Testing the Empirical Validity of CAPM in Shorter Periods Evidence from Indian Capital Market

DR D. LAZAR* and YASEER K.M

Abstract:
Capital asset pricing Model (CAPM) provides an equilibrium linear relationship between risk and expected return of an asset. The purpose of this study is to examine the risk return relationship with the CAPM frame work by Using Black, Jensen and Scholes (1972) Methodology. The study is conducted for seven sub periods comprising of three years each .The total period covers 9 years and used data from the year 01-01-2001 to 3112-2009, which includes the recession period. The study used the data of 70 companies which are the part of BSE100 and tested the validity of CAPM, test of SML, test of Non- linearity. Further the study compared the relationship between beta and portfolio return. The analysis gives mixed result and we couldnt find conclusive evidence in support of CAPM in the selected study periods.

Key words: CAPM, Intercept, Security Returns, Beta, Portfolio Returns, SML, Black Jensen and Scholes
Methodology (1972)

1. Introduction
Indian Capital Market has a long tradition and is one of the oldest in Asia and the history records back to nearly 200 years ago. The first stock exchange in India is the Bombay stock exchange which begins its operation in organized form from the year 1875. Indian capital market growing and so far it is one of the developed markets in the world. The growth of the capital market in India witnessed unique changes in the last two decades and there was an unprecedented growth, in terms of the number of companies listed, total market capitalisation, number of brokers and also in the number of participants. In India, Capital market is one of the most important parts of the financial system and the stock exchanges plays an important role in the economic growth and the growth in the stock market is symbolised as a barometer of economic growth. Currently there are 23 stock exchanges and one over the counter exchange operating in India. Out of which the National Stock Exchanges (NSE) and the Bombay Stock Exchange (BSE) are the biggest in number of companies listed and also in the market capitalization.

*Dr. Daniel Lazar, Pondicherry University, India., lazar.dani@gmail.com

The capital market is an interface, where investors can buy and sell stocks and other selected financial instruments. In India the growth of technology brought indefinite opportunities and opened an arena for investors particularly in the capital market. Today a cursory glance provides wildering collection of securities from the market and also a numerous number of financial instruments and the stock market becomes an investment avenue for FIIs, Institutional investors as well as individual investors. Today the investment arena is very complex and the capital market is over flooded with many financial instruments and large number of securities. Further the Indian market is highly volatile and selecting suitable securities became a quite complex exercise. The investors are risk averse and he expects additional return for the risk he bears. The market is highly complex, highly volatile and unpredictable and a simple mistake and lack of attention may lead to loss of money. The investor should be very alert and a suitable model which will help the investor to pick the best investment option or one which is helpful for analysing the various alternatives will be useful in decision making. Today investors and analysts are practicing different techniques and models to find out the best investment opportunity which will help reduce the risk and bag more return. CAPM is one of the important and widely used models for investment analysis. This model can be used to evaluate any investment project and it provides an equilibrium linear relationship between risk and expected return of an asset. Every investment is characterised by risk and return and the element of risk is always akin with every investment. The investment returns measure the financial results and may be historical or prospective. The return may be in the form various types of yield and capital appreciation, that the return is the sum of the benefits received (interest, dividend etc) while he own the asset and the change in value of the asset in the form of capital gain or loss, which is realised at the time of disposal of the asset. Risk is a mix of threat and opportunities which literally means the possibility of danger. The risk is defined as the potential for variability in return (Rao, Ramesh.K.S, 1989). Number of factors will affect the risk return relation and the various factors which are external that affects large number of securities simultaneously are known as systematic risk and is denoted as beta () .These types of risks are mostly uncontrollable and One can examine the individual stock return to the overall market return by
comparing how an individual company stock reacts to overall market fluctuation. CAPM is one of the models widely used throughout the world to explain the risk return relationship and the theory postulate that there is a linear relationship between beta and return. But the literature shows that the model gives different results for different market and thereby it is crucial to test the validity of this

before applying to the field concerned. This study is attempt to test the empirical validity of the one factor basic CAPM model in Indian Capital market by using three years data. This study is organized as follows. Section 2 deals with the review of previous empirical work, section 3 with methodology and the empirical work and results are explained in section4 and the section 5for summary and conclusion.

2. Previous research
The modern portfolio theory explains that there is a clear trade of between risk and return .The Markowitz portfolio selection model helps one to plot the efficient frontier of risky assets and provides a useful framework for selecting an optimal combination of risky funds. But this model however does not provide guidance with respect to the risk-return relationship for individual assets. The Capital asset pricing Model which was contributed by Jack Treynor(1961, un published), William Sharpe (1964), John Lintner (1965), and Jan Mosssin (1966), explains the equilibrium relationship between the expected return on risky assets .The model provide a mechanism to assess the role of a particular asset in the overall portfolio risk and return and it uses the result of capital market theory to derive the relationship between expected return for the risky assets. The literature showed that large number of studies has been conducted to test the applicability of the model in different markets and found different results for different markets. The empirical validity of this model was widely challenged in the late of Seventies, Eighties and Nineties by roll, Fama French etc. But at the same time there are number of studies which are in favor and supported the usage one factor model CAPM in developing and emerging markets. Literature showed that Sauer and Murphy (1992) are definite about the applicability of CAPM in describing risk return relation in the German Stock Market data. Similarly the studies conducted by Black and fisher (1993), Daniel and Titman(1997), Gyorgy Andor et.al (1999) for the Hungarian capital market.Ming-Hsiang Chen (2003) established that empirical performance of the CAPM is encouraging and the CAPM outperforms the CCAPM in terms of goodness of fit . Similarly Daniel Suh (2009)opined that in a highly volatile market Parameter estimates of the CAPM are generally superior to those of the Fama French three factor model At the same time the studies conducted by roll (1977), Harris and et al.(2003)argued against CAPM, Nopbhanon Homsud and et. al. (2009) found that Fama French model explain risk in

stock return better than the traditional one factor Capital Asset Pricing model. Yan Li and Liyan Yang( 2008) found that the conditional CAPM fails miserably to explain the size effect, the value effect, and the momentum effect. Pablo Rogersand et.al (2007) found that the results of their study propose and supported the explanatory power of Fama French model. Cudi Tuncer Gursoy and Gulnara Rejepova (2007): found that their test result weakens the validity of single index CAPM model in Turkey market over the analysis period. Further Grigoris Michailidis and et.al (2006): Xi Yang, Donghui Xu (2006), Medvedev A. (2004) Arduino Cagnetti (2001). Elsas Rand et.al (2000) etc tested CAPM and found

evidence against CAPM. Besides this Jan Bartholdy and Paula Peare (2004),Samit Maunder and Frank W. Bacon (2007) neither support nor reject the Capital Asset Pricing Model pIn Indian context, only few studies were conducted for analyzing risk return relationship in Indian capital market and studies by Varma (1988), Srinivasan (1988) have generally supported CAPM. The studies by Rao and Bhole (1990), Palaha(1991), Vaidyanadathan (1995), Sehgal (1997) Sehgal (2001,2003), Mohanthy (2006) questioned the validity of CAPM in Indian context. From the literature it is clear that there is a mixed opinion about the validity of one factor CAPM model. But the fact is that only few researches were conducted to test the applicability of the one factor capital asset pricing Model in Indian capital market by using daily data. Therefore the present study is proposed to test validity of the one factor Capital Asset Pricing Model by using daily data of 70 companies listed in BSE100. Index (2002) Mallikarjunappa and et.al

3. Scope of the study

The present study will test the suitability of the CAPM frame work in Indian context by using daily data of 70 companies. Since Indian capital market is one of the fastest developing markets in the world, it is very important to suggest how far the western portfolio theories are suitable to explain the differential return on financial assets in Indian capital market and also to suggest the suitability of the model which will help the investors, fund managers and the analysts. Further the study period also covers the recession period were we can see abnormal fluctuation in the stock market and the period comprises from 01-01-2008 to 31-03-2009.This will help the investors and fund managers to understand the applicability of this model in such situations.

