Comp.fundamental And Equity Returns Indian Markets

Submission Cover

st

21 Australasian Finance and Banking Conference 1. Title

Company Fundamentals and Equity Returns in India

2. Primary Author Dr. Vanita Tripathi

3. Co-Authors (separate with comma)

4. Prizes Select the prizes for which you would like to be considered (you may pick more than one). (For more information about prizes please see the conference web site: www.banking.unsw.edu.au/afbc) Prize Barclay's Global Investors Australia Prize

BankScope Prize

Yes/No Yes No

Sirca Research Prize Australian Securities Exchange Prize

No No

5. Journals Select the journals for which you would like to be considered (you may pick more than one). Journal Journal of Banking and Finance Journal of Financial Stability

Yes/No Yes No

6. Conference Proceedings

Would you like your paper (if accepted) to be published by World Scientific Publishing Co Ltd as a review volume compiling selected papers?

1 Electronic copy available at: http://ssrn.com/abstract=1247717

Yes/No Yes

Company Fundamentals and Equity Returns in India

Author: Dr. Vanita Tripathi Senior Lecturer Department of commerce Delhi School of Economics University of Delhi Delhi-110007

Telephone: 91-011-27667891 9213269951 Email: [email protected] Date of submission: May 13,2008 JEL Classification Number: G12, G14

Keywords: Size Effect, Value effect, P/E effect, Leverage effect, CAPM, Asset pricing.

Acknowledgements: The author is thankful to Indian Council of Social Science Research (ICSSR) New Delhi for providing financial support to carry out this study Abstract

This paper examines the relationship between four company fundamental variables (viz. market capitalization, book equity to market equity ratio, price earnings ratio and debt equity ratio) and equity returns in Indian stock market using monthly price data of a 2 Electronic copy available at: http://ssrn.com/abstract=1247717

sample of 455 companies forming part of S&P CNX 500 Index over the period June 1997 to June 2007. We also investigate whether the inclusion of any one or more of these fundamental variables can better explain cross sectional variations in equity returns in India than the single factor CAPM. We find that market capitalization and price earnings ratio have statistically significant negative relationship with equity returns while book equity to market equity ratio and debt equity ratio have statistically significant positive relationship with equity returns in India. The investment strategies based on these variables produced extra risk adjusted returns over the study period. Using Davis Fama and French(2000) methodology we find that Fama-French three factor model ( based on market risk premium, size premium and value premium) explains cross sectional variations in equity returns in India in a much better way than the single factor CAPM. These results have important implications for market efficiency, asset pricing and market microstructure issues in Indian stock market.

.

3 Electronic copy available at: http://ssrn.com/abstract=1247717

Company Fundamentals and Equity Returns in India

Abstract

This paper examines the relationship between four company fundamental variables (viz. market capitalization, book equity to market equity ratio, price earnings ratio and debt equity ratio) and equity returns in Indian stock market using monthly price data of a sample of 455 companies forming part of S&P CNX 500 Index over the period June 1997 to June 2007. We also investigate whether the inclusion of any one or more of these fundamental variables can better explain cross sectional variations in equity returns in India than the single factor CAPM. We find that market capitalization and price earnings ratio have statistically significant negative relationship with equity returns while book equity to market equity ratio and debt equity ratio have statistically significant positive relationship with equity returns in India. The investment strategies based on these variables produced extra risk adjusted returns over the study period. Using Davis Fama and French(2000) methodology we find that Fama-French three factor model ( based on market risk premium, size premium and value premium) explains cross sectional variations in equity returns in India in a much better way than the single factor CAPM. These results have important implications for market efficiency, asset pricing and market microstructure issues in Indian stock market.

4

Company Fundamentals and Equity Returns in India

I INTRODUCTION In an emerging stock market like India, investment analysts and market participants are continuously in search for investment strategies that can outperform the market. Efficient Market Hypothesis (EMH) rules out the possibility by anybody to consistently earn extra normal return in an efficient stock market. According to this hypothesis securities are correctly priced and return is solely determined by the amount of risk one assumes (as per the standard Capital asset Pricing Model – CAPM). However a plethora of empirical studies doubts such a phenomenon and documents the availability of extra normal returns by using investment strategies based on firm specific variables such as size (Banz (1981)), leverage (Bhandari (1988)), price earnings ratio (Basu (1977)), book equity to market equity ratio (Stattman (1980), Rosenberg, Reid and Lanstein (1985)) etc. These empirical evidences have been commonly cited as anomalies to CAPM based on company fundamentals and popularly known as the size effect (small capitalization stocks outperform large capitalization-stocks), leverage effect (high debt-equity stocks outperform low debt-equity stocks), Price Earnings Effect (low P/E stocks outperform high P/E stocks) and value effect (high book equity to market equity stocks outperform low book to market equity stocks). Two schools of thought have emerged in search for possible explanation of persistent departure from the standard CAPM. One argument is that CAPM is mis specified; there is/are some missing risk factor(s) which beta fails to capture. Hence there is a move towards multifactor asset pricing framework as specified by Fama and French (1996). The other school blames the investors' irrationality for the existence of the phenomenon. Whatever be the cause, the presence of these CAPM anomalies provide gainful investment opportunities to the investing community. The robustness of size and value effects in US stock market motivated Fama and French (1992, 1993, 1996) to suggest the inclusion of a size and value factor in asset pricing model. A number of research studies have explored the economic feasibility of investment strategies based on fundamental 5

variables, but most of these studies relate to US and other mature markets. Similar evidence for emerging markets including India is limited and more recent in origin. As a result of financial sector reforms initiated since early 1990s the Indian stock market has witnessed metamorphic changes as regards to the size, structure and turnover. With more than 4700 listed companies, 20 millions shareowners and a market capitalization of Rs.30,257,720 million (in 2005-06) developments in Indian stock markets are now comparable to those in other mature markets. Hence there is a felt need for a study which can examine the relationship between various company fundamentals and equity returns in Indian stock market in this changed regime and test for the economic feasibility of fundamentals based investment strategies in the advent of technological up gradation. The results of the study are of pertinent use by investment analysts, mutual fund managers as well as marginal investors in devising fundamentals based investment strategies to earn extra-normal returns in Indian stock market. II

RESEARCH OBJECTIVES

The primary objectives of the study are :


to examine the relationship between four company fundamentals (size, leverage, P/E ratio and Book to market equity ratio) and equity returns in India.




to test whether the investment strategies based on these company fundamentals yield any extra risk adjusted returns in Indian stock market.




to check whether the inclusion of any or more of these fundamental variables can better explain cross sectional variations in average equity returns in India. In other words whether a multifactor model can better explain cross-sectional variations in equity returns in India or not.

The study also attempts to examine the following research issues :


Whether arbitrage opportunities are available in Indian stock market.




Whether company fundamentals can explain variation in average stock return in a better way than market factor in Indian context.

III

RESEARCH HYPOTHESES

Following hypotheses have been tested in the study – 6

I.

Regarding company fundamentals and equity returns in India. (i)

There is a statistically significant relationship between various company fundamentals and equity returns in India.

(ii)

Stocks of small companies outperform the stocks of large companies in Indian stock market.

(iii)

Low P/E stocks outperform the stocks of high P/E stocks.

(iv)

High BE/ME stocks outperform low BE/ME stocks.

(v)

Stocks of companies with high D/E ratio outperform the stocks of low D/E ratio companies.

(vi)

The investment strategy based on company size yields extra normal returns in Indian stock market.

(vii)

The investment strategy based on P/E ratio of companies yields extra normal returns in Indian stock market.

(viii) The investment strategy based on BE/ME ratio of companies yields extra normal returns in Indian stock market. (ix)

The investment strategy based on D/E ratio of companies yields extra normal returns in Indian equity market.

II.

