider the limitations of available price series. 5. Inadequate

Infrastructure

The infrastructure in the stock markets in India are woefully inadequate. The stock exchanges are faced with inadequate office space, lack of computerization and communication system, etc. These inadequacies in turn has affected the quality of the investor service provided by the members of the exchanges. Though the number of investors as well as the volume of transaction has gone up many fold in recent years, the basic infrastructure and system has almost remained unchanged. Besides problems of office space and inadequate technology the present number of brokers in the exchanges are also not sufficient to provide proper (L.C. Gupta, 1992) services to the vast investor population. Moreover as the number of brokers are small, they almost have an assured volume of business. As the brokers now do not have to compete with each other much to get business there is no incentive for them to improve the quality of the investor service. The cost of the service provided by them also tend to be high. The monopoly of the brokers increases the transaction cost of investors.

Capital Asset Pricing Model: The Indian Context The settlement system in India is antiquated, requiring large volumes of paperwork, which results in inordinate delays. The settlement period of 14 days is also very long compared to settlement periods in exchanges in developed countries. The lack of infrastructure adds to the transaction cost of the investors. Moreover inadequate infrastructure and delays in settlement can slow down the absorption of price sensitive information in the market, affecting its overall efficiency. Both increased transaction cost and low operational efficiency violates the assumptions of CAPM.

Conclusion
In order to carry out a through analysis of the capital market we need to strengthen the database. Unlike the U.S. context where data is available dating back to 1920s, we do not have historical data on tapes. We also do not have data adjusted for bonus/rights, etc. over a long period. This is one area where analysts, market participants and regulators should give more attention. A weak database, combined with lack of sense of history can only add to market inefficiency. To facilitate the making of a twenty-first century market, we need to strengthen the infrastructure, improve liquidity, minimize insider trading and enhance transparency. Till then asset pricing will continue to be inefficient and provide opportunities for extra gains to the initiated - those initiated into the intricacies of the imperfections of the Indian market!

References
Agarwal, R.N., 1991, Dividends and Stock Prices: A Case Study of Commercial Vehicle Sector in India 1966-67 to 1986-87, Indian Journal of Finance and Research, January, VoU, No.1, pp.61-67. Alxander, Gordon J. and Jack Clark Francis, 1986, Portfolio Analysis, 3rd Edition, Prentice-Hall, Englewood Cliffs, New Jersey 07632. Bark, Hee-Kywzg K, 1991, Risk, Return,and Equilibrium in the Emerging Markets: Evidence from the Korean Stock Market, Journal of Economics and Business, November, Vol. 43, No.4, pp. 353-362. Barua, S.K, 1981, The Short-Run Price Behavior of Securities: Some Evidence on Efficiency of Indian Capital Market, Vikalpa, April, Vol.6, No.2, pp. 93-100. Barua, S.K and V. Raghunathan, 1986, Inefficiency of the Indian Capital Market, Vikalpa, July-September, Vol.11, No.3, pp. 225-229. Barua, S.K.. and V.Ralghunathan, 1987, Inefficiency and Speculation in the Indian Capital market, Vikalpa, July-September, Vol. 12, No.2, pp. 53-58. Barua, S.K and V. Raghunathan, 1988, Testing Stock Market Efficiency Using Risk Return Parity Rule: A Reply, Vikalpa, July- September, Vol. 13, No.3, pp. 82-83. Barua. S.K and V. Raghunathan. 1990, Soaring Stock Prices: Defying Fundamentals. Economic and Political Weekly, November 17, Vol. 25, No.46:pp. 2559-61.