3.1. Objectives of the study

The primary objective of the study is to test the empirical validity of the CAPM frame work in Indian context by using Black etal (1972) methodology. The main objectives of the study are described below. 1. To revisit the empirical validity of CAPM frame work in Indian capital market by using portfolios having different number of Securities. 2. To check whether higher or lower risk generate higher or lower rate of return. 3. To ascertain the relationship between return of securities and market return

4. To check whether expected rate of return is linearly related with systematic risk.

3.2. Source and Period of Data

In this study the test is organized to examine the suitability of CAPM models in Indian context by considering daily data of 70 companies which are the part of BSE 100 stock Index, a broad-based index, launched in 1989 as the base year 1983-84. The sample for the study covers nine years daily data for a period from 01-01-2001 to 31-12-2009 and the data used in this study were sourced from RBI , SEBI, BSE websites and Prowess- a data base of CMIE., (Center for Monitoring Indian Economy) a leading private sector economic research data provider in India. The study will consider 70 actively traded stocks listed in the BSE100 index including financial institutions. Brown and Warner (1985) suggest that the daily price are better for auto correlation in event methodology and is felt that quarterly , monthly and weekly data do not provide a very meaningful relationship between risk and return and hence daily data is used in this study. Further the study considers 91 day Treasury bill rate as the proxy for the risk free assets, which is available in weekly format in the Reserve Bank of India site. The 91 day Treasury bill is specifically chosen because it will better reflects the short term changes in the financial market and also a number of studies used the same

3.3. Research Method

Since the main objective of the study is to examine the suitability of CAPM in Indian context, it is proposed to collect nine years data of actively traded companies of BSE 100. The average percentage daily return of shares is put to use for the study to calculate the return and risk of the companies. Share prices returns in Prowess are calculated by considering all benefits accrued / losses incurred by the share holder by way of change in price on the

exchange, benefits received or losses incurred due to bonus issues, rights issues, and adhoc gains/losses. The return calculation also ensures that any split or consolidation event which happens has no effect on the return, except in the event of the prices changing due to market activity, return is calculated on closing prices. Here the data has been analyzed in two stages, in the first stage the daily percentage return has been used for all the scrip throughout the study period. In the second part of the analysis attempts are made to work out to test the applicability of Capital Asset Pricing Model in Indian context

3.4. Methodology for the basic Capital Asset Pricing Model

Black, Jensen and Scholes (1972) introduced a time series test of the CAPM and the relationship between risk and return has been analysed systematically. They carried out the study by using 1931-1965 data of all the NYSE stocks and they form portfolios and regressed them on beta. They had tested whether the relationship is linear and also whether any firmspecific volatility of a securitys return has an impact on the return of securities. Mallikarjunappa (2007), Valeed A Ansari (2000), in their studies in Indian capital market and Xi Yang (2006) Chinese stock market Grigoris Michailidis (2006) in Greek market etc. used the same methodology. The present study also follows a similar methodology followed by the Black and etal (1972).Further the study will also us Fama Macbeth (1973) methodology to test the Non-Linearity.

3.5 Testing CAPM with portfolios having 10 Securities

This study will test the CAPM model for the period from 2001 to 2009 and used the same method followed by the Black, Jenson and Scholes in (1972). This methodology use portfolio technique and also time series regression of excess portfolio return on excess market return and also cross sectional regression in risk premium form, which can be express by the equation below. The study will also use Fama and Macbeth methodology to test the non linearity .

For this, in the first step betas (also known as the systematic risk) of individual securities are measured and the beta coefficients of individual securities were calculated for the seven portfolio formation periods. A time series regression between the daily percentage return against the market return is used to get the beta coefficient of each security in the sample and the model is shown bellow.

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

----------------------

(1)

Where: Rit is the rate of return on asset i (or portfolio) at time t, Rft is the risk-free rate at time t, Rmt is the rate of return on the market portfolio at time t. i is the beta of stock i, eit

is the is the error term in the regression equation at time t. The equation can also expressed as

rit = i + i rmt + eit

Where: Rit Rft = r it and Rmt Rft = r mt r it is the excess return of stock i

----------------------

(1A)

r mt is the average risk premium and the i is the intercept The study will use the percentage daily return of security return on index (BSE 100) and the risk free return. The daily return of securities and the market for the period are regressed by taking the company return as dependent variables and the market return as the independent variable. In the second stage, the portfolios are constructed by using the calculated betas. For the formation of portfolios the individual beta for each stock is the arranged on ascending order and the stocks were grouped in to portfolios having 10 stocks each according to their beta value .The first portfolio comprises the first 10 securities with the lowest beta, the next portfolio with the next 10 securities. The same method is followed for the formation of other portfolios and there by the last portfolio is formed with the securities having the highest beta. In this stage the portfolio betas are calculated by using the following regression model.

rpt = p + p rmt + ept

Where

----------------------

(2)

rpt is the average excess portfolio return on time t, p is the estimated portfolio beta, and e pt is the error term in the regression equation at time t. In the third step in order to estimate the ex post security market line for each testing period the portfolio return are regressed against portfolio betas. The model for the calculation is

rp = 0 + 1 p + e p
Where

----------------------

(3)

rp = is the average excess return of the portfolio P

p is the beta of the portfolio P, and

ep is the error term in the regression equation The theory says that if the CAPM is true 0 should be equal to zero and the slope SML, 1 is the average risk premium of the market portfolio. Further the study will also test the non- linearity between the total portfolio return and betas by using the following equation.

rp = 0 + 1p + 2p + ep
related with each other and 2 will be equal to zero.

-------------------

(4)

Here the theory says that if the CAPM is true, the portfolio returns and its betas are linearly

3.6 The Statistical test of the CAPM

The t -Test Further the validity of the CAPM is statistically tested by using the t- test at different levels of significance, say- 99%, 95%. (This study will not consider 90% confidence level for interpretation even though some of the coefficients are significant at 90% level) the The t test has been introduced by W.S Gosset and the distribution of the ratio t has been derived for normally distributed population. It is most commonly applied when the test statistic would follow a normal distribution and the analysis is commonly used to compare and evaluate the difference in means between two groups. Theoretically, the t-test can be used even if the sample sizes are very small and as long as the variables are normally distributed within each group and the variation of scores in the two groups is not reliably different.

3.7 Why Black, Jenson and Sholes Methodology

Miller and Scholes (1972) diagnosed that while using individual stock betas, there is problem because of the betas are measured with error and the measurement error in right hand variable biases down regression coefficients. Fama and MacBeth (1973), Black, Jenson and Scholes (1972) addressed this problem by grouping stocks in to portfolios. Portfolio betas are better measured because the portfolio has lower residual variance. Further the individual betas vary over the time as the size, leverage and risk of the business change. Secondly the individual stock return is so volatile that you cannot reject the hypothesis that all average returns are the same (Asset pricing: John H. Cochrane, Princeton University Press, USA, P- 434-435).There by the present study planned to use this methodology.

3.8Limitation of the study

The size of the sample and the number of companies used to construct the portfolio is one of the important limitations. Only seven portfolios are formed and tested in the present study and this may affect the statistical result and may be biased in limited observations. The literature says that the CAPM tests realized in international scope use more than 30 years of observations and the market portfolio plays an important role in the test results. But the present study used 9 years data and conducted tested with return of only one index.

4. Testing CAPM in Different study Periods

In order to dress up the question of the validity of the CAPM in Indian context ,the test is conducted by dividing the entire nine year period in to seven different sub periods comprising three years each and the sub periods includes 2001-2003, 2002-2004, 2003-2005, 2005-2007, 2006-2008 and 2007-2009. The outline of the study is summarised in the Table 1 below. Table .1 Table Showing the Different Portfolio Formation Periods and Testing Periods

Period Period Range Portfolio Formation Testing period

1 01-03 2001 2003

2 02-04 2002 2004

3 03-05 2003 2005

4 04-06 2004 2006

5 05-07 2005 2007

6 06-08 2006 2008

7 07-09 2007 2009

In the first step of the empirical testing the systematic risk (beta), also known as the un diversifiable risk which unanimously affects the prices of all securities in the market is measured. The beta coefficient shows the risk associated with a security or portfolio and as per the theory the investor should be bothered only about the systematic risk which cannot be diversified away. The basic CAPM theory clearly argues that the efficient market is expected to compensate only the systematic risk which is denoted by beta (). In the first step, the beta coefficients of individual securities are calculated for the seven sub periods. A time series regression model (1) is run between the daily percentage return against the percentage market return is used to get the beta coefficient of each security in the sample Rit - Rft = i + i ( Rmt- Rft) + eit -------(1)

4.1 Period Wise Distribution of Beta

The betas for individual securities by using the above model were calculated for different study periods. The result shows that the range of estimated beta for the sub period 1 is in between 0.15575 minimum and the maximum 2.0056.The range of beta for the sub period 2 is in between 0.16846 and 1.74350 and for the sub period 3, the beta lies between 0.166861 minimum and 1.68426 maximum .The range of beta for the period 4 shows that the minimum beta is 0.27854 and the maximum is 1.62242 and the beta for the period 5 lies between 0.30657 minimum and 1.57201 maximum . For the sub period 6 the minimum beta is 0.28916 and the maximum 1.61827 and for the seventh sub periods the range of beta is in between 0.04611 and 1.66231. Here we can see variation in the range of beta in different Study periods.

4.2 Average Excess Portfolio Return and Beta

Different studies shows that combining securities in to portfolios will definitely helps to diversify the risks due to the firm specific factors and will enhance the precision of estimates of beta and the expected return on the portfolios. At this stage of the study, the portfolios are constructed by using the calculated betas. The same procedure is repeated for the whole sample period, for the adjusted period and also for different sub periods. The average excess return was calculated for each portfolio and the following regression model (2) is used to calculate the portfolio beta. rpt = p + p rmt + ept ------------(2)

On the basis of the regression results the CAPM is tested for different sub periods.

4.3. Testing CAPM in the First Sub Period (2001-2003)

In the first sub period the analysis is carried out on the data of 70 companies listed BSE 100 and covers the period 1st Jan - 2001 to 31st Dec 2003. For the first sub period, the study used 753 daily observations and the test is repeated with the same test procedures used for the whole and adjusted periods. For this sub period it is also noted that, BSE 100 index was (2023.82) in the beginning and it was (3074.87) at the end of the study period the total gain in the index was (1051.05) points during this period.