Regarding cross sectional variations in equity returns in Indian stock market. (x)

Company size can better explain cross sectional variations in equity returns in Indian stock market than market factor.

(xi)

P/E ratio can better explain cross sectional variations in equity returns in Indian stock market than market factor.

(xii)

BE/ME ratio can better explain cross sectional variations in equity returns in Indian stock market than market factor.

(xiii) D/E ratio can better explain cross sectional variations in equity returns in Indian stock market than market factor.

7

(xiv)

A two factor model can better explain cross sectional variations in equity returns in Indian stock market than the single factor CAPM.

(xv)

A three factor model can better explain cross sectional variations in equity returns in India than single factor CAPM or two factor model.

(xvi)

A four factor model can better explain cross sectional variations in equity returns in Indian stock market than any of the single factor, two or three factors model.

(xvii) A five factor model (based on excess market return, size premium, P/E risk premium value premium and leverage risk premium) can better explain cross sectional variations in equity returns in Indian stock market than any of the single factor, two factors, three factors or four factors model. IV

DATA AND THEIR SOURCES

The data comprises of monthly closing adjusted share prices of 455 listed companies/stocks in India (as included in S&P CNX 500 index) over the most recent 10 years period June 1997 to June 2007 (See Annexure I for List of Sample Companies). The monthly price data have then been converted into monthly return data using the following equation : R it =

Pit − Pi ( t −1) Pi ( t −1)

for i = 1 to 455

for t = 1 to 120

(1)

where Rit = Return on stock i in the month t Pit = Closing adjusted share price of stock i in month t Pi(t-1) = Closing adjusted share price of stock i in month t-1. This gives us a monthly return series of 120 observations for every stock (or company). Monthly return on market portfolio (proxied by S&P CNX Nifty) have also been calculated using equation (1) except that in place of closing adjusted share prices we have used closing Index Values. 8

Rate of returns on 91-days Treasury Bills has been used as a proxy for risk free return and S&P CNX NIFTY, a broad based market index has been used as a proxy for the market portfolio. The study also uses various accounting and financial information regarding the sample companies such as market capitalization, P/E ratio, BE/ME ratio and D/E ratio. The data have been primarily collected from PROWESS (a financial database of Centre for Monitoring Indian Economy) and web sources such as rbi.org, sebindia.com and nseindia.com. It is important to mention here that the entire data set (regarding four company fundamentals as well as closing adjusted share prices) was not available for all sample companies throughout the sample period of 10 years. Hence effective number of companies used in the analysis ranges from 295 to 455.

V

OPERATIONAL

DEFINITIONS

OF

VARIOUS

COMPANY

FUNDAMENTALS USED IN THE STUDY As mentioned earlier we have used four company fundamentals in the study. The selection of these fundamentals is based on the fact that robust CAPM anomalies have already been detected in developed countries using these variables. Table 1 provides operational definitions of various company fundamentals used in the study. VI

RESEARCH METHODOLOGY

Internationally accepted methodology as used by Davis Fama and French (2000) and Chan, Hamao and Lakonishok (1991) has been used to test various research hypotheses regarding relationship between company fundamentals and equity returns. (i)

Construction of Portfolios

In June-end of year T all the sample companies are ranked on the basis of size (measured by market capitalization : MC). The ranked sample companies are then divided into 5 equally weighted portfolios namely P1MC, P2MC, P3MC, P4MC and P5MC. P1MC is the smallest sized portfolio consisting of 20 percent of companies with lowest size while P5MC consists of top 20 percent companies with largest size. The process is repeated 9

using P/E ratio, BE/ME ratio and D/E ratio as the sorting variable. Since the study uses four company fundamental variables (MC, P/E, BE/ME and D/E) there are four sets of 5 portfolios each or in total 20 portfolios. Portfolios sorted on the basis of P/E ratio have been specified as P1PE (lowest), P2PE, P3PE, P4P3 and P5PE (highest). Portfolios sorted on the basis of BE/ME ratio are named as P1BEME (lowest), P2BEME, P3BEME, P4BEME and P5BEME (highest), while those sorted on the basis of D/E ratio are specified as P1DE (lowest), P2DE, P3DE, P4DE and P5DE (highest). Portfolios are rebalanced on annual basis. Then monthly equally weighted returns on all portfolios including market portfolio (proxied by S&P CNX NIFTY) have been calculated from July 1997 till June 2007 giving a total of 120 monthly observations. The relationship between company fundamentals and stock returns has been tested using time series regression as implied in the famous market model equation i.e. R pt − R ft = a p + b p (R mt − R ft ) + e t (for t = 1 to 120)

(2)

(for p = 1 to 20) where R pt − R ft =

Excess return on portfolio i.e. return on portfolio P minus risk for return in month t.

R mt − R ft =

Excess return on market portfolio in month t.

ap =

Intercept term

bp =

Slope coefficient (or beta coefficient) of the market factor.

et =

error term

It must be mentioned here that if ap = 0 then equation (2) reduces to Black Jensen Scholas (1972) version of single factor CAPM. The null hypothesis is that there are no extra normal returns earned on portfolios sorted on the basis of various company fundamentals which is equivalent to testing ap = 0 for all sorted portfolios. The alternate hypothesis is ap ≠ 0. The hypothesis is tested at 5 percent level of significance.

10

In order to test whether the investment strategy based on company fundamentals yields any extra normal returns in Indian equity market, equation (2) is estimated for the following portfolios. (i)

Portfolio consisting of long position in P1MC and a short position in P5MC which is SMB (small minus big) i.e. size based investment strategy.

(ii)

Portfolio consisting of long position in P1PE and short position in P5PE and which is LMH (low minus high) i.e. the P/E ratio based investment strategy.

(iii)

Portfolio consisting of long position in P5BEME and short position in P1BEME which is HML (high minus low) i.e. BE/ME ratio based investment strategy.

(iv)

Portfolio consisting of long position in P5DE and short position in P1DE which is LEVG (high leverage minus low leverage) i.e. leverage or D/E ratio based investment strategy.

In order to test whether inclusion of any one or more of the four company fundamentals (viz. Market capitalization : MC, Price Earnings ratio : P/E, Book equity to market equity ratio : BE/ME and Financial Leverage : D/E ratio) can better explain cross sectional variations in average equity returns in Indian stock market we have used the methodology followed by [Davis, Fama and French (DFF) : 2000] with the following modifications. (i)

DFF (2000) constructed and used only nine portfolios based on size and book to market equity. We have constructed, and used 20 portfolios based on size, P/E ratio, book to market equity ratio and D/E ratio.

(ii)

DFF (2000) used the following 3 factors and tested Fama-French three factor asset pricing model equation :

Factors used by DFF (2000) (a)

Market Risk Premium = (RM – RF)

(b)

Size Premium = SMB = Return differential between small & large firms portfolios.

(c)

Value Premium = HML (High Minus Low) = Return differential between high BE/ME stocks portfolio and low BE/ME stocks portfolio. 11

Instead we have used the following five factors : Factors used in the present study (a)

Market Risk Premium = R M − R F

(b)

Size Premium = SMB (Small Minus Big) = Monthly Return differential between PIMC (Smallest Stocks Portfolio) and P5MC (Largest Stocks Portoflio).

(c)

P/E risk premium = LMH (Low Minus High) = Monthly return differential between PIPE (lowest P/E stocks portfolio) and P5PE (highest P/E stocks portfolio).

(d)

Value risk premium = HML (High Minus Low) = Monthly return differential between P5BEME (Highest BE/ME stocks portfolio) and P1BEME (Lowest BE/ME stocks portfolio)

(e)

Leverage risk premium = LEVG = Monthly Return differential between P5DE (Highest D/E stocks portfolio) and PIDE (Lowest D/E stocks portfolio).