Research Papers in Applied Finance Black, Fisher, 1972, Capital Market Equilibrium with Restricted Borrowing, Journal of Business, July. *** Black, Fisher, Michael e. Jensen, and Myron Scholes, 1972. The Capital Asset Pricing Model: Some Empirical Tests. in Studies in the Theory of Capital Markets, Ed. Michael e. Jensen, New York: Praeger Publishers, Inc. Dimson, Elroy, 1979, Risk Measurement when Shares are Subject to Infreque1lt Trading,Journal of Financial Economics, June, Vol.7, No.2, pp. 197-226. Fama, Eugene F. and James D. MacBeth, 1973, Risk, Return, and Equilibrium: Empirical Tests. Journal of Political Economy, May/June, Vol.71, No.3, pp. 607-36. Fama, Eugene F. and Kenneth R. French, 1988, Permanent and Temporary Components of Stock Prices, Journal of Political Economy, April, Vol. 96, No.2, pp. 24673. Fama, Eugene F. and Kenneth R. French, 1992, The Cross-Section of Expected Stock Retums, Journal of Finance, June, Vol. 47, No.2, pp. 427-465. Francis, Jack Clark and Frank J. Fabozzi, 1979, the Effects of Changing Macroeconomic Conditions on the Parameters of the Single Index Market Model, Journal of Financial and Quantitative Analysis, June, Vol. 14, No.2, pp. 351-360. Friend, Irwin, Randolph Westerfield and Michael Granito, 1978, New Evidence on the Capital Asset Pricing Model. The Journal of Finance, vol. 33, June, pp. 903-917. Gupta, L.C., 1981, Rates of Return on Equities, Oxford University Press. Jaffe, Jeffrey F., Randolph Westerfield and MA Christopher, 1989, a Twist on the Monday Effect in Stock Prices: Evidence from the U.S. and Foreign Stock Markets, Journal of Banking and Finance, September, Vol. /3, No.4/5, pp. 641-50. Lintner, John, 1965, Security Prices, Risk and Maximal Gains from Diversification Journal of Finance, December; Vol.20, pp. 79-96. Litzenberger, Robert H. and Krishna Ramaswamy, 1979, the Effect of Personal Taxes and Dividends on Capital Asset Prices: Theory and Empirical Evidence, Journal of Financial Economics, June, Vol.17, No.2, pp. 163-95. Maheshwari, G.e. and K.R. Vanjara, 1989, Risk Return Relationship: A Study of Selected Equity Shares, in Stock Market Efficiency and the Price Behaviour (the Indian Experience), Edited by O.? Gupta, First Eddition, Anmol Publication, New Delhi, pp. 335-52. Miller, Merton and Myron Scholes, 1972, Rates of Return in Relation to Risk: A Reexamination of some Recent Findings, in Studies in the Theory of Capital Markets. Edited by Michael Jensen. New York: Praeger. Markowitz, Hary, 1952, Portfolio Selection, Journal of Finance, Vol. 7, March, pp. 77-91. Mossin, Jan, 1966, Equilibrium in a Capital Asset market. Econometrica, vol. 34, October, pp. 768-783. Obaidullah, M., 1991 "The Distribution of Stock Returns-Chartered Financial Analyst,

Capital Asset Pricing Model: The Indian Context November 1991. Obaidullah, M., 1991a, Earnings, Stock Prices & Market Efficiency: Indian Evidence,Securities Industry Review, Journal of the Singapore Securities Research Institute, October. Obaidullah, M. 1991b, The Price-Earnings Ratio-Anomaly in Indian Stock Markets Decision, July-Sept. Obaidullah, M., 1993, Does the CAPM Explain Actual Price Behaviour, Chartered Financial Analyst, Nov. Obaidulllah, M., 1994, Indian StockMarket: Theories and Evidence, Hyderabad: ICFAI. Palaha, Sa tinder, 1991, Cost of Capital and Corporate Policy, Anmol Publications. Pandey, I.M. and Ramesh B.hat, 1989, Efficient Market Hypothesis: Understanding and Acceptance in India, in Stock Market Efficiency and the Price Behaviour (the Indian Experience), Edited by O.P. Gupta, First Edition, Anmol Publication, New Delhi, pp. 27993. Raghunathan. V, 1991. The Capital Asset Pricing Model, Investment Week, March 1824. Ramachandran. J., 1988, Behaviour of Stock Market Prices, Information Assimilation and Market Efficiency, unpublished FPM thesis, Indian Institute of Management Ahmedabad. Rao, N. Krishna and K. Mukherjee, 1971, Random Walk Hypothesis: An Empirical Study, Arthaniti, Vol. 14, Nos. 1 & 2, pp. 53-58. Reproduced in Stock Market Efficiency and the Price Behaviour (the Indian Experience), Edited by O.P. Gupta, First Edition, 1989, Anmol Publication, New Delhi, pp. 99-105. Rao, N. Krishna, 1988, Stock Market Efficiency: The Indian Experience, Paper presented at 'National Seminar on Indian Securities Markets: Thrusts and Challenges', Organized by the Department of Commerce, University of Delhi, South Campus, March 24-26,1988. Reproduced in Stock Market Efficiency and the Price Behavior (The Indian Experience), Editedby O.P. Gupta, First Edition, 1989, Anmol Publication, New Delhi, pp. 203-18. Ray, Subrata, 1994, "Capital Asset Pricing Model: The Indian Context", Unpublished Doctoral Dissertation, Indian Institute of Management, Bangalore. Roll, Richard, 1977, A Critique of the Asset Pricing Theory's Tests Part I: On Part and Potential Testability of the Theory, Journal of Financial Economics, March, Vol.4, No.2, pp.129-176. Roll, Richard, 1981, A Possible Explanation of the Small Firm Effect, Journal of Finance, September, Vol.36, No.4, pp. 879-88. Ross, Stephen A., 1978, The Current Status of the CapitalAsset Pricing Model (CAPM), Journal of Finance, June, Vol.33, No.3, pp. 885-901. Sears, R. Stephen and Gary L Trennepohl, 1993, Investment Management: Fort Worth The Dryden Press.