4.4 Testing CAPM through Portfolios

For the sub period 1st- Jan-2001 to 31stDec-2003, the test considers 753 observations and the results are shown in the Table 2 below. From the table, it is clear that portfolio 1 (P1) with lowest beta earned the minimum return of (0.135846) while the portfolio 6 (P6) with the highest beta (1.08355) receives the maximum return (0.21964).But the portfolio seven with higher beta bagged nearly half of the return than the portfolio 6 and hence the argument of CAPM that higher risk beta is associated with higher rate of return is violated. Out of the seven portfolios, both the beta and the return shows an increasing trend up to the portfolio 6 ,but in portfolio seven, the return (0.128447) is decreasing while the beta (1.57857) shows an increase from (1.08355). The R2 value for the first six portfolios lies between (0.27150) and (0.596), which indicates less than adequate correlation with the market index. But in portfolio 7, R2 value is (0.76875), which indicates that above 76 per cent of the variation in the scrip has been explained by the relationship with the index. Further from the Table 2, it is noted that the all the constant except portfolio 7 are statistically significant and also have positive values. That means the first six constants are statistically significant and the alpha coefficient is significantly different from zero, there by reject the null hypothesis. Further the estimated betas of portfolios are found to be statistically significant at 99% level; thereby we reject the null hypothesis that the portfolio beta is not a significant determinant of portfolio return. Thus from the analysis it clear that the can be used for predicting risk return relationship in Indian stock market for the sub period 2001-2003

Table 2 Table Showing Average Excess Portfolio Return and Portfolio Betas for the Sub period (2001 -2003) (N = 753)

Port folio P1 P2 P3 P4 P5 P6 P7

Portfolio Return (rp) 0.135846 0.198832 0.143898 0.182184 0.216427 0.219646 0.128447

Constant

Beta

Standard Error 0.82728 1.09264 0.94657 1.62623 1.14613 1.29599 1.25793

R2

F value

P Value of beta at 99% 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

0.11887 0.17091 0.10840 0.14148 0.17020 0.16675 0.05138

0.34760 0.57192 0.72707 0.83370 0.94681 1.08355 1.57857

0.27150 0.36644 0.55466 0.55408 0.59027 0.59606 0.76875

279.895 434.367 935.384 933.163 1081.92 1108.22 2496.59

Avg Rf

0.01681

Average rm = (Rm-Rf)

0.04881

The value of constants of P1, P2, P3, P4, P5, P6, are significant at 99 % level but P7 is insignificant

4.5. Estimation of Security Market Line (2001 - 2003)

The result for the first sub period is shown in the Table 3.and it is clear that the t-test rejects the null hypothesis that 0 is not significantly different from zero. Here the calculated value of the intercept is (0.17757) and it is significantly different from zero. Statistically, the result shows that the t- value is greater than (2.57) at 95% confidence level and hence the o is statistically inconsistent with CAPM. Further from table it is clear that 1 is negative (0.0029) and it is nearly equal to zero and the Absolute t- value is less than (2.57), this means that 1 is not significantly different from zero. But as per CAPM the 1 should be greater than zero, there by the result is inconsistent with the CAPM hypothesis and the CAPM is rejected during this period.

Table 3. Table showing the result of the test of SML for Sub period (2001 - 2003) Coefficients 0 1 0.17757 0.00291 Std error 0.04133 0.04380 t- value 4.296*** 0.066 p-value 0.0077 0.9495

*** shows significant at 99% level

Critical value for t test with 5-Degrees of freedom at 95 % level (2.57)

4.6. Test of Non-Linearity (2001-2003)

The result of the test of non-linearity for the sub period 1 is summarised bellow in the Table 4.The result shows that the intercept (0.03810) of the model is not significantly different from zero. Statistically the t- value is (0.5678), which is less than (2.7765) at 5% significant level and thereby it is consistent with the argument of CAPM.
Table 4 Table showing the result of the test of Non-Linearity for the Sub period (2001 - 2003) Coefficients Std error t- value p-value 0 1 2 0.03810 0.33520 0.17365 0.06711 0.14884 0.07466 0.5678 2.252 2.326 0.6005 0.0874 0.0806

Critical value of t test for 4-Degrees of freedom at 95% (2.7765)

In the case of 1, the t- value is (2.252), which is less than (2.7765), and it is significantly not different from zero. As per the CAPM, the 1 should be equal to the average risk premium; hence we can conclude that result is inconsistent with the CAPM hypothesis. In the case of 2, the value of coefficient is (0.17365) and the absolute t- value is less than (2.7765) at 5% significance level, 2 is consistent with the CAPM hypothesis. Thus we can say that the betas are linearly related with return and hence CAPM is can be accepted during the first sub period but still the data showed weakness to fully explain the model.

4.7. Section IV: CAPM in Second Sub Period (2002-2004)

The data used in the second sub period consists of 759 daily observations of a sample of 70 companies listed in BSE 100. The second period covers the period from 01-01-2002 to 3112-2004. Further, in the beginning of this sub period BSE 100 index was (1557.22) points and it was (3580.34) at the end. The total gain in the index was (2023.12) points during this

period, which shows that, there is an increase of 972 points in the index when compared with the previous period. The various test results for the period is described below.

4.8. Testing CAPM through Portfolios

As per the Capital asset pricing theory, the higher risk beta is associated with higher rate of return. But from the Table 5 it is clear that the portfolio 2 (P2) with lowest beta earned the minimum return of (0.132295) and the portfolio 6(P6) earned more return than the other portfolios. During the study period all the portfolios including the portfolio with lowest beta earned more return than the average excess market return and also the risk free return. Further the positive constants suggest that the portfolios have earned higher returns than the CAPM has predicted Table 5 Table Showing Average Excess Portfolio Return and Portfolio Betas for the Sub period (2002-2004) (N =759)
Port folio P1 P2 P3 P4 P5 P6 P7 Avg Rf Portfolio Return(rp) 0.18812 0.13299 0.27030 0.20948 0.23339 0.27087 0.27020 0.014238 Constant 0.14555 0.06508 0.18199 0.10713 0.11769 0.14154 0.11722 Average (Rm-Rf) Beta 0.40544 0.64687 0.84115 0.97486 1.10212 1.23187 1.45715 rm = Standard Error 0.77423 0.67298 0.85477 1.01318 0.78913 0.93917 1.17759 0.10498 R2 0.34023 0.63469 0.64552 0.63516 0.78577 0.76390 0.74222 F value 390.375 1315.22 1378.53 1317.89 2776.67 2449.39 2179.65 P Value of beta at 99% 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

The values of constants of P1, P2, P3, P4, P5, P6, P7are significant at 99 % level

In the case of portfolio1, the value of R2 is (0.34023), and in all other case the R2 value is in between (0.63) and (0.78) which indicates that above 63 to78 % of the variation in the scrip has been explained by the relationship with the index. If we look further in to the results of the test for alpha and the slope coefficients of portfolios, the result shows that the constant (alpha) values are significantly different from zero, and thereby the null hypothesis is rejected. Further the p value of slope coefficient are greater than the level of significance in all the cases and thereby we reject the null hypothesis that beta does not significantly explain the variation in portfolio return. Thus the conclusion from this analysis is that beta can explain the portfolio return as suggested by CAPM during the second sub period.

4.9. Estimation of Security Market Line (2002-2004)

The Table 6 describes the result for the sub period 2 which compose the values of 0 and 1.From the table it is clear that the t- test rejects the null hypothesis that 0 is not significantly different from zero. Here the calculated value of the intercept is (0.12526) and is significantly different from zero. Statistically, the result shows that the t- value is greater than (2.57) at 95 % confidence level and the o is significant. It means that the result is statistically inconsistent with CAPM. Table 6 Table showing the estimation of SML for Second Sub Period (2002 - 2004)
Coefficients 0 1 0.12526 0.10488 Std error 0.04608 0.04577 t- value 2.718 ** 2.292 p-value 0.0419 0.0705

** Shows significant at 95% level.

Critical value for t test with 5-Degrees of freedom at 95 % level (2.57)

Further the value of the 1 is (0.10488) and the t- value is less than (2.57) this means that 1 is not significantly different from zero .As per CAPM the 1 should be greater than zero, there by the result is inconsistent with the CAPM hypothesis and the data shows its weakness to fully explain the CAPM during this sub period.

4.10. Test of Non-Linearity (2002-2004)

The result of the non-linearity test for the sub period 2 is summarised bellow in the table 7.The result shows that the value of the intercept is (0.12757) and is not significantly different from zero. Statistically the t- value is (1.015) and is less than the table value, hence we accept the null hypothesis that 0 is not significantly different from zero. Thus it is consistent with the argument of CAPM. Table 7
Table showing the result of the test of Non-Linearity for the Sub period (2002 - 2004)
Coefficients 0 1 2 0.12757 0.09918 0.00307 Std error 0.12565 0.28795 0.15287 t- value 1 .015 0.3444 0.0201 p-value 0.3674 0.7479 0.9849

Critical value for t-test with 4-Degrees of freedom at 95% level (2.7765)

In the case of 1, the t- value is (0.3444) which is less than (2.7765), and it is not significantly different from zero. Hence we can conclude that result is inconsistent with the CAPM

hypothesis because the value 1 should be equal to the average risk premium. In the case of 2, the value is (0.00307) and the t- value is less than (2.7765), at 5% significance level, we can say that it is consistent with the CAPM hypothesis. From the analysis it is clear that the value of the 2 is not significantly different from zero .Thus we cannot clearly reject the CAPM is during the second sub period.