Then we have estimated the following first pass time series regression equations for each portfolio over the 120 months between July 1997 – June 2007. The standard notations used in equations (3 to 33) are given below : R pt

= Portfolio return in month t

R ft

= Risk free return in month t

R mt

= Return on market portfolio in month t

SMBt = Size risk premium in month t

LMHt = P/E risk premium in month t HMLt = Value premium in month t LEVGt= Leverage risk premium in month t a

= Intercept

b

= slope coefficient of market risk premium i.e. beta

s

= Slope coefficient or factor loading of size risk premium

p

= Slope coefficient or factor loading of P/E risk premium 12

h

= Slope coefficient or factor loading of value risk premium

λ

= Slope coefficient or factor loading of leverage risk premium

et

= error term

Statistically significant values of slope coefficient of various factors would indicate that those factors are important in explaining cross sectional variations in portfolio returns otherwise not. Moreover whether independent variable(s) in a particular model significantly explain cross sectional variations in equity portfolio returns or not can be detected by looking at its adjusted R2 value. The higher the value of adjusted R2 the greater is the explanatory power of the independent variable(s) included in the model. I.

Single Factor Model

Here one independent factor is used to estimate portfolio excess returns i.e. the dependent variable. We have used all four company fundamentals separately for this purpose and compared the results with the results of the single factor market model. (i)

Market alone R pt − R ft = a + b(R mt − R ft ) + e t

for t = 1 … 120

(3)

p = 1 … 20 (ii)

Size alone R pt − R ft = a + s(SMB t ) + e t

(iii)

(4)

P/E risk premium alone R pt − R ft = a + p(LMH t ) + e t

(iv)

(5)

Value premium alone R pt − R ft = a + h (HML t ) + e t

(v)

(6)

Leverage risk premium alone R pt − R ft = a + λ(LEVG t ) + e t

II.

(7)

Two Factor Model :

13

Here we have used two independent variables to estimate portfolio excess returns i.e. the dependent variable. (i)

Market and Size R Pt − R ft = a + b(R mt − R ft ) + s(SMB t ) + e t

(ii)

Market and P/E risk premium R Pt − R ft = a + b(R mt − R ft ) + p(LMH t ) + e t

(iii)

(12)

Size and value premium R pt − R ft = a + s[SMB t ] + h[HML t ] + e t

(vii)

(11)

Size and P/E risk premium R pt − R ft = a + s[SMB t ] + p[LMH t ] + e t

(vi)

(10)

Market and Leverage (D/E ratio) risk premium R pt − R ft = a + b(R m − R ft ) + λ(LEVG t ) + e t

(v)

(9)

Market and Value Premium R pt − R ft = a + b(R mt − R ft ) + h [HMLt ] + e t

(iv)

(8)

(13)

Size and Leverage risk premium R pt − R ft = a + s[SMB t ] + λ[LEVG t ] + e t

(14)

(viii) P/E risk premium and value premium R pt − R ft = a + p[LMH t ] + h[HML t ] + e t

(ix)

P/E risk premium and leverage premium R pt − R ft = a + p[LMH t ] + λ[LEVG t ] + e t

(x)

(16)

Value premium and leverage risk premium R pt − R ft = a + h[HML t ] + λ[LEVG t ] + e t

III.

(15)

Three Factor Model

14

(17)

Here we have included three independent factors to explain portfolio excess returns (i.e. the dependent factor). (i)

Market, Size and P/E risk premium R pt − R ft = a + b[R mt − R ft ] + s[SMB t ] + p[LMH t ] + e t

(ii)

(18)

Market, Size and value premium R pt − R ft = a + b[R mt − R ft ] + s[SMB t ] + h[HML t ] + e t

(19)

This is the famous Fama-French three factor asset pricing model equation. (iii)

Market, Size and Leverage R pt − R ft = a + b[R mt − R ft ] + s[SMB t ] + λ[LEVG t ] + e t

(iv)

Market, P/E and value premium R pt − R ft = a + b(R mt − R ft ) + p(LMH t ) + h (HML t ) + e t

(v)

(22)

Market, Value and Leverage Premium R pt − R ft = a + b(R mt − R ft ) + h (HML t ) + λ (LEVG t ) + e t

(vii)

(21)

Market, P/E and Leverage premium R pt − R ft = a + b(R mt − R ft ) + p(LMH t ) + λ (LEVG t ) + e t

(vi)

(20)

(23)

Size, P/E and value premium R pt − R ft = a + s(BMB t ) + p(LMH t ) + h (HML t ) + e t

(24)

(viii) Size, P/E and Leverage premium R pt − R ft = a + s(SMB t ) + p(LMH t ) + λ (LEVG t ) + e t

15

(25)

(ix)

Size, Value and Leverage premium R pt − R ft = a + s(SMB t ) + h (HML t ) + λ (LEVG t ) + e t

(x)

P/E, Value and Leverage Premium R pt − R ft = a + p(LMH t ) + h (HML t ) + λ (LEVG t ) + e t

IV.

(26)

(27)

Four Factor Model

Here we have included four independent variables to explain the dependent variable i.e. portfolio excess returns. (i)

Market, Size, P/E and Value Premium R pt − R ft = a + b(R mt − R ft ) + s(SMTt ) + p(LMH t ) + h (HML t ) + e t

(ii)

Market, Size, Value and Leverage Premium R pt − R ft = a + b(R mt − R ft ) + s(SMB t ) + h (HML t ) + λ (LEVG t ) + e t

(iii)

(31)

Size, P/E, Value and Leverage R pt − R ft = a + s(SMB t ) + p(LMH t ) + h (HML t ) + λ (LEVG t ) + e t

V.

(30)

Market, P/E, Value and Leverage Premium R pt − R ft = a + b(R mt − R ft ) + p(LMH t ) + h (HML t ) + λ(LEVG t ) + e t

(v)

(29)

Market, Size, P/E and Leverage R pt − R ft = a + b(R mt − R ft ) + s(SMB t ) + p(LMH t ) + λ(LEVG t ) + e t

(iv)

(28)

(32)

Five Factor Model

Here we have included all five factors under study as independent variables to estimate the portfolio excess returns (i.e. dependent variable). The estimated regression equation is R pt − R ft = a + b(R mt − R ft ) + s(SMB t ) + p(LMH t ) + h (HML t ) + λ (LEVG t ) + e t

16

(33)

We have used Statistical Package for Social Sciences (SPSS) and Excel for the purpose of data analysis. VII EMPIRICAL RESULTS

Table 2 presents cross correlation matrix of fundamental variables used in the study. As is evident there is positive but low relationship between size and P/E ratio. There is negative but low relationship between size and D/E ratio; and size and BE/ME ratio. However a statistically significant positive relationship exists between D/E ratio and BE/ME ratio; and a statistically significant negative relationship exists between D/E ratio and P/E ratio and BE/ME ratio and P/E ratio. Thus it may be said that in Indian context BE/ME ratio, P/E ratio and D/E ratio tend to capture almost similar firm characteristics. VIIa

Relationship Between Company Fundamentals and Equity Returns in Indian Stock Market

First of all the relationship between four company fundamentals ((viz. market capitalization, P/E ratio, BE/ME ratio and debt equity ratio) and average stock returns are analysed using correlation coefficients. The results are presented in Table 2. It can be observed that there exists a (i)

statistically significant negative relationship between company size and average stock returns over the study period.

(ii)

Statistically significant negative relationship between P/E ratio and average stock returns over the study period.

(iii)

Statistically significant positive relationship between BE/ME ratio and average stock returns over the study period.

(iv)

Statistically significant positive relationship between Debt Equity ratio and average stock returns over the study period.

These results are further substantiated by constructing various portfolios on the basis of these four company fundamentals and then analyzing the pattern of mean monthly returns.

