S C M S J o u r n a l o f I n d i a n M a n a g e m e n t , April - June, 2 0 1 3.

90

Fama and French:

Three Factor Model
Renuka Sharma, Dr. Kiran Mehta

A
b
s
t
r
a
c
t

Fama and French (1993) elaborated the use of firm specific characteristics in explaining the return behaviour of
different types of portfolios and described that only market factor cannot describe the return behaviour of the stocks
in a significant manner but a blend of market factor with the size and book to market ratio jointly has more power
to elucidate the behaviour of stock returns. The Fama-French Three Factor Model compares returns of portfolio to
three distinctive types of risk found in the equity market to assist in categorizing returns. The present study is destined
to empirically test the three factor model suggested by Fama and French on Indian stock market and to document the
evidences how firm characteristics are used as a better explanation of stock return behaviour.

Key words: Portfolio, return behaviour, categorising returns, size premium, value premium

T
Dr. Renuka Sharma, Associate Professor,
Chitkara Business School, Chitkara University,
Punjab, Mb. 09416281206, 09467014669.
Email :- bhavyarenuka@rediffmail.com
Dr. Kiran Mehta,
Associate Professor in Finance,
Chitkara Business School,
Chitkara University, Punjab.
E-mail. ujjawalakiran@rediffmail.com

he past two decades of finance research have

strongly documented that the beta or systematic risk
of a security or portfolio cannot be relied to explain
the return behaviour of securities as well as returns of a
portfolio. Further evidences have been documented by
academic research done in this area indicating that certain
other parameters should also be added to an asset pricing
model which is based upon single factor like market risk. In
June 1992, Eugene F. Fama and Kenneth R. French documented
that on an average only 70% return of a portfolio are explained
by the beta of the portfolio and rest 30% returns are expressed
by some other factors. The Capital Asset Pricing Model
(CAPM) is extensively used by researchers to explain the
returns of a portfolio as well as for individual assets. Later on
this model was found to be limited use due to ignorance of
certain other fundamental parameters which also affect the
return behavior of the security or portfolio. Incorrectly
predicting results compared to realize returns and the affect
of other risk factors have put this model under criticism.
Ideally the inclusion of firm specific parameters when included
in an asset pricing model, which is based upon any single

A Quarterly Journal

S C M S J o u r n a l o f I n d i a n M a n a g e m e n t , April - June, 2 0 1 3

factor, must increase the explanatory power of that single

factor model. Fama and French (1993) elaborated the use of
firm specific characteristics in explaining the return behavior
of different types of portfolios. He extended the CAPM
model and described that only market factor cannot describe
the return behaviour of the stocks in a significant manner but
a blend of market factor with the size and book to market
ratio jointly have more power to explain the behaviour of
stock returns. The Fama-French Three Factor Model is used
to predict the risk and returns of equity portfolios. Its a
model that compares a portfolio to three distinctive types of
risk found in the equity market to assist in categorizing returns.
The return of any stock portfolio can be explained almost
entirely by two factors: Market cap (size) and book/market
ratio (value). The smaller and the median market cap of
your portfolio, the higher its expected return.
The present study is destined to empirically test the three
factor model suggested by Fama and French on Indian stock
market and to document the evidences how firm characteristics
are used as a better explanation of stock return behaviour.
Review of Literature
Fama and French (1992) provided a strong support to the
relationship between size and B/M ratio and stock returns. In
their univariate and multivariate tests they found a significant
positive relationship between B/M value and stock returns
and a negative relation between size and stock returns. They
studied the joint effect of beta, size, E/P ratio, leverage and
B/M ratio on the cross-sectional stock returns. Their results
showed that both size and B/M ratios were significant when
included together, and they dominated other variables. In their
study leverage and P/E ratio were significant by themselves
or when considered with size, but they became insignificant
when both size and B/M ratio were considered. The strong
size effect has also been documented in the succeeding works
of Fama and French. Lakonishok, Schleifer and Vishny (1994)
defined value strategies as buying shares with low prices
compared to some indicator of fundamental values such as
earnings, book value, dividend and cash flow. Glamour stocks
grew faster for the first couple of years but after that the
growth rates of the two groups were essentially the same.
Value strategies using both past low growth and low current
multiplies outperformed glamour strategies by an impressive
10-11% per year. Among the various measures of fundamental
values, P/E did not produce as large an effect as price-tobook value or price-to-cash flow, possibly because stocks
with temporarily depressed earnings were lumped together