4.11. Section V: CAPM in Third Sub Period (2003-2005)

The study investigated the applicability of CAPM and the data used in this study consisted 70 stocks listed in the BSE 100 Index over the period 01-01-2003 to 31-12-2005. For the third sub period, the study used 755 daily observations and the test is repeated with the same methodology and test procedures used for the whole period and adjusted period. For this sub period, it is also noted that BSE 100 index was (1664.67) in the beginning and it was (4953.28) at the end .The total gain in the index during this period was (3288.61) points.

4.12. Testing CAPM through Portfolios

For the third sub period, the test considers 755 observations over the period 01-01-2003 to 3112-2005. From the Table 8, it is clear that beta of the portfolio increases from portfolio 1 to portfolio 7, but we cannot see such trend in portfolio return. Beta of the portfolio 3 (P3) and portfolio 6 (P6) earns less when compared to portfolio two and portfolio 4 & 5(P4 & P5) which shows that the result contradicts the CAPM. The R2 value for the first portfolios is (0.39166), which indicates less than adequate correlation with the market index. But in portfolio 2 to 7, R2 value is in between (0.64385) and 0.84625) which indicates that 64 to 84 per cent of the variation in the scrip has been explained by the relationship with the index. If we further look in to the Table 8, it is noted that the values of constants are significant at different level (P6 & P7 at 90% level and all others at 99%level) and also have positive values, which suggests that, we reject the null hypothesis that the alpha is not significantly different from zero. Further all the estimated betas of portfolios are found to be statistically significant at the 99% level, and we reject the null hypothesis that the portfolio beta is not a significant determinant of portfolio return. Thus from the analysis it is clear that the can predict the risk return relationship in the Indian market during the sub period 2003-2005

Table 8
Portfolio P1 P2 P3 P4 P5 P6 P7 Portfolio Return (rp) 0.19576 0.23213 0.19582 0.24712 0.24341 0.23194 0.27509 Constant 0.13121 0.13598 0.07629 0.11524 0.09436 0.05893 0.06909 Beta 0.46072 0.69667 0.84995 0.94406 1.06785 1.24122 1.47422 Standard Error 0.78158 0.70528 0.69089 0.82561 0.61953 0.97754 1.00364 0.13860 R2 0.39166 0.64385 0.73712 0.70782 0.84625 0.74919 0.79990 F value 487.370 1368.55 2122.70 1833.88 4166.87 2261.24 3026.13 P Value of beta at 99% 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

Avg Rf

0.01366

Average m = (Rm-Rf)

The values of constants of P1, P2, P3, P4, P5 are significant at 99 % level; the constants of P6 and P7 are significant at 90 %Level

Table Showing Average Excess Portfolio Return and Portfolio Betas for the Sub period (2003 -2005) (N = 755)

4.13. Estimation of Security Market Line (2003-2005)

From the Table 9, it is clear that the t-test rejects the null hypothesis that 0 is not significantly different from zero. Here the calculated value of the intercept is (0.16826) and it is significantly different from zero. Table 9
Table showing the result of the test of SML for the Sub period (2003 - 2005)
Coefficients 0 1 0.16826 0.06585 Std error 0.02357 0.02329 t- value 7.138 *** 2.826 ** p-value 0.0008 0.0368

** Shows significant at 95% level. *** Shows significant at 99% level.

Critical value for t test with 5-Degrees of freedom at 95 % level (2.57)

Statistically, the result shows that the t- value is greater than (2.57) at 95% confidence level and the p value is significant at 99% level, hence the result do not support the CAPM. Further looking in to the table it is clear that the slope (1) is significantly different from zero. Here the t- value is greater than (2.57) at 95% confidence level. As per the CAPM, 1 should be equal to the average risk premium, which should be greater than zero and it is concluded that the result is consistent with the CAPM. Hence the CAPM is accepted for the third sub period by rejecting Ho that 0 = 0

4.14. Test of Non-Linearity (2003-2005)

The results of the estimated values for the test of non - linearity is summarised in the Table 10. The result shows that the intercept is (0.18392) and 0 is not significantly different from zero. Statistically the t- value is (2.605), which is less than (2.7765) at 5% significant level and there by the null hypothesis is accepted and is consistent with the CAPM hypothesis. Table 10
Coefficients 0 1 2 0.18392 0.02979 0.01858 Std error 0.07060 0.15312 0.07777 t -value 2.605 0.1946 0.2389 p-value 0.0597 0.8552 0.8229

Table showing the test of Non-Linearity for the Sub period (2003 - 2005)
Critical value for t-test with 4-Degrees of freedom at 95% level (2.7765)

In the case of 1, the t- value is (0.02979) is smaller than (2.7765), and it is not significantly different from zero. As per the CAPM, the 1 should be equal to the average risk premium. Hence we can conclude that result is inconsistent with the CAPM hypothesis. The value 2 is (0.01858) and the t- value is less than (2.7765), at 5% significance level that means it is not significantly different from zero. Hence, we can say that it is consistent with the CAPM hypothesis. Hence the CAPM is accepted but the data shows weakness to fully explain the postulates of CAPM.

4.15. Section VI CAPM in Fourth Sub Period (2004 -2006)

This sub period considered the daily data for the period from 01-01-2004 to 31-12-2006 and the dataset consists of 755 daily observations of 70 companies which have been the part of BSE 100 index. In the beginning of the test period the BSE 100 index was (3074.87) points and it was (6982.56) at the end. The total gain in the index was (3907.69) points during this period,

4.16. Testing CAPM through portfolios

For the fourth sub period 755 observations are used and the result shows that portfolio 1 (p1) with lowest beta (0.56299) received maximum return when compared to the other portfolios especially P5, P6, and P7 which are having beta values above one. Here all the portfolios

including the portfolio 2, with lowest beta bags more return than the average excess market return and also the risk free return. Table 11 Table Showing Average Excess Portfolio Return and Portfolio Betas for Sub Period (2004 2006) (N = 755)
Port folio Portfolio Return (rp) 0.19367 0.13098 0.18110 0.16977 0.13713 0.17221 0.17639 Standard Error F value P Value of beta at 99%

Constant 0.13865 0.04788 0.09071 0.07077 0.02389 0.04885 0.02408

Beta

R2

P1 P2 P3 P4 P5 P6 P7 Avg Rf

0.56299 0.81057 0.89768 0.97381 1.1060 1.20218 1.48129

0.78490 0.64327 0.69737 0.69060 0.79109 0.77701 0.95771 0.10505

0.53150 0.77784 0.78512 0.81428 0.81169 0.84072 0.84064

854.283 2636.53 2751.36 3301.59 3245.86 3974.74 3972.28

0.0000 0.0003 0.0000 0.0000 0.0000 0.0000 0.0000

0.01496

Average rm = (Rm-Rf)

The constants of P1, P3, P4, are significant at 99 % level; P2 at 95 % Level and P6 at 90% significant level. P5,P7 are insignificant

The CAPM explains that, higher risk is associated with higher rate of return and the result of the study does not find any support for this argument because ,we cannot find any positive correlation (-0.09531) between beta and the average portfolio excess return. Here all the portfolios including the portfolio 2, with lowest beta received more return than the average excess market return and also the risk free return. In the case of portfolio 1, the R2 value is (0.53150), which indicates less than adequate correlation with the market index. But for other portfolios, the R2 value is above (0.777) to (0.957), which indicates that above 77 % to 95% of the variation in the scrip has been explained by the relationship with the index. All the values of the constants except p5 and p7 are statistically significant and all are positive. It indicates that, the alpha coefficients are significantly different from zero and hence we reject the null hypothesis that the intercept is not significantly different from zero. Further the positive constants suggest that the portfolios earned higher returns than the CAPM has predicted. All the p values of estimated betas are found to be statistically significant at the 99% level; thereby we reject the null hypothesis that the portfolio beta is not a significant determinant of portfolio return. Thus from the analysis

we can say that the is a predictor of return for the Indian market during the sub period 2004-2006.

4.17. Estimation of Security Market Line (2004-2006)

The estimated result of the SML for the sub period 4 is shown in the Table 12 below. From this it is clear that the t-test rejects the null hypothesis that 0 is not significantly different from zero. Here the calculated value of the intercept is (0.17341) and it is significantly different from zero. Statistically, the result shows that the t- value is greater than (2.57) at 95% confidence level; hence the o is statistically inconsistent with CAPM. Table 12
Table showing the result of the test of SML for the Sub period (2004- 2006)
Coefficients 0 1 0.17341 0.00748 Std error 0.03639 0.03494 t- value 4.765 *** 0.2141 p-value 0.0050 0.8389

*** Shows significance at 99% level. Critical value for t test with 5-Degrees of freedom at 95 % level (2.57)

Further from table it is clear that 1 is negative (0.00748) and it is nearly equal to zero and the absolute t- value is less than (2.57), this means that 1 is not significantly different from zero. But as per CAPM the 1 should be greater than zero, there by the result is inconsistent with the CAPM hypothesis and the model is fully rejected during the sub period.