17

Table 3 provides summary statistics of monthly excess returns of all 20 portfolios sorted on the basis of four company fundamentals, while Table 4 provides results of the market factor model. (i)

Regarding Company Size and Equity Returns

The results regarding size sorted portfolios are presented in Panel A of Table 3 and Panel A of Table 4. It is clearly visible that mean monthly excess returns of smallest size portfolio (PIMC) is much higher than that of largest sized portfolio (P5MC). The mean excess return of PIMC was found to be 3.34 percent per month as against 1.02 percent per month for P5MC. This clearly provides a size premium (the return differential between PIMC and P5MC) of 2.32 percent per month (t-value 5.800) or about 24 percent per annum which is quite robust. However the standard deviation of PIMC is also higher than that of P5MC pointing towards the intuitive fact that small firms are more risky than their large counterparts. Panel A of Table 4 presents the results of the market model equation used to check for the relationship between company size and equity returns in Indian stock market. It is clear that intercept value (i.e. a) decline monotonically as one moves from PIMC to P5MC. The smallest sized portfolio has provided an extra normal return of 3.06 percent per month over the study period as revealed by its "a" value which is statistically significant (t-value 2.573 as against its critical value of 1.96). Thus we can reject null hypothesis (i.e. ap = 0) as intercept value for this portfolio is positive and statistically significant. The same is true for P2MC. However as one moves from P1MC to P5MC there has been a sharp decline in intercept value and for P3MC, P4MC and P5MC we do not find any statistically significant extra normal returns. These findings indicate that the stocks of small firms outperformed those of large firms over the study period. These results are in line with the results presented earlier by Mohanty (2001) Sehgal & Muneesh (2002), Sehgal and Tripathi (2005) and Tripathi(2007) for Indian stock market. A look at the R2 value reveals the fact that market factor is important in capturing a large amount of variation in equity returns especially for the large stocks portfolios. It is important to note here that R2 value (coefficient of determination) is low for small stocks

18

portfolio (e.g. 50.5 percent for PIMC as against 63.1 percent for P5MC) suggesting that the portfolio of small stocks have larger unexplained variations in their returns. The slope coefficient "b" (i.e. commonly known as beta coefficient) of all the portfolios have been statistically significant but there has been no substantial difference between the beta coefficient of small and large stocks portfolios. This might indicate that market risk of small firms is not substantially higher than that of large firms. (ii)

Regarding Price Earnings Ratio (P/E ratio) and Equity Returns

These results are presented in Panel B of Table 3 and Panel B of Table 4. It is clear that low P/E stocks provided a statistically significant mean monthly excess return of 3.01 percent (t value 3.040) as against 1.33 percent per month by high P/E stocks portfolio over the study period. The mean monthly excess return declines as one moves from PIPE to P5PE. However as was the case with size based portfolios, portfolio returns of low P/E stocks have also shown higher standard deviation (or variability) than those of high P/E stocks. The LMH (low minus high) risk premium based on P/E ratio has been found to be a statistically significant 1.68 percent per month (t value 3.111) or about 20 percent per annum over the study period. If one looks at the intercept values of the market model results presented in Panel B of Table 4, one finds that the "a" values decline monotonically as one moves from PIPE to P5PE, showing that low P/E stocks portfolio provided the investors with statistically significant extra risk adjusted returns over the study period. The lowest P/E stocks portfolio i.e. PIPE provided an extra risk adjusted return of 2.17 percent per month (t value 2.879) as against 0.45 percent per month (t value 1.015) on highest P/E stocks portfolio. (iii)

Regarding Book Equity to Market Equity Ratio (BE/ME Ratio) and Equity Returns

The summary statistics and market model results of portfolios sorted on the basis of BE/ME ratio are presented in Panel C of Table 3 and Panel C of Table 4 respectively. As expected mean monthly excess return of high BE/ME stocks portfolio (P5BEME) is much larger and statistically significant (3.06 percent per month with t value 3.091) than that of low BE/ME stocks portfolio (P1BEME : 1.49 percent per month with t value 1.886). Moreover the standard deviation of high BE/ME stocks portfolio is also higher 19

than that of low BE/ME stocks portfolio. The intercept value "a" as shown in Panel C of Table 4 also increases monotonically from 0.67 per cent per month (t value 1.434) for P1BEME to 2.23 percent per month (t value 2.957) for P5BEME suggesting that high BE/ME stocks portfolio generated higher risk adjusted extra return during the study period. The value premium (i.e. the return differential between P5BEME and P1BEME) is as high as 1.57 percent per month (t value 2.492) which is also statistically significant. Hence we can conclude that during the study period a strong value effect existed in the Indian stock market. However the intensity of this effect is slightly lower as found by Muneesh Kumar and Sehgal (2004) and Sehgal and Tripathi (2007). (iv)

Regarding Debt-Equity Ratio (D/E ratio) and Equity Returns

Bhandari (1988) found leverage effect in equity returns implying that stocks of firms having high financial leverage provide higher risk adjusted returns than those of firms having low financial leverage. The results of our analysis regarding financial leverage (as measured by Debt Equity ratio) and equity returns in India are presented in Panel D of Table 3 and Panel D of Table 4. Panel D of Table 3 shows that mean monthly excess return of high D/E stocks portfolio (P5DE) has been 2.71 percent (t value 2.823) as against 1.40 percent (t value 1.750) on low D/E stocks portfolio (PIDE). As expected the standard deviation of high D/E stocks portfolio is also higher than that of low D/E stocks portfolio. The return differential between high and low D/E stocks portfolio, popularly known as leverage risk premium (LEVG) has been found to be 1.31 percent per month (t value 2.673) which is also statistically significant. Panel D of Table 4 shows that the intercept terms 'a' of P3DE, P4DE and P5DE are higher and statistically significant than those of PIDE and P2DE. The extra normal return of P5DE is 1.86 percent per month (t value 2.845) as against 0.61 percent per month (t value 1.186) for PIDE. This suggests that during the study period stocks of high financial leverage firms outperformed those of low financial leverage firms implying the presence of a "leverage effect" in the Indian stock market. A peculiar feature of all the above results has been that the slope coefficients (or beta coefficients) of all portfolios have been statistically significant but R2 values have been lower for PIMC, PIPE, P5BEME and P5DE portfolios and high for P5MC, P5PE, 20

PIBEME and PIDE portfolios suggesting that portfolios of small capitalization stocks, low P/E stocks, high BE/ME stocks and high D/E stocks have larger unexplained variations in their returns than those of large capitalization stocks, high P/E stocks, low BE/ME stocks and low D/E stocks, although market factor has been important in capturing cross sectional variations in average stock returns of all portfolios. VIIb Economic Evaluation of Company Fundamentals based Investment Strategy

The statistically significant relationship between company fundamentals and equity returns in India gives rise to arbitrage opportunities which can be used to earn extra returns on risk adjusted basis in Indian stock market. Table 5 presents results regarding the extra returns on a risk adjusted basis which can be generated by investment strategies based on four company fundamentals viz. Market capitalization, P/E ratio, BE/ME ratio and D/E ratio. It can be observed that size based investment strategy generated a statistically significant extra normal return of 2.43 percent per month (t value 2.273), P/E ratio based investment strategy provided 1.71 percent per month (t value 3.157), BE/ME based strategy gave 1.56 percent per month (t value 2.474) and D/E ratio based strategy provided the investors with an extra normal return of 1.25 percent per month (t value 2.502) over the study period. The fact that all these investment strategies generated statistically significant extra risk adjusted returns points towards the fact that arbitrage opportunities were present in Indian stock market during the study period. VIIc : Cross sectional variations in Equity returns