Research Papers in Applied Finance Sehgal Sanjay - 1994 "The Distribution of Stock Market Returns: Tests of Normality, Indian Journal of Finance and Research Vol. V No.2 July. Sharma J.L., 1983, Efficient Capital Markets and Random Character of Stock Price Behavior in a Developing Economy, Indian Economic Journal, October-December, Vol.31, No.2, pp. 53-70, Reproduced in Stock Market Efficiency and the Price Behaviour (the Indian Experience), Edited by O.? Gupta, First Edition, 1989, Anmol Publication, New Delhi, pp. 219-43. Sharma, J.L. and Robert E Kennedy, 1977, a Comparative Analysis of Stock Price Behavior on the Bombay, London, and New York Stock Exchanges, Journal of Financial and Quantitative Analysis, September, Vol. 12, No.3, pp. 391-413. Sharpe, Willian F., 1964, Capital Asset Prices: A Theory of Market Equilibrium UndeConditions of Risk, Journal of Finance, September, Vol. 19, No.3, pp. 425-42. Shashikant, Uma, 1988, A Study of the Rates of Return on Equity, Paper presented at 'National Seminar on Indian Securities markets: Thrusts and Challenges', Organized by the Department of Commerce, University of Delhi, South Campus, March 24-26, 1988. Reproduced in Stock Market Efficiency and the Price Behavior (the Indian Experience), Edited by O.P. Gupta, First Edition, 1989, Anmol Publication, New Delhi, pp. 317-34. Srinivasan, N. P and M.S. Narasimhan, 1988, Testing Stock Market Efficiency Using Risk-Return Parity Rule, Vikalpa, April-June, vol. 13, No.2, pp. 61-66. Vaidyanathan, R. and Subrata Ray, 1992, Estimation of Market Risk of Securities, Banking Finance, July. Vaidyanathan, R. and Kanti Kumar Gali, 1993, Stock Market Returns and "Vallan" (Settlement Period) Effect, Chartered Financial Analyst, December, vol. 8, no. 12, pp. 510. Vaidyanathan, R. and Kanti Kumar Gali, 1994a, Efficiency of the Indian Capital Market Indian Journal of Finance and Research Vol. V. No.2 July, 1994. Vaidyanathan, R. and Kanti Kumar Gali, 1994b, Market Indices, Working Paper of Center for Capital Market Research (CCMR) of llM, Bangalore. Varma, Jayant Rama, 1988, Asset Pricing under Parameter Non-stationarity, FPM Thesis, Indian Institute of Management, Ahmedabad. Vasal, V.K., 1988, Corporate Financial Decisions and Share Price Behavior in India, Paper presented at 'National Seminar on Indian Securities Markets: Thrusts and Challenges', Organized by the Department of Commerce, University of Delhi, South Campus, March 24-26, 1988. Reproduced in Stock Market Efficiency and the Price Behavior (the Indian Experience), Edited by O.? Gupta, First Edition, Anmol Publication, New Delhi, pp. 245-53. Yalawar, Y.B., 1985, Bombay Stock Exchange: Rates of Return and Efficiency, Paper Published by Indian Institute of Management, Bangalore.