4.18.Test of Non-Linearity (2004-2006)

While testing the non-linearity, as per the CAPM the 0 and 2 will be equal to zero and the 1 should be equal to the average risk premium. The results of the estimated values are summarised bellow in the Table 13. Table 13
Table showing the result of the test of Non-Linearity for the Sub period (2004- 2006)
Coefficients 0 1 2 0.29847 0.26997 0.12792 Std error 0.10790 0.21703 0.10452 t -value 2.766 1.244 1.224 p-value 0.0505 0.2815 0.2881

Critical value for t-test with 4-Degrees of freedom at 95% level (2.7765)

The result shows that the intercept of the model is greater than the risk free interest rate and the constant 0 is significantly different from zero. Statistically the t - value is (2.766), which is greater than (2.7765) at 5% significant level and there by the null hypothesis is rejected and hence inconsistent with the argument of CAPM. The absolute t- value for the 1 is (1.244) which is less than (2.7765) and the value is not significantly different from zero. As per the CAPM, 1 should be equal to the average risk premium; hence we can conclude that result is inconsistent with the CAPM hypothesis. The t- value of 2 (1.224) is less than (2.7765) and hence value is not significantly different from zero, which is consistent with the CAPM. Thus the CAPM couldnt clearly be rejected during the sub period.

4.19. Section VII: CAPM in the Fifth Sub Period (2005 -2007)
This sub period considered the daily data for the period from 01-01-2005 to 31-12-2007 and the dataset consists of 750 daily observations of 70 companies which have been the part of BSE 100 index. It is also noted that in the beginning of this study period the BSE 100 index was at (3580.34) points and at the end of the study period it is (11154.28) resulting a total gain of (7573.94) points throughout the period

4.20. Testing CAPM through Portfolios

The study in the sub period 5 used 750 observations and the data covers the period from 101-2005 to 31-03-2007.The estimates of the study is reported in the table 14 below. The table reveals that all the constants are positive. During this period the portfolios bags higher rate of return when compared with the other study periods. Further The CAPM postulates that higher risk beta is associated with higher rate of return and the result of the study partially support this argument because we can see high positive correlation between beta and average excess return on portfolios. Further it is also interesting to note that all the beta values are in between (1.49240) and (1.52382).Out of the seven portfolios, the beta shows an increasing trend, P7 with high beta (1.5203) earned more return than others and the R2 explains that 76.17% of the variation in the scrip has been explained by the relationship with the index. In the case of portfolio 1, the R2 value is (0.8002), which indicates that adequate correlation with the market index. If we further look in to the Table 14, it is noted that the constants of P1 and P2 are of statistically insignificant but all others are significant at 95 and 90% level. Further the positive constants suggest that the portfolios have earned higher returns than the CAPM has

predicted. All the p values of estimated betas are found to be statistically significant at 99% level; thereby we reject the null hypothesis that the portfolio beta is not significant determinant of portfolio return. Thus from the analysis we can say that beta can predict the risk return relation in the Indian capital market during the sub period 2005-2007.

Table 14 Table Showing Average Excess Portfolio Return and Portfolio Betas for the Sub Period (2005 2007) (N = 750)

Portf olio P1 P2 P3 P4 P5 P6 P7 Avg Rf

Portfolio Return (rp) 0.27919 0.28746 0.30661 0.31259 0.30497 0.30914 0.31722 0.01724

Constant 0.06130 0.06720 0.08589 0.09113 0.08421 0.08654 0.09722 Average (Rm-Rf)

Beta 1.49240 1.50999 1.51668 1.52038 1.52228 1.52323 1.52382 rm =

Standard Error 1.07047 1.13101 1.16175 1.22087 1.17057 1.15407 1.18991 0.14487

R2 0.8002 0.7860 0.7784 0.7617 0.7771 0.7822 0.7717

F value 2997.27 2748.69 2628.29 2391.53 2608.02 2686.45 2529.03

P Value of beta at 99% 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

The value of constants P3, P4, P6 and P7 are significant 95 % level and P5, at 90% significant level. P1 ,P2 are insignificant

4.21. Estimation of Security Market Line (2005-2007)

The result for the fifth sub period is shown in the Table 15 and it is clear that the t-test rejects the null hypothesis that 0 is not significantly different from zero. Here the value of the intercept is (1.41536) and it is significantly different from zero. Statistically, the result shows that the t- value is greater than (2.57) at 95% confidence level and hence the o is statistically inconsistent with CAPM.

Table 15
Table showing the result of the test of SML for the Sub period (2005 - 2007)
Coefficients 0 1 1.4153 1.13347 Std error 0.32306 0.21316 t- value 4.381 *** 5.317 *** p-value 0.0071 0.0031

*** Shows significance at 99% level. Critical value for t test with 5-Degrees of freedom at 95 % level (2.57)

Further from table it is clear that 1 is (1.13347) and it is significantly different from zero. Here the t- value is greater than (2.57) which means that it is consistent with CAPM hypothesis

4.22.Test of Non-Linearity (2005-2007)

The results of the estimated values for the test of non - linearity are summarised in the table 16. The result shows that the intercept (60.2641) of the model is 0 is significantly different from zero. Statistically the t- value is (1.055), which is less than (2.7765) at 5% significant level and thereby we cannot reject the null hypothesis. Thus it is consistent with the CAPM hypothesis. In the case of 1, the absolute t- value is (1.065) is smaller than (2.7765), since it is not significantly different from zero. As per the CAPM, the 1 should be equal to the average risk premium; hence we can conclude that result is inconsistent with the CAPM hypothesis. Table 16 Table showing the test of Non-Linearity for the Sub period (2005- 2007)
Coefficients 0 1 2 60.2641 80.6609 27.1154 Std error 57.1167 75.7428 25.1092 t - value 1.055 1.065 1.080 p-value 0.3509 0.3469 0.3410

Critical value for t-test with 4-Degrees of freedom at 95% level (2.7765)

The value 2 is (27.1154) and the t- value is less than (2.7765) at 5% significance level that means it is not significantly different from zero. Hence, we can say that it is consistent with the CAPM hypothesis. Hence, the relationship is linear but the data is weak to explain the CAPM during the study period.

4.23. Section VIII: CAPM in the Sixth Sub Period (2006 -2008)
The data used in the seventh sub period consists of 745 daily observations of a sample of 70 companies listed in BSE 100. This sub period covers the data from 01-01-2006 to 31-122008. It is also noted that in the beginning of this study period the BSE 100 index was at (4953.28) points and at the end of the study period it is (4988.04) resulting a total gain of (34.76) points. Further it is noted that the period includes a part of the recession period.

4.24. Testing CAPM through Portfolios

The table 17 deals with the test results for the constant alpha and the beta coefficient of the portfolio for the sub period 01-01-2006 to 31-12-2008.In the case of portfolio return the portfolio 4 earns the least return (-0.01330) and the value of the constant is also negative. Further the value of the R2 shows high correlation between the market return and the portfolio return .For all the portfolios the value of R2 is in between (0.593) and (0.905), which indicates that adequate correlation with the market index. ie is 59% to 90% of the variation in the scrip has been explained by the relationship with the index. The table shows that most of the constants are insignificant and thereby we cannot reject the null hypothesis. Table 17 Table Showing Average Excess Portfolio Return and Portfolio Betas for the Sub period (2006-2008)

Port folio P1 P2 P3 P4 P5 P6 P7

Portfolio Return (rp) 0.05494 0.00965 0.03012 -0.01330 0.05991 0.06231 0.12303

Constant 0.05317 0.00700 0.02690 0.01678 0.05603 0.05787 0.11767

Beta 0.47559 0.71177 0.86554 0.93899 1.04399 1.19443 1.44127

Standard Error 0.82658 0.80501 0.86240 0.77934 0.70999 0.98340 1.15324 0.00372

R2 0.59363 0.77526 0.81634 0.86496 0.90513 0.86683 0.87329

F value

P Value of beta at 99% 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

1085.42 2563.14 3302.58 4759.44 7088.83 4836.66 5120.82

Avg Rf

0.01939

Average

rm = (Rm-Rf)

The value of constants of P7 is significant at 99% level P5 at 95 % Level and P1 at 90% significant level. P2,P3,P4,P6 are insignificant

Further the positive constants suggest that the portfolios have earned higher returns than the CAPM has predicted. All the p values of estimated betas are found to be statistically significant at the 99% level; thereby we reject the null hypothesis that the portfolio beta is not significant determinant of portfolio return. Thus the analysis do not gives a firm result in support of CAPM.