The empirical results regarding the role of company fundamentals in explaining cross sectional variations in equity returns in Indian stock returns are presented in Table 4 and from Table 6 to 10. It is clearly visible from Table 4 that market factor (excess return on market portfolio) captures the most part of cross-sectional variations in equity returns in India, but not all. Moreover Panel A to D of Table 6 shows that no other factor (be it size premium, P/E risk premium, value premium or leverage premium) can capture any significant portion of cross sectional variations in average equity returns, in isolation, as all other single factor models have very low R2 values as compared to the single factor market model. 21

Hence we conclude that the company fundamentals, per se, are not capable of explaining cross sectional variations in equity returns in India. They must be clubbed with market factor (or some other independent variable) to check whether a multifactor model can better explain cross sectional variations in equity returns in India or not. The results regarding two factor model based on market and size factors are presented in Table 7. It can be observed that there has been considerable improvement in adjusted R2 value when both excess market return and size premium are used as independent variables. This can also be confirmed by the fact that the slope coefficient of size premium i.e. s is statistically significant for all 20 portfolios while all (except six) intercept values i.e. 'a' values are very low and not statistically significant. Hence we conclude that size and market factors together can better explain cross sectional variations in equity returns in India than the market factor alone. As far as other company fundamentals are concerned we found an improvement in adjusted R2 value when they are used in addition to the market factor but such an improvement has not been as large as the one produced by market and size factors. Hence detailed results are not provided here. Regarding various three factor models, Fama French three factor model (based on market , size and value premium) turned out to be the best in explaining cross sectional variations in equity returns in India. The results of this model are presented in Table 8. These results are in line with those found by Connor and Sehgal(2003). In case of various four factor models the one based on market, size, P/E and value premium provides the best results as shown in Table 9.However here adjusted R2 values have improved only marginally as compared to the three factor model based on market, size premium and value premium. This might be due to the overlapping effect of value and P/E risk premium in Indian context. Finally, the results of the five factor model have been provided in Table 10. It is clearly visible that the five factor model does not show any substantial improvement in explaining cross sectional variations in equity returns in India over three or four factor models, as adjusted R2 values have improved only marginally with the inclusion of two additional factors (P/E risk premium and leverage premium). 22

VIII

SUMMARY OF THE RESEARCH RESULTS

On the basis of the empirical results presented in this paper, following conclusions may be drawn. (i)

There existed a statistically significant negative relationship between company size (Market Capitalisation) and equity returns in India over the study period. The smallest stocks portfolio (PIMC) outperformed largest stocks portfolio and provided the investors with an extra risk adjusted return of 3.06 percent per month (t value 2.573) as against 0.63 percent per month (t value 1.329) on P5MC i.e. largest stocks portfolio. The size premium (i.e. the return differential between smallest and largest stocks portfolios) has been found to be 2.32 percent per month (t value 5.800) over the study period which is quite robust.

(ii)

There existed a statistically significant negative relationship between P/E ratio and equity returns in India over the study period. The lowest P/E ratio stocks portfolio (PIPE) outperformed the highest P/E stocks portfolio (P5PE) and provided the investors with an extra risk adjusted return of 2.17 percent per month (t value 2.879). The P/E risk premium (i.e. the return differential between PIPE and P5PE) has been found to be statistically significant over the study period (1.68 percent per month with t value 3.111).

(iii)

There existed a statistically significant positive relationship between BE/ME ratio and equity returns in India over the study period. The highest BE/ME stocks portfolio outperformed the lowest BE/ME stocks portfolio. The highest BE/ME stocks portfolio (P5BEME) produced an extra risk adjusted return of 2.23 percent per month (t value 2.957) as against 0.67 percent per month (t value 1.434) on P1BEME i.e. the lowest BE/ME stocks portfolio. The HML premium or value premium (i.e. the return differential between P5BEME and P1BEME) has been found to be 1.57 percent per month (t value 2.492) which is also statistically significant.

(iv)

There existed a statistically significant positive relationship between D/E ratio and equity returns in India over the study period. The highest D/E stocks portfolio (P5DE) outperformed the lowest D/E stocks portfolio (PIDE) and provided an 23

extra risk adjusted return of 1.86 percent per month (t value 2.845) as against 0.61 percent per month (t value 1.186) on lowest D/E stocks portfolio i.e. PIDE. The leverage risk premium has been found to be statistically significant at 1.31 percent per month (t value 2.673) over the study period. (v)

The investment strategies based on size, P/E ratio, BE/ME ratio and D/E ratio would have provided the investors with statistically significant extra risk adjusted returns of 2.43 percent (t value 2.273), 1.71 percent (t value 3.157), 1.56 percent (t value 2.474) and 1.25 percent (t value 2.502), respectively over the study period. It shows that opportunities are available for Indian investors to earn extra returns on a risk adjusted basis by following investment strategies based on company fundamentals.

(vi)

Excess market return has been found to be an important factor in explaining cross sectional variations in equity returns in India although it is not capable of explaining all such variations. However none of the company fundamentals, in isolation, could explain cross sectional variations in equity returns in India in any significant way. It implies that other company fundamentals should be added to the asset pricing model in order to explain cross-sectional variations in equity returns in India in a better way.

(vii)

The three factor model based on market, size premium and value premium (Popularly known as Fama-French Multifactor asset Pricing Model) explained cross-sectional variations in equity returns in India in a much better way than the single factor CAPM or any two factor model. Four factors or five factors models did not improve the results regarding cross-sectional variations in equity returns in India in any significant manner and hence we may conclude that the three factor Fama-French model works well in Indian context.

IX

POLICY IMPLICATIONS OF THE FINDINGS

The findings of this research paper have important policy implications and are of pertinent use for equity analysts, fund managers and investing community at large. 24

(i)

Implications for Market Efficiency : We have found a statistically significant

relationship between four company fundamentals (viz. market capitalization, P/E ratio, BE/ME ratio and D/E ratio) and equity returns in India over the study period of 1997-2007. This implies that a strong size effect, P/E effect, value effect and leverage effect existed in Indian stock market over the most recent ten years period. This further implies that Indian stock market is still not semi strong efficient because publically available financial information can be used to earn extra risk adjusted return in Indian stock market. Although the intensity or robustness of these effects have been lower than those detected by earlier studies during the decade of 1990s. Hence although the efficiency level has been increasing in Indian stock market, it has still not become fully semi-strong efficient. (ii)

Implications for Investment Strategies : Presence of strong size effect, P/E

effect, value effect and leverage effect are indicative of the fact that arbitrage opportunities are available in Indian stock market and gainful investment strategies can be formulated and used by equity analysts, fund managers and investing community at large based on these company fundamentals. Since the research results are of the most recent ten years period their utility gets further enhanced in this context. (iii)

Implications for Asset Pricing Framework

a. We have found that no single factor asset pricing model (be it based on market risk premium or any of the company fundamentals) works well in explaining cross sectional variations in equity returns in India. Hence more factors should be included in the asset pricing framework. We have found that addition of two company fundamentals (viz. size premium and value premium) in the asset pricing model can substantially explain cross-sectional variations in equity returns in India. This lends further support to the Fama French three factor asset pricing model in Indian stock market in a recent time period. b. Contrary to Fama & French (1996) we have found that market risk premium is still ‘the’ most important independent factor in asset pricing framework although its relative importance has substantially declined since the decade of 25

1980’s or 1990’s. This implies that company fundamentals are gaining importance in explaining cross sectional variations in equity returns in India. (iv)

Implications for Market Microstructure : The findings also have important

implications for market microstructure aspects as we have found the presence of strong size effect, P/E effect, value effect and leverage effect on National Stock Exchange (i.e. the prominent exchange having much higher turnover than that of Bombay Stock Exchange). Earlier studies have found the presence of size and value effect on Bombay Stock Exchange (Sehgal and Tripathi (2005,2007). This implies that unlike Reinganum (1990) market microstructural aspects do not affect the relationship between company fundamentals and equity returns in India in any substantial manner.