91

with well performing glamour stocks in the high expected

growth/low E/P category. They found little support to the
view that value strategies were fundamentally riskier. Value
stocks outperformed glamour stocks quite consistently and
did particularly well in bad state of the world.
Fama and French (1995) studied the behaviour of stock prices,
in relation to the size and book-to-market-equity (BE/ME)
of companies listed New York Stock Exchange, American stock
exchange and NASDEQ for the period from 1963 to 1992.
They found, consistent with rational pricing, high BE/ME (a
low stock price relative to book value) indicated persistent
poor earnings and low BE/ME (a high stock price relative to
book value) indicated strong earnings. Within book-to-market
groups, small stocks tended to be less profitable than big
stocks. Berk (1995) argued that the market value of a firm
was inherently negatively related with its common stock
return and Berk (1996) reported that size effect disappeared
when some non market based size measure was used. He used
five different measures of firm size viz market capitalization
(MC), book value of total assets (BVA), book value of undepreciated property, plant and equipment (PPE), annual
sales value (sales) and number of employees in order to check
the existence of size effect.
Fama and French (1996) studied the value and growth stocks
in markets around world for the period 1975-1995. During
the study period the difference between average returns on
global portfolios of high and low book-to-market stocks was
7.60% per year and value stocks outperformed growth stocks
in 12 of 13 major markets. There were similar value premiums
when stocks were sorted on earnings/price, cash flow/price,
and dividend/price and found a value premium in emerging
markets. Since these results were out-of-sample relative to
earlier tests on U.S. data, they suggested that the return
premium for value stocks was real. An international CAPM
was not able to explain the value premium, but a two factor
model that included a risk. Kim (1997) examined the
explanatory power of beta, firm size, book-to-market equity
and the earning price ratio during the period July 1958 to
December 1993. The study found stronger support for the
betas pricing theory and concluded market betas had
economically and statistically significant force regardless of
the presence and absence of the firm size, book-to-market
equity and earning-price ratio. But unlike the firm size and
earning price, book to market had significant explanatory
power to average stock returns. In particular, firm size was
barely significant using monthly returns, but no longer

A Quarterly Journal

S C M S J o u r n a l o f I n d i a n M a n a g e m e n t , April - June, 2 0 1 3.

significant using quarterly returns. However, book-to-market

equity still had significant explanatory power for average stock
returns. Daniel and Titman (1997) confronted that firm
characteristics rather than factor loadings on the SMB and
HML portfolios determine expected returns. Within
portfolios formed on size, there was basically no relation
between returns and loadings on the SMB factor. This
suggested that expected stock returns were related to their
characteristics for reasons that might have nothing to do with
the covariance structure of returns and constituted evidence
against a financial distress interpretation of the SMB factor.
Connor and Sehgal (2001) examined the Fama-French threefactor model of stock returns for India using a sample of 364
companies from June 1989 to March 1999. They analyzed
whether the market, size and value factors were pervasive in
the cross-section of random stock returns and investigated
whether there were market, size and value factors in corporate
earnings similar to those in returns, and whether the common
risk factors in earnings translated in to common risk factor in
returns. They found evidence for pervasive market; size and
book-to-market factors in Indian stock returns and found
cross-sectional mean returns were explained by exposure to
these three factors and not by the market factor alone. They
found mixed evidences for parallel market, size and book-tomarket factors in earnings and did not find any reliable link
between the common risk factors in earnings and those in
stock returns. As a whole the empirical results were reasonably
consistent with the Fama-French Three-Factor model. Pandey
(2001) studied panel data set of 1729 observations to identify
variables that could explain expected returns of Malaysian
stocks. The study was based on the fixed effects regression
model as it performed better than the random effect model
and OLS models without the firm effects. Results of the fixedeffect univariate regression indicated that beta, size, bookto-market value (B/M) ratio, earning-price (E/P) ratio and
dividend yield individually played significant role in explaining
stock returns and payout and leverage had no effect. The
explanatory power of size (natural log of market
capitalization) was the highest. Beta was found to have
consistently a positive relation with stock returns by itself
and other variables. But this explanatory power was less than
size and other variables. Contrary to the results of Fama and
French (1992) B/M ratio was not persistently a significant
variable; its significance disappeared when they incorporated
size and E/P ratio in regression.

92

Drew, Naughton and Veeraraghavan (2003) studied firm size,

book-to-market equity and security returns on Shanghai Stock
Exchange (China) and tested multifactor approach to asset
pricing in one of the most challenging international market,
the Shanghai Stock Exchange, China for the period December,
1993 to December, 2000 by making various type portfolios.
The study concluded that mean-variance efficient investors
in china was able to select some combinations of small and
low book-to-market equity firms in addition to the market
portfolio to generate superior risk-adjusted returns. Moreover
they found no evidence to support the view that seasonal
effects could explain the findings of the multifactor model. In
summary, the study found the market factors alone was not
sufficient to describe the cross-section of average stock returns
in China. Gaunt (2004) studied the evidences of size effect,
BE/ME effect and the application of the Fama and French
factor model in the Australian market. He found that beta was
less than one which was contrary to Fama and French who
found beta to be close to one. Risk inclined to be greater for
smaller size firms and low BE/ME ratios like the findings of
Fama and French. There were evidences that there was a
monotonic increase in the HML factor loading from low to
high BE/ME portfolios implying that the HML factor played
a significant role in asset pricing. The author found an
inconsequential small firm effect and no large firm effect. He
found an improvement in the explanatory power of the three
factor model over and above the one factor CAPM when
compared to prior studies in the Australian setting. Vassalou
and Xing (2004) investigated the relation between the size
and book-to-market effects and default risk, defined as the
risk that a firm failed to service its debt obligations. The
authors estimated the default likelihood for up to 4,200 U.S.
firms over the period 1971-1999 on the basis of contingent
claims theory. The study showed that while the SMB and
HML factors contain some default-related information, default
risk could not account for the explanatory power of the FamaFrench three-factor model. Nartea and Djajadikerta (2005)
found a significant size effect and a weak BE/ME effect in the
case of New Zealand. According to them, the three factor
models explanatory power was not as big an improvement
over the CAPM as was for the Australian case. Sehgal and
Triphati (2005) examined the size effect in the Indian stock
market using data of top 482 Indian companies for the period
of 1990-2003. They found a strong size premium using six
alternative measures of company viz. - Market capitalization,
Enterprise value, Net Fixed Assets, Net Annual Sales, Total
assets and net working capital. Further the size based
investment strategy seemed to be economically feasible as it