4.25. Estimation of Security Market Line (2006-2008)

From the Table 18, we can see that, the value of the intercept is (0.03215) statistically; the result shows that the absolute t- value is less than (2.57) at 95% confidence level, and the o is not significantly different from zero. Thus the result is consistent with the CAPM. As per CAPM 1 should be equal to the average risk premium and here the t- value is (1.657), which is less than the table value is not significantly different from zero and should be greater than zero. Hence it is concluded that the result is inconsistent with the CAPM and hence there is mixed result and we dont have conclusive evidence in support of CAPM in the sixth sub period. Table 18
Table showing the result of the test of SML for the Sub period (2006- 2008)
Coefficients 0 1 0.03215 0.08270 Std error 0.04977 0.04992 t- value 0.6461 1.657 p-value 0.5467 0.1585

Critical value for t test with 5-Degrees of freedom at 95 % level (2.57)

4.26. Test of Non-Linearity (2006-2008)

The result for the sub period 6 is summarised bellow in the Table 19. The result shows that the intercept (0.20379) of the model 0 is significantly different from zero. Statistically the t value is (2.393), which is less than (2.7765) at 5% significant level and thereby we cannot reject the null hypothesis. Thus it supports the argument of CAPM. In the case of 1, the absolute t- value is (2.485) which is less than (2.7765) and it is not significantly different from zero. As per the CAPM, the 1 should be equal to the average risk premium and hence we can conclude that result is inconsistent with the CAPM hypothesis.

Table 19
Table showing the result for the test of Non-Linearity for the Sub period (2006- 2008)
Coefficients 0 1 2 0.20379 0.45958 0.28262 Std error 0.08517 0.18494 0.09500 t- value 2.393 2.485 2.975** p-value 0.0750 0.0678 0.0410

** Shows significant at 95% level. Critical value for t-test with 4-Degrees of freedom at 95% level (2.7765)

In the case of 2, the value is (0.28262) and the t- value is greater than (2.7765) at 5% significance level, we can say that it is inconsistent with the CAPM hypothesis. From the analysis it is clear that the value of the 2 is significantly different from zero .Thus we cannot say that the betas are linearly related with each other and hence CAPM is rejected during the sixth sub period.

4.27. Section IX: CAPM in the Seventh Sub Period (2007 -2009)
In the seventh sub period the analysis is carried out on the data of 70 companies listed BSE 100 and covers the period from 01-01-2007 to 31-12-2009, the study used 738 daily observations and the test is repeated with the same test procedures used for other test period. In the beginning of the test period the BSE 100 index was (6982.56) points and it was (9229.71) at the end. The total gain in the index was (2247.15) points during this study period,

4.28. Testing CAPM through Portfolios

The study in the sub period 7 used 738 observations and the data covers the period from 101-2007 to 31-03-2009.Further the study period includes the period which is excluded for defining the adjusted period due to the high fluctuation in the capital market. It will be interesting to check the result during this period and the various estimates are reported in the Table 20 below. The table reveals that all the constants are positive and all the portfolios except the portfolio 2 (P2) bags higher return than the average excess market return

Table 20 Table Showing Average Excess Portfolio Return and Portfolio Betas for Sub Period (2007 2009) (N = 738)
Port folio P1 P2 P3 P4 P5 P6 P7 Avg Rf Portfolio Return (rp) 0.08502 0.04054 0.08185 0.08081 0.11927 0.16448 0.12786 0.01702 Standard Error 0.82383 0.92283 0.93873 0.94318 0.88599 1.07297 1.20925 0.04611 F value 860.651 2035.23 2927.60 3723.10 5719.99 5010.55 5624.87 P Value of beta at 99% 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

Constant 0.06672 0.00929 0.04320 0.03690 0.06948 0.10719 0.05648 Average (Rm-Rf)

Beta 0.393851 0.67844 0.82771 0.93784 1.04698 1.23770 1.47794 rm =

R2 0.53903 0.73441 0.79910 0.83494 0.84949 0.87192 0.88429

The values of constants of P6 are significant at 99% level P1, P5 at 95 % significance level .P2, P3, P4, P7are insignificant.

The CAPM postulates that higher risk beta is associated with higher rate of return but from the result we cannot see any upward trend in the portfolio return. In the case of portfolio 1, the value of R2 is greater than (0.73441) and in the case portfolio7 it (0.88429) the maximum, which shows that adequate correlation with the market index. If we further look in to the Table 20, it is noted that most of the constants are insignificant and al1 are positive, which suggests that we cannot reject the null hypothesis that the alpha is not significantly different from zero. Further the estimated betas of portfolios are found to be statistically significant at the 99% level; thereby we reject the null hypothesis that the portfolio beta is not a significant determinant of portfolio return. Thus the analysis gives a mixed result and we cannot clearly accept the CAPM for this sub period 2007-2009 and apart from other period it may be due to the recession effect.

4.29. Estimation of Security Market Line (2007-2009)

The estimated result of the SML for the sub period 7 is shown in the Table 21 below. The table shows that the t-test accept the null hypothesis that 0 is not significantly different from zero. Here the calculated value of the intercept is (0.02385) and it is not significantly

different from zero. Statistically, the result shows that the t- value is less than (2.57) at 95% confidence level and hence it is insignificant, thus consistent with CAPM.

Table 21 Table showing the estimation of SML for the Sub Period (2007- 2009)
Coefficients 0 1 0.02385 0.08072 Std error 0.03514 0.03516 t-value 0.6789 2.296 p-value 0.5274 0.0701

Critical value for t test with 5-Degrees of freedom at 95 % level (2.57)

Further from table it is clear that 1 is (0.08072) and the t- value is greater than (2.57). Hence, 1 is significantly different from zero. As per CAPM the 1 should be greater than zero, there by the result is inconsistent with the CAPM hypothesis. Thus the CAPM is rejected in the seventh sub period.

4.30. Test of Non Linearity (2007-2009)

Test for the non-linearity is used to check whether there exists non- linearity between portfolio return with beta. As per theory, if CAPM holds true 0 and 2 will be equal to zero and the 1 will be equal to the average risk premium. Table 22
Table showing the result of the test of Non-Linearity for the Sub period (2007- 2009)
Coefficients 0 1 2 0.06469 0.01859 0.05287 Std error 0.08969 0.20112 0.10512 t- value 0.7213 0.0924 0.5030 p-value 0.5106 0.9308 0.6414

Critical value for t-test with 4-Degrees of freedom at 95% level (2.7765)

The results of the estimated values for the sub period 7 are summarised bellow in the Table 22. The result shows that the intercept (0.06469) of the model is not significantly different from zero. Statistically the t- value is (0.7213), which is less than (2.7765) at 5% significant level and there by the null hypothesis is accepted and is consistent with the argument of CAPM.The absolute t- value for the 1, is (0.0924) which is less than (2.7765), and the value of intercept is not significantly different from zero. As per the CAPM, the 1 should be equal to the average risk premium; hence we can infer that result is inconsistent with the CAPM

hypothesis. The t- value of 2 is (0.503), which is less than (2.7765) and the value is not significantly different from zero. Thus we can say that beta is linearly related with return. Hence, we cannot fully reject CAPM during this sub period.

5. Summary and Conclusion

This study examined the empirical validity of CAPM, which was questioned in home security market as well as throughout the world markets. The present study used daily return of 70 securities listed in BSE 100 index. The CAPM is tested for different study period through different methods by using portfolios having 10 securities. The results of the different tests for different study periods are summarized below in Table 23.-29. The Form the table, following conclusion can be derived. 1. The test for portfolios based on percentage return with equally weighted portfolios having 10 securities mostly in support of CAPM but do not give a conclusive evidence in favor of CAPM 2. For the sub periods, the test gives mixed result and in some period the test clearly rejects the CAPM hypothesis and in few sub periods it partially supports the CAPM hypothesis. 3. In almost all the cases the constant have positive values, which suggest that the portfolio bagged more return than the CAPM has predicted. 4. In analyzing the risk - return relationship, for most of the cases the R2 shows a high value over .65 (approximate), which shows that above 65% of the variation, has been explained by the relationship with index. 5. From the analysis, it is found that, generally higher beta provides higher return to the investor , in most of the case beta explain the variation in portfolio returns.( it does not mean it is fully true in 100% cases) 6. Test for SML and Non linearity support CAPM but do not give conclusive evidence in favor of CAPM in different sub periods.