26

References : Banz, Rolf W. 1981The Relationship between Return and Market Value of Common Stock, Journal of Financial Economics, March :3-18. Basu, Sanjoy.1977.Investment Performance of Common Stocks in Relation to their PriceEarnings Ratios: A Test of Efficient Market Hypothesis, Journal of Finance, (32):663-682. Bhandari, L.C.1988. Debt-Equity Ratio and Expected Common Stock Returns: Empirical Evidence", Journal of Finance ( 43):507-528. Black, F., Jensen, M., and Scholes, M.1972. The Capital Asset Pricing Model: Some Empirical Tests, in Studies in the Theory of Capital Markets, M. Jensen (ed.), Praeger, New York:79-121. Chan, L., Hamao, Y., and Lakonishok, J.1991.Fundamentals and Stock Returns in Japan", Journal of Finance,(46):1739-1764. Connor G. and Sehgal S.2003.Tests of Fama and French Model in India.Decision, 30 (2), :1-30. Davis J.L., Fama E.F. & French K.R. 2000.Characteristics, Covariances and Average Returns, 1929 to 1997, Journal of Finance, 55(1):389-406. Fama Eugene F, and Frnech Kenneth R., 2004. The Capital Asset Pricing Model : Theory and Evidence, Journal of Economic Perspectives 18 (3).25-46 Fama, E. and K. French .1992, The Cross Section of Expected Stock Returns, Journal of Finance,.47:427-466.

Fama, E. and K. French .1993, Common Risk Factors in the Returns of Stock and Bonds, Journal of Financial Economics,33: 3-56.

Fama, E. and K. French.1996, Multifactor Explanations of Asset Pricing Anomalies, Journal of Finance, 51:55-84.

27

Muneesh Kumar, Sehgal S.2004.Company Characteristics and Common Returns : The Indian Experience, Vision, July-December: 34-45. Mohanty P.,2001.Efficiency of the market for small stocks, NSE research initiative, April series. Reinganum, Marc R.1990. Market Microstructure and Asset Pricing: An Empirical Investigation of NYSE and NASADQ Securities, Journal of Financial Economics, 28:127-148. Rosenberg, B., Reid, K., and Lanstein, R.1985.Persuasive Evidence of Market Efficiency, Journal of Portfolio Management, 11:.9-17.

Sehgal S and Tripathi V.2007.Value Effect in Indian Stock Market, The ICFAI Journal of Applied Finance .13 (1): 23-36.

Sehgal S. and Tripathi V., 2005. Size Effect in Indian Stock Market. Vision,9(4):27-42. Sehgal S. Munnesh Kumar.2002.The Relationship Between Company Size, Relative Distress and Returns in Indian Stock Market, The ICFAI Journal of Applied Finance,8(2):41-50.

Stattman, D.1980.Book Values and Stock Returns, The Chicago MBA - A Journal of Selected Papers, 4:25-45.

Tripathi V, 2007, Size Effect in Indian Stock Market, Serials Publications, New Delhi.

28

Table 1 Operational Definitions of Various Company Fundamentals used in the Study S.No. Fundamental Variable

Measured by

1.

Size

Market capitalization (MC) as on June end every year

2.

Price Earnings ratio

Price Earnings ratio (P/E ratio) as on June end every year

3.

Book Equity to Book equity to market equity ratio (BE/ME ratio) as on Market Equity Rate June end every year. This is calculated as inverse of Price to Book value ratio (PB ratio) provided by PROWESS database as on that date.

4.

Financial leverage

Debt equity ratio (D/E ratio) as on March end every year

Table 2 Cross Correlation-Matrix of Various Fundamental Variables and average Portfolio Returns [Pearson's coefficient of correlation] D/E

P/E

BE/ME Average Portfolio Return

MC

-0.209 .414

-.307

D/E

-

-0.608** +.897** .821**

P/E

-

-

-.764**

BE/ME

-.716**

-0.816** +0.765**

Note : Correlations are calculated across portfolios over the study period. ** Significant at 5% level

29

Table 3 Summary Statistics of monthly excess returns of Portfolios sorted on the basis of various company fundamentals (Total Period July 1997- June 2007) Panel A : Size Based (Firm Size increases as one moves from P1MC to P1MC) Portfolio P1MC P2MC P3MC P4MC P5MC SMB (P1MC-P5MC)

Mean .0334 .0245 .0237 .0127 .0102 .0232

SE (Mean) .0085 .0088 .0088 .0084 .0077 .0040

t(Mean) 3.929* 2.784* 2.693* 1.512 1.325 5.800*

S.D. .0986 .0967 .0964 .0924 .0848 .0442

Panel B : P/E Ratio Based (P/E ratio increases as one moves from P1PE to P5PE) Portfolio P1PE P2PE P3PE P4PE P5PE LMH (P1PE-P5PE)

Mean .0301 .0241 .0206 .0160 .0133 .0168

SE (Mean) .0099 .0081 .0082 .0083 .0081 .0054

t(Mean) 3.040* 2.975* 2.512* 1.927 1.6419 3.111*

S.D. .1089 .0883 .0902 .0907 .0891 .0588

*Significant at 5 percent level

Panel C : BEME Ratio Based (BEME ratio increases as one moves from P1BEME to P5BEME) Portfolio P1BEME P2BEME P3BEME P4BEME P5BEME HML (P1BEME-P5BEME)

Mean .0149 .0195 .0171 .0243 .0306 .0157

SE (Mean) .0079 .0088 .0084 .0085 .0099 .0063

t(Mean) 1.886 2.216* 2.036* 2.858* 3.091* 2.492*

S.D. .0868 .0908 .0924 .0930 .1084 .0685

*Significant at 5 percent level Panel D : Financial Leverage or D/E ratio based (D/E ratio increases as one moves from P1DE to P5DE) Portfolio P1DE P2DE P3DE P4DE P5DE LEVG (P5DE-P5DE)

Mean .0140 .0151 .0212 .0240 .0271 .0131

SE (Mean) .0080 .0080 .0082 .0086 .0096 .0049

30

t(Mean) 1.750 1.887 2.585* 2.791* 2.823* 2.673*

S.D. .0877 .0873 .0898 .0943 .1056 .0543

Table 4 Results of the Market Model

Rpt – Rft = a + b (RMt – Rft) + et Panel A : Portfolios sorted on the basis of Size (MC) Portfolio P1MC P2MC P3MC P4MC P5MC

a .0306 .0163 .0053 .0082 .0063

B .932 .973 1.010 1.016 .943

t(a) 2.573* 2.643* 1.600 1.317 1.329

t(b) 11.068* 11.303* 12.341* 12.904* 14.310*

Adj R – Square .505 .516 .560 .618 .631

Panel B : Portfolios sorted on the basis of P/E Ratio Portfolio P1PE P2PE P3PE P4PE P5PE

a .0217 .0167 .0127 .0101 .0045

B 1.008 .890 .948 .986 1.042

t(a) 2.879* 2.957* 2.314* 1.925 1.015

t(b) 9.613* 11.335* 12.439* 13.494* 16.675*

Adj R – Square .434 .517 .563 .603 .700

Panel C : Portfolios sorted on the basis of BE/ME Ratio Portfolio P1BEME P2BEME P3BEME P4BEME P5BEME

a .0067 .0110 .0091 .0166 .0223

B .9850 1.018 .958 .920 .993

t(a) 1.434 1.207 1.595 2.74* 2.957*

t(b) 15.153* 14.639* 12.076* 10.929* 9.439*

Adj R – Square .658 .642 .549 .499 .425

Panel D : Portfolio sorted on the basis of D/E Ratio Portfolio P1DE P2DE P3DE P4DE P5DE

a .0061 .0123 .0132 .0156 .0186

B .947 .942 .960 1.003 1.023

t(a) 1.186 1.398 2.478* 2.770* 2.845*

*Significant at 5 percent level

31

t(b) 13.257* 13.231* 12.929* 12.786* 10.481*

Adj R – Square .595 .594 .583 .577 .478

Table 5 Evaluation of Investment Strategy Based on Various Company Fundamentals Strategy based on ‘a’ differential t (‘a’ differential)

MC

.0243

2.273*

P/E Ratio

.0171

3.157*

BE/ME Ratio

.0156

2.474*

D/E Ratio

.0125

2.502*

* Significant at 5% level.