A Quarterly Journal

S C M S J o u r n a l o f I n d i a n M a n a g e m e n t , April - June, 2 0 1 3

93

provided extra normal returns on risk adjusted basis. Frequent

rebalancing of size based portfolio was however fount to be
undesirable. The size effect did not seem to be owing to any
seasonality or business cycle factors. The presence of a strong
size premium also raised doubts the informational efficiency
of Indian Stock market. The authors found strong size effect
over the study period which had become more pronounced
during recent time period.

sampled duration. The companies listed under the head

Finance Related was excluded as these companies may differ
from the other companies in terms of their market cap and
other financial parameters. A few companies were further
eliminated due to non availability of data related to one or
more parameters (EP or Size). So the final sample consisted
of a total of 219 companies having consistent availability of
monthly data for their prices, EP ratio, & size factors.

Bhel (2006) studied the Fama and French three-factor model

of stock returns along with its variants, including the onefactor CAPM for 79 stocks listed on the BSE-100 stock
market index for India from July 2001 to June 2006. These
sample stocks were split in to six portfolios sorted on size
and book-to-market equity ratio. The factor portfolios that
explained the returns were the market factor, size factor (SMB)
and value factor (HML). The author found strong evidences
for the market factor in all the portfolios, it being regarded
with had the highest explanatory power. The SMB and HML
factors could not be clearly ranked in this regard. On the
basis of the adjusted R 2 it was confirmed that the three-factor
model captured better the common variations in the stock
returns than the CAPM. It was found that the three-factor
model of Fama-French fairs better in explaining the crosssection of returns in the portfolios than its variants and the
CAPM.

All the portfolios are constructed and revised on first trading

day of July. All the data required for the study under
consideration have been obtained from PROWESS database
provided by CMIE, Mumbai.

Data Inputs and Research Methodology

The Fama and French Three Factors Model

To represent the market cap (size) and book/market ratio
(value) returns, Fama and French modified the original
CAPM with two additional risk factors: Size risk and Value
risk.
The original CAPM equation:
i (E(R m) - R f))
E (R i) = R f +
Where R f is the risk free rate and E (R m) is the expected excess
return of the market portfolio beyond the risk-free rate, often
called the equity risk premium.
The Fama and French equation:

The present study has considered a sample of 219 stocks

which are listed in BSE 500 index. The monthly observations
have been considered for all parameters used in Fama and
French three factors model. As the base year of BSE-500
index was February 1999, so the study under consideration
has taken data from Feb 1999 to December 2007. This time
period has been studied in different time segments as Indian
stock market passed through different phases of market cycle
during the overall study period.


First Phase, July 1997-March 2000 (End of declining

trend and growth phase),

Second Phase, April 2000-March 2003 (Bear Market

Phenomenon), and

Third Phase, April 2003-December 2007 (Bull Market

Phenomenon).

Further, the companies were first shortlisted in consideration

with the availability of the data for a regular period of the

E (R i) = R f + i (E(R m) - R f) + s i SMB + h i HML)

Where SMB is the Small Minus Big market capitalization
risk factor and HML is the High Minus Low value premium
risk factor.
SMB, Small Minus Big, measures the additional return
investors have historically received by investing in stocks of
companies with relatively small market capitalization. This
additional return is often referred to as the size premium.
HML, which is short for High Minus Low, has been
constructed to measure the value premium provided to
investor for investing in companies with high book-to-market
values (essentially the value placed on the company by
accountants as a ratio relative to the value the public markets
placed on the company, commonly expressed as B/M).The
key point of the model is that it allows investors to weight
their portfolios so that they have greater or lesser exposure
to each of the specific risk factors, and therefore can target
more precisely different levels of expected return.

A Quarterly Journal

S C M S J o u r n a l o f I n d i a n M a n a g e m e n t , April - June, 2 0 1 3.

So in order to test the explanatory power of different variants

of Fama and French three factor model, the study under
consideration constructed the six size and value sorted
portfolios.
As size in terms of market cap and value premium in terms of
book to market ratio are the two additional parameters
considered in the Fama and French Three Factors Model to
improve basic CAPM equation