I. Through Portfolios
Table 23 Table Showing Consolidated Results for Different Study Periods by Using 10 Securities
Port folio Constant Sub period 1 (20001-2003) F Value R2 P value Beta

P1 P2 P3 P4 P5 P6 P7

0.1189 0.1709 0.1084 0.1415 0.1702 0.1668 0.0514

279.90 434.37 935.38 933.16 1081.92 1108.22 2496.59

0.2715 0.3664 0.5547 0.5541 0.5903 0.5961 0.7688

0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

Table Showing Consolidated Results for Different Study Periods by Using 10 Securities
Table 24
Sub period 2 (2002-2004) Port folio Constant F Value R
2

Sub period 3( 2003-2005) P value Beta Constant F Value R

2

Sub period 4 (2004-2006) P value Beta Constant F Value

R
2

P value Beta

P1 P2 P3 P4 P5 P6 P7

0.1456 0.0651 0.1820 0.1071 0.1177 0.1415 0.1172

390.38 1315.22 1378.53 1317.89 2776.67 2449.39 2179.65

0.3402 0.6347 0.6455 0.6352 0.7858 0.7639 0.7422

0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

0.1456 0.0651 0.1820 0.1071 0.1177 0.1415 0.1172

390.38 1315.22 1378.53 1317.89 2776.67 2449.39 2179.65

0.3402 0.6347 0.6455 0.6352 0.7858 0.7639 0.7422

0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

0.13865 0.04788 0.09071 0.07077 0.02389 0.04885 0.02408

854.28 2636.53 2751.36 3301.59 3245.86 3974.74 3972.28

0.5315 0.7778 0.7851 0.8143 0.8117 0.8407 0.8406

0.0000 0.0003 0.0000 0.0000 0.0000 0.0000 0.0000

I. Through Portfolios
Table 25
Sub period 5 (2005-2007) Port folio Constant F Value R2 P value Beta Constant Sub period 6 (2006-2008) F Value R2 P value Beta Constant Sub period 7 (2007-2009) F Value R2 P value Beta

P1 P2 P3 P4 P5 P6 P7

0.0613 0.0672 0.0859 0.0911 0.0842 0.0865 0.0972

2997.27 2748.69 2628.29 2391.53 2608.02 2686.45 2529.03

0.8002 0.7860 0.7784 0.7617 0.7771 0.7822 0.7717

2

0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

0.0532 0.0070 0.0269 0.0168 0.0560 0.0579 0.1177

1085.42 2563.14 3302.58 4759.44 7088.83 4836.66 5120.82

0.5936 0.7753 0.8163 0.8650 0.9051 0.8668 0.8733

0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

0.0667 0.0093 0.0432 0.0369 0.0695 0.1072 0.0565

860.65 2035.23 2927.60 3723.10 5719.99 5010.55 5624.87

0.5390 0.7344 0.7991 0.8349 0.8495 0.8719 0.8843

0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

Note: The Values of Constant, F, and R are adjusted to 4 digits.

Table Showing Consolidated Results for Different Study Periods by Using 10 Securities
II. Test of Security Market Line
Table 26
Sub Period1(2001-2003) Coefficients
Constant t- value P value Constant

Sub Period 2 (2002-2004)

tvalue P value

Sub Period 3( 2003-2005)

Constant t- value P value

Sub Period 4 (2004-2006)

Constant t- value P value

0.1776 0.0029

4.2960 *** 0.066

0.0077 0.9495

0.12526 0.10488

2.718 ** 2.292

0.0419 0.0705

0.16826 0.06585

7.138 *** 2.826**

0.0008 0.0368

0.1734 0.00748

4.765 *** 0.2141

0.0050 0.8389

*** Significant at 99 %level ** Significant at 95% level

Table 27
Sub Period 5 (2005-2007) Coefficients Constant 1.4153 1.13347 t- value 4.381*** 5.317*** P value 0.0071 0.0031 Sub Period 6 (2006-2008) Constant 0.03215 0.08270 t- value 0.6461 1.657 P value 0.5467 0.1585 Sub Period 7 (2007-2009) Constant 0.02385 0.08072 t- value 0.6789 2.296 P value 0.5274 0.0701

*** Significant at 99 %level

III. Test of Non Linearity

Table Showing Consolidated Results for Different Study Periods by Using 10 Securities
Table 28
Sub Period 1(2001-2003) Coefficient Constant 0.03810 0.33520 0.1736 t- value 0.5678 2.252 2.326 P value 0.6005 0.0874 0.0806 Constant 0.12757 0.0991 0.0030 t- value 1.0150 0.3444 0.0201 P value 0.3674 0.7479 0.9849 Constant 0.1839 0.0298 0.0186 t- value 2.6050 0.1946 0.2389 P value 0.0597 0.8552 0.8229 Sub Period 2(2002-2004) Sub Period 3(2003-2005)

III. Test of Non Linearity

Table 29
Sub Period 4(2004-2006) Coefficient Constant 0.29847 0.26997 0.12792 t- value 2.766 1.244 1.224 P value 0.0505 0.2815 0.2881 Sub Period 5 (2005-2007) Constant 60.2641 80.6609 27.1154 t- value 1.055 1.065 1.080 P value 0.3509 0.3469 0.3410 Sub Period 6 (2006-2008) Constant 0.2037 0.4595 0.2826 t- value 2.393 2.485 2.975** P value 0.0750 0.0678 0.0410 Sub Period 7(2007-2009) Constant 0.0647 0.0185 0.0529 t- value 0.7213 0.0924 0.5030 P value 0.5106 0.9308 0.6414

The sub period 6 & 7 which covers the recession period generally in Support of CAPM but in the sub period 1 and 4 the test of non linearity shows that beta is not linearly related with expected return. The findings of the study shows that, the test in the Indian market by using 70 securities listed in the BSE 100 index is mostly supportive in different test periods to the hypothesis of Capital Asset Pricing Model, which says that higher beta provides higher return to the investor and the study reveals that while using percentage return and portfolios with equal weight, in most of the case beta explain the variation in portfolio returns. Regarding the security market line, The CAPM predicts that 0 (the intercept) should be equal to zero and the 1 (the slope of SML) should be equal to the average risk premium. The result for the SML for the whole period support the CAPM but for the adjusted period the 0 is inconsistent with CAPM and thereby we cannot say that CAPM is fully accepted for the adjusted period. The result for the different sub periods by using portfolios with 10 securities mostly rejected CAPM. Five out of Seven test results clearly reject the CAPM hypothesis while two partially support CAPM hypothesis. From the above result, we cannot give conclusive evidence in favor of CAPM.

The test for non- linearity between beta and stock return is tested by including beta square coefficient. As per CAPM the portfolio return and its betas are linearly related with each other when the 0 and 2 is equal to zero. The test for the non- linearity tells that, for the whole and adjusted period the result is in support of the CAPM hypothesis. For the adjusted period we cannot give conclusive evidence in support of the CAPM hypothesis, but the model supports the non linearity of the CAPM factors in most of the cases, which explains the beta estimates. Further the high value of the estimated correlation coefficient between the intercept and the slope indicates that the model explains excess returns. However in most of the case, the intercept have value near to zero, weakens above explanation.

In short most of the test result supports the CAPM and is in favor of the model but we cannot see conclusive evidence in support of CAPM to wrap up the question of the validity of CAPM in Indian context

Reference
1. Andor Gyorgy, Ormos Mihly and SzAbo Balazs, 1999, Empirical Testing of Capital Asset Pricing Model(CAPM) In the Hungarian Capital Market,Periodica Polytechnica, Ser. Soc. Man. Scl.Nol.7, Vol. 1, PP. 4761. 2. Ansari, Valeed A (2000). Capital Asset Pricing Model: Should We Stop Using It , Vikalpa, Vol.25(1), PP-55-64. 3. Bark, Hee-Kyung K (1991). Risk, Return, and Equilibrium in the Emerging Markets: Evidence from the Korean Stock Market, Journal of Economics and Business, Vol 43(4), pp-353-362. 4. Bartholdy Jan, Peare, Paula, 2004, Estimation Of Expected Return: CAPM Vs Fama And French, Working Paper Series no 176, Centre For Analytical Finance, University of Aarhus. 5. Bello, Zakri.Y, 2008, A Statistical Comparison of the CAPM to the Fama-French Three Factor Model And the Caharts Model, Global journal of Finance and Banking Issues, Vol.2, pp14-24. 6. Black, F., Jensen, M. C. and Scholes, M. (1972), The Capital asset pricing model: Some empirical tests, Studies in the Theory of Capital Markets, New York: Praeger. pp-79-121. 7. Black, Fischer (1993). Beta and Return, Journal of Portfolio Management, Vol. 20(1), 8-18. 8. Brav, Alon, Lehavy, Reuven, Michaely, Roni, 2005 Using Expectations to Test Asset Pricing Models, Financial Management, Vol. 34, No. 3, PP. 31-64. 9. Cagnetti Arduino,2001 ,Capital Asset Pricing Model and Arbitrage Pricing Theory in the Italian Stock Market: an Empirical Study, http://www.ssrn.com/. 10. Chan, Cheong ,Yue , 1997 Multivariate testing of the capital asset pricing model in the Hong Kong stock market, Applied Financial Economics, Vol.7: 3, PP.311 - 316. 11. Chen, Ming-Hsiang, 2003, Risk and return: CAPM and CCAPM, The Quarterly Review of Economics and Finance, Vol.43, PP.369393. 12. Cheng, Tsung-Chi, Neng, Hung, Tsai Lai,Pei-Fen,2005, On the Two-Stage Estimation of Fama-French Three Factor Models: Evidence from Taiwan, http://www.airitilibrary.com/. 13. Cochrane, John H. (2001) Asset Pricing. New Jersey: Princeton University Press. 14. Connor, Gregory and Sehgal,Sanjay, 2001, Tests of the Fama and French Model in India, http://www.ifa.com/. 15. Daniel, Kent and Titman, Sheridan (1997). Evidence on the Characteristics of Cross Sectional Variation in Stock Returns, Journal of Finance, Vol.52(1), pp.1-33.

16. Fama, Eugene F and French, Kenneth R (1992). The Cross-section of Expected Stock Returns, Journal of Finance, Vol.47 (2), pp.427-465. 17. Fama, Eugene F and French, Kenneth R (2004). The CapitalAsset Pricing Model: Theory and Evidence, Journal of Economic Perspectives, Vol.18 (3), pp. 25-46. 18. Fan, Stephen C, , 2004,Have we misinterpreted CAPM for 40 years? a theoretical proof, Working Paper Series, http://www.ssrn.com/. 19. Fletcher Jonathan, Kihanda, Joseph, 2005, An examination of alternative CAPMbased models in UK stock returns, Journal of banking and finance, Vol.29, pp.29953014. 20. Glabadanidis, Paskalis, 2009, A Dynamic Asset Pricing Model with Time Varying Factor and Idiosyncratic Risk, Journal of Financial Econometrics, Vol. May 6, pp.1-18. 21. Grauer, Robert R., Janmaat, Johannus A. b , 2010, Cross-sectional tests of the CAPM and FamaFrench three-factor model, Journal of Banking & Finance, Vol.34 ,pp. 457470 22. Gursoy Cudi Tuncer, Rejepova , Gulnara , 2007, Test of Capital Asset Pricing Model In Turkey, Dogus Universitesi Dergisi, Vol.8 (1), pp.47-58 , 23. Hamao, Yasushi , 1988,An Empirical Examination Of The Arbitrage Pricing Theory Using Japanese Data, Japan And The World Economy,Vol 1, pp.45-61. 24. Harris, Robert S; Marston, Felicia C; Mishra, Dev R and OBrien, Thomas J (2003). Ex Ante Cost of Equity Estimates of S&P 500 Firms: The Choice between Global and Domestic CAPM, Financial Management, Vol. 32(3), pp51-66. 25. Homsud, Nopbhanon , Wasunsakul, Jatuphon , Phuangnark, Sirina, and Joongpong , Jitwatthana, 2009, A Study of Fama and French Three Factors Model and Capital Asset Pricing Model in the Stock Exchange of Thailand, International Research Journal of Finance and Economics ,Issue 25. 26. Hsiang, Ming Chen, 2003, Risk and return: CAPM and CCAPM The Quarterly Review of Economics and Finance, Vol. 43, pp.369393. 27. Jagannathan, R., Wang, Z., 1996, The conditional CAPM and the cross-section of expected returns, Journal of Finance, Vol.51, pp.353. 28. Javid, Attiya Y. 2009, Test of higher moment capital asset pricing model in case of pakistani equity markrt, European journal of Economics, Finance and Administrative Sciences.Vol(15) pp.144-162. 29. Koch Stefan and Westheide Christian, The Conditional Relation between FamaFrench Betas and Return, G August 4, 2008 30. Kothari, S P; Shanken, J and Sloan, R P (1995). Another Look at the Cross -section of Expected Stock Returns, Journal of Finance, Vol. 50(1), PP185-224.

31. Lewellen, Jonathan, Nagel, Stefan, 2006, The conditional CAPM does not explain asset-pricing anomalies, Journal of Financial Economics ,Vol.82, pp.289314, 32. Li, Yan and Yang, Liyan, 2009, Under conditioning and Over conditioning: Testing the Conditional CAPM and the Conditional Fama-French Three-Factor Model ttp://www.ssrn.com/, 33. MacKinlay, A. Craig, 1995, Multifactor models do not explain deviations from the CAPM, USA Journal of Financial Economics, Vol.38, pp. 3-28. 34. MacKinlay, A. Craig, 1987,Mackinlay, C.A. 1995, Multifactor models do not explain deviations from the CAPM, Journal of Financial Economics, vol. 38, pp. 3 28 36. Madhusoodanan, T P (1997). Risk and Return: A New Look at the Indian Stock Market, Finance India, Vol.11 (2), pp.285-304. 37. Majumdar, Samit, Bacon, Frank .W, 2007, Multi Factor Pricing Model: An Alternative Approach to CAPM, ASBBS E-Journal, Volume 3, No. 1. 38. Malin Mirela , Veeraraghavan, Madhu, 2004, On the Robustness of the Fama and French Multifactor Model: Evidence from France, Germany, and the United Kingdom International Journal of Business and Economics, , Vol. 3, No. 2, pp.155176 . 39. Manjunatha, T; Mallikarjunappa, and Begum, Mustiary(2007). Capital Asset Pricing Model: Beta and Size Tests, AIMS International Journal of Management, Vol1 (1), pp.71-87. 40. Michailidis, G, Tsopoglou, S., Papanastasiou, D. and Mariola, E. (2006)Testing the Capital Asset Pricing Model (CAPM): The Case of the Emerging Greek Securities Market International Research Journal of Finance and Economics, Vol.4, PP.78-91. 41. Mohanty, Pitabas (2002). Evidence of Size Effect on Indian Stock Returns, Vikalpa, 27(2), pp.27-37. 42. Pettengill, G., Sundaram, S., & Mathur, I. (1995). The conditional relation between beta and returns, Journal of Financial and Quantitative Analysis, Vol.30, pp.101116. 43. Rahman, Moztafizur,Baten,Azizul, Alam,Ashraf-Ul, 2006, An Empirical Testing Of Capital Asset Pricing Model In Bangladesh, Journal of Applied Sciences Vol.6(3),pp.662-667. 44. Reinganum, Marc R., 1981, A New Empirical Perspective on the CAPM, Journal of Financial and Quantitative Analysis, Vol. XVI, No.4. 45. Roll, R. (1977). A critique of the asset pricing theorys tests, Journal of Financial Economics, Vol.4, PP.129-176.

46. S.Shijin, Kumar, Arun G. and Sanghamitra B,2007, Relationship between Size, Value and Market Risk: Some Evidence, International Journal of Investment Management and Financial Innovations,p47. Samit Majumdar ,Frank W. Bacon, 2007 , Multi Factor Pricing Model:An Alternative Approach To CAPM , ASBBS E-Journal, Volume 3, No. 1. 48. Sauer, A. and A. Murphy, 1992, An Empirical Comparison of Alternative Models of Capital Asset Pricing in Germany, Journal of Banking and Finance, PP.16-47 49. Srinivasan, S (1988). Testing of Capital Asset Pricing Model in Indian Environment, Decision, Vol.15(1), pp.51-59. 50. Suh, Daniel The Correlations and Volatilities of Stock Returns: The CAPM beta and the Fama French Factors, http :// ssrn.com/1364567 51. Terregrossa, S J (2001). Robust Informational Tests on the CAPM, Applied Economics, Vol.8(2),pp. 121-124. 52. Wong,Kie Ann and Meng Siong Lye, 1990, Market Values, Earnings Yields And Stock Returns Evidence From Singapore,Journal of Banking and Finance, Vol.14 (2-3) :3 ,pp.I l-326. 53. Yalwar, Y B (1988). Bombay Stock Exchange: Rates of Return and Efficiency, Indian Economics Journal, Vol.35(4), pp.68-121. 54. Yang, Xi and Donghui Xu . 2006. Testing the CAPM modelA Study of the Chinese Stock Market, Master Thesis Essay. Sweden: UMEA School of Business.
55. Yonezawa ,Yasuhiro Hin, Tio Kia, 1992,An empirical test of the CAPM on the

stocks listed on the Tokyo Stock Exchange,Japan and the World Economy,vol.4 , pp.145-161,

Books &websites
1. 2. 3. 4. Asset Pricing,Princton Universty Press, John.H. Cochrane,P-434,435 Economic Outlook for 2010-11 Fundamentals of Financial Management, Rao, Ramesh.K.S, p-389, 1989 Investments a Global Perspective, jack C. Francis and Roger Ibbotson, prentice Hall , New Jersey 5. Asset Pricing, John.H. Cochrane, Princton University Press,P-434,435. 6. Investments, William.F.Sharpe and et.al, prentice hall India Pvt. Limited, 1996. 7. Investments, ZVI Bodie and et.al.Tata McGraw hill publishing company, Newdeli, 2008. 8. Rao, Ramesh.k.s Fundamentals of Financial Management, page 389, 1989sey,2002. 9. S.Kevin, Security analysis and portfolio Management, 2008. 10. Security analysis and portfolio Management, S.Kevin, PHI L Ltd, New Delhi, 2008. 11. Statistics-Handbook of Statistics on Indian Economy. 12. http://benmcclure.mp/. 13. http://bseindia.com/about/abindices/bse30.asp# reconstitution. 14. http://www.crisil.com/.

15. http://www.essays.se/about/empirical+test+of+CAPM/. 16. http://www.imf.org. 17. http://www.rbi.org.in/. 18. http://www.sebi.gov.in/. 19. http://www.wikipedia.org. 20. https://cdbmsi.reservebank.org.in/cdbmsi/servlet/login/. 21. https://sebi.ac.in/ monthly bulletins.