32

Table 6 Single Factor Model Regression Results Panel A : Size as Independent Factor R pt − R ft = a + b(SMB t ) + e t a

s

t(a)

t(s)

Adj. R2

P1MC

.0151

.90

2.904

5.099

.174

P2MC

.0184

.662

2.133

3.453

.082

P3MC

.0189

.524

2.160

2.687

.050

P4MC

.0172

.369

2.029

1.95

.023

P5MC

.0151

-.100

1.904

-.569

-.006

P1PE

.0242

.643

2.459

2.934

.060

P2PE

.0199

.459

2.474

2.563

.045

P3PE

.0165

.446

2.003

2.434

.040

P4PE

.0147

.401

1.763

2.163

.030

P5PE

.0096

.403

1.17

2.22

.032

P1BEME

.0105

.480

1.336

2.738

.052

P2BEME

.0158

.404

1.899

2.179

.031

P3BEME

.0135

.394

1.589

2.087

.027

P4BEME

.0200

.458

2.362

2.426

.039

P5BEME

.0249

.458

2.362

2.426

.039

P1DE

.0094

.497

1.187

2.813

.055

P2DE

.0156

.492

1.187

2.813

.055

P3DE

.0177

.385

2.144

2.097

.028

P4DE

.0196

.475

2.281

2.482

.042

P5DE

.5224

.504

2.326

2.344

.036

Portfolio

*all t values above 1.96 statistically significant. 33

Panel B : P/E Risk Premium as Independent Factor R pt − R ft = a + p(LMH t ) + e t a

p

t(a)

t(p)

Adj. R2

P1MC

.0131

.609

1.590

4.502

.139

P2MC

.0150

.561

1.734

3.948

.109

P3MC

.0151

.513

1.729

3.579

.090

P4MC

.0130

.454

1.545

3.276

.076

P5MC

.0062

.474

.811

3.781

.108

P1PE

.0121

1.068

1.429

7.671

.327

P2PE

.0143

.585

1.84

4.596

.145

P3PE

.0122

.496

1.505

3.716

.097

P4PE

.0117

.395

1.400

2.880

.058

P5PE

.0121

.0683

1.429

.490

-.006

P1BEME

.0149

.013

1.801

.010

-.008

P2BEME

.0138

.339

1.637

2.444

.040

P3BEME

.0075

.568

.919

4.216

.129

P4BEME

.0123

.714

1.549

5.496

.197

P5BEME

.0139

.990

1.602

6.914

.282

P1DE

.0093

.278

1.135

2.059

.027

P2DE

.0134

.399

1.675

3.027

.064

P3DE

.0126

.512

1.564

3.865

.015

P4DE

.0143

.575

1.705

4.179

.122

P5DE

.0128

.848

1.445

5.823

.217

Portfolio

*all t values above 1.96 statistically significant.

34

Panel C : Value Premium as Independent Factor R pt − R ft = a + h (HML t ) + e t

a

h

t(a)

t(h)

Adj. R2

P1MC

.0148

.543

1.836

4.709

.151

P2MC

.0171

.469

1.996

3.825

.103

P3MC

.0162

.476

1.905

3.904

.107

P4MC

.0138

.434

1.684

3.609

.096

P5MC

.0065

.485

.893

4.629

.147

P1PE

.0167

.856

1.93

6.942

.218

P2PE

.0147

.599

2.00

5.708

.210

P3PE

.0132

.470

1.670

4.146

.120

P4PE

.0127

.357

1.553

3.037

.065

P5PE

.0113

.127

1.352

1.070

.001

P1BEME

.0157

-0.50

1.928

-.433

-.007

P2BEME

.0148

.303

1.777

2.552

.044

P3BEME

.0088

.523

1.111

4.573

.143

P4BEME

.0135

.683

1.793

6.326

.247

P5BEME

.0157

.950

1.928

8.144

.354

P1DE

.0102

.243

1.259

2.097

.028

P2DE

.0147

.344

1.865

3.040

.065

P3DE

.0138

.477

1.748

4.242

.125

P4DE

.0149

.573

1.859

4.981

.167

P5DE

.0150

.772

1.744

6.280

.244

Portfolio

*all t values above 1.96 statistically significant.

35

Panel D : Leverage as Independent Factor R pt − R ft = a = l(LEVG t ) + e t

a

l

t(a)

t(l)

Adj. R2

P1MC

.0157

.584

1.892

3.910

.107

P2MC

.0175

.534

2.010

3.412

.082

P3MC

.0159

.597

1.858

3.878

.106

P4MC

.0129

.592

1.58

4.025

.113

P5MC

.0066

.580

.886

4.338

.130

P1PE

.0174

.970

1.934

5.994

.227

P2PE

.0156

.650

2.043

4.731

.152

P3PE

.0143

.480

1.758

3.274

.075

P4PE

.0119

.490

1.46

3.328

.078

P5PE

.0093

.300

1.132

2.017

.025

P1BEME

.0127

.174

1.595

1.188

.003

P2BEME

.0146

.378

1.748

2.521

.043

P3BEME

.0101

.534

1.222

3.590

.091

P4BEME

.0139

.786

1.793

5.612

.204

P5BEME

.0173

1.017

1.967

6.420

.253

P1DE

.0128

.0875

1.557

.589

-.006

P2DE

.0145

.429

1.830

3.03

.063

P3DE

.0135

.590

1.707

4.145

.120

P4DE

.0148

.695

1.827

4.741

.153

P5DE

.0128

1.088

1.557

7.321

.307

Portfolio

*all t values above 1.96 statistically significant.

36

Table 7 Results of Two Factor Model based on Market and Size R Pt − R ft = a + b(R mt − R ft ) + s(SMB t ) + e t

a

b

s

t(a)

t(b)

t(s)

Adj.R2

P1MC

.0070

.942

.926

1.439

14.264

8.652

.691

P2MC

.0090

.980

.690

1.76

12.738

5.535

.672

P3MC

.0101

1.0160

.553

1.824

13.382

4.49

.694

P4MC

.0084

1.021

.398

1.654

14.619

3.520

.732

P5MC

.0069

.942

-.073

1.438

14.264

-.688

.815

P1PE

.0154

1.015

.671

2.148

10.349

4.224

.721

P2PE

.0122

.895

.484

2.248

12.119

4.045

.710

P3PE

.0083

.953

.473

1.575

13.294

4.074

.690

P4PE

.0061

.991

.429

1.207

14.317

3.825

.670

P5PE

.0005

1.046

.433

.128

18.141

4.634

.721

P1BEME

.0020

.991

.508

.462

16.931

5.36

.731

P2BEME

.0070

1.023

.433

1.462

15.654

4.092

.732

P3BEME

.0052

.963

.422

.929

12.670

3.423

.752

P4BEME

.0121

.926

.485

2.059

11.580

3.741

.631

P5BEME

.0163

1.000

.648

2.249

10.101

4.036

.693

P1DE

.0012

.953

.525

.255

14.609

4.963

.726

P2DE

.0075

.948

.519

1.560

14.555

4.914

.739

P3DE

.0093

.964

.413

1.808

13.629

3.598

.732

P4DE

.0109

1.008

.504

2.031

13.746

4.237

.728

P5DE

.0136

1.029

.533

1.989

11.038

3.528

.731

Portfolio

*all t values above 1.96 statistically significant.

37

Table 8 Three Factor Model Results based on Market, Size & Value Premium R pt − R ft = a + b[R mt − R ft ] + s[SMB t ] + h[HML t ] + e t

a

b

s

h

t(a)

t(b)

t(s)

t(h)

Adj.R2

P1MC

.000

.938

.859

.486

.001

18.545

10.436

9.148

.822

P2MC

.0038

.976

.631

.424

.761

14.445

5.740

5.973

.701

P3MC

.0038

1.012

.492

.439

.785

15.419

4.601

6.37

.717

P4MC

.0026

1.017

.405

.342

.588

16.874

3.486

6.406

.740

P5MC

.000

.938

.486

-.141

.001

18.545

9.148

-1.717

.783

P1PE

.0037

1.007

.558

.815

.746

15.016

5.111

11.578

.768

P2PE

.0040

.890

.405

.568

.985

16.317

4.564

9.931

.767

P3PE

.0020

.949

.412

.438

.441

15.63

4.174

6.864

.723

P4PE

.0014

.988

.384

.326

.306

15.619

3.727

4.909

.703

P5PE

-.0008

1.046

.420

.0943

-.190

18.238

4.503

1.567

.747

P1BEME

.0032

.992

.521

-.089

.747

17.025

5.496

-1.456

.725

P2BEME

.0031

1.020

.395

.272

.679

16.709

3.980

4.238

.724

P3BEME

-.0019

.95

.353

.459

-.409

15.320

3.465

7.529

.720

P4BEME

.0027

.920

.394

.652

.649

16.472

4.336

11.131

.780

P5BEME

.0032

.991

.521

.911

.747

17.025

5.495

14.901

.824

P1DE

-.0017

.951

.496

.206

-.368

15.113

4.844

3.119

.686

P2DE

.0030

.945

.476

.308

.681

15.893

4.917

4.935

.717

P3DE

.0029

.960

.35

.448

.667

16.285

3.65

7.247

.737

P4DE

.0032

1.00

.428

.540

.757

17.892

4.696

9.177

.784

P5DE

.0030

1.022

.43

.7738

.599

15.223

3.938

10.466

.753

Portfolio

*all t values above 1.96 statistically significant.

38

Table 9 Four Factor Model Regression Results based on Market, Size, P/E & Value R pt − R ft = a + b(R mt − R ft ) + s(SMTt ) + p(LMH t ) + h (HML t ) + e t a

b

s

p

h

t(a)

t(b)

t(s)

t(p)

t(h)

Adj.R2

P1MC

-.001

.94

.82

.22

.32

-.27

18.85

9.97

1.92

.322

.83

P2MC

.002

.99

.57

.40

.13

.41

14.96

5.26

2.56

1.03

.71

P3MC

.002

1.02

.45

.23

.26

.56

15.58

4.23

1.5

2.04

.72

P4MC

.001

1.02

.31

.21

.25

.36

17.04

3.14

1.54

2.07

.74

P5MC

-.001

.95

-.17

.22

.32

-.27

18.85

-2.08

1.92

3.22

.79

P1PE

-.0003 1.04

.43

.89

.17

-.07

18.06

4.56

6.58

1.50

.83

Portfolio

P2PE

.003

.89

.39

.04

.53

.92

16.22

4.38

.38

4.87

.76

P3PE

.001

.96

.38

.21

.28

.23

15.78

3.82

1.47

2.37

.73

P4PE

.0002

.99

.34

.26

.13

.05

15.87

3.34

1.80

1.07

.71

P5PE

-.003

1.04

.43

-.11

.17

-.07

18.06

4.56

-.79

1.50

.75

P1BEME

.0020

1.00 .485 .259 -.276 .476 17.32 5.075

1.91

-2.398

.73

P2BEME

.001

1.03

.36

.26

.08

.42

16.96

3.58

1.80

.71

.73

P3BEME

.003

.97

.31

.30

.27

-.70

15.65

3.04

2.06

2.25

.73

P4BEME

.001

.93

.36

.23

.48

.39

16.72

3.94

1.778

4.38

.78

P5BEME

.002

1.00

.48

.25

.72

.47

17.32

5.07

1.91

6.29

.83

P1DE

-.002

.96

.46

.21

.05

-.56

15.26

4.49

1.40

.45

.69

P2DE

.001

.95

.43

.29

.09

.38

16.25

4.48

2.01

.84

.73

P3DE

.001

.97

.31

.26

.26

.39

16.57

3.25

1.90

2.23

.74

P4DE

.002

1.00

.41

.13

.45

.61

17.90

4.41

.97

3.99

.78

P5DE

.001

1.04

.37

.42

.43

.21

15.82

3.43

2.73

3.32

.77

*all t values above 1.96 statistically significant.

39

Table 10 Five Factor Model Results based on Market, Size Value, P/E and Leverage R pt − R ft = a + b(R mt − R ft ) + s(SMB t ) + p(LMH t ) + h (HML t ) + λ (LEVG t ) + e t Portfolio

a

B

s

h

p

l

t(a)

t(b)

t(s)

t(h)

t(p)

t(l)

Adj.R2

P1MC

-.0012

.942

.833

.300

.215

.05779

-.329

18.523

9.962

2.833

1.7780

.637

.825

P2MC

.0019

.989

.579

.124

.393

.0315

.379

14.707

5.237

.884

2.479

.262

.713

P3MC

.0022

1.01

.472

.213

.209

.140

.452

15.285

4.342

1.5

1.344

1.185

.721

P4MC

.0009

1.010

.329

.178

.186

.212

3.324

16.766

1.415

1.257

1.723

.748

.732

P5MC

-.0012

.942

-.167

.300

.214

.0579

-.329

18.523

-1.998

2.834

1.790

.636

.787

P1PE

-.0009

1.028

.45

.111

.857

.163

-.216

17.767

4.732

.920

6.283

1.571

.833

P2PE

.0034

.885

.407

.499

.0306

.0884

.832

15.917

4.448

4.313

.234

.889

.765

P3PE

.0012

0.962

.378

.309

.224

-.0601

.277

15.646

3.736

2.41

1.54

-.546

.724

P4PE

-.0002

.988

.358

.088

.241

.120

-.046

15.561

3.43

.668

1.609

1.056

.709

P5PE

.0001

1.028

.450

.111

-.143

.163

-.216

17.7767

4.732

.919

-1.049

1.571

.749

P1BEME

.0013

.987

.502

-.345

.220

.184

.319

17.046

5.278

-2.865

1.614

1.773

.736

P2BEME

.0019

1.031

.359

.0902

.260

-0.010

.428

16.731

3.542

.703

1.788

-.698

.727

P3BEME

.0031

.975

.306

.301

.313

-.0612

-.650

15.525

2.959

2.302

2.118

-.544

.726

P4BEME

.009

.914

.379

.416

.193

.183

.232

16.453

4.151

3.594

1.477

1.836

.7788

P5BEME

.0013

.987

.502

.655

.220

.184

.319

17.047

5.278

5.429

1.614

1.773

.831

P1DE

-.0010

.993

.426

.220

.3

-.434

-.234

16.649

4.343

1.772

2.134

-.4060

.725

P2DE

.0015

.953

.439

.083

.282

.0407

.347

15.969

4.479

.667

2.007

.381

.723

P3DE

.0013

.96

.326

.214

.235

.123

.292

16.26

3.35

1.736

1.692

1.159

.743

P4DE

.0019

.994

.428

.379

.0899

.180

.453

17.630

4.616

3.232

.678

1.786

.788

P5DE

-.0010

.993

.426

.22

.300

.566

-.234

16.649

4.343

1.773

2.135

5.302

.811

*all t values above 1.96 statistically significant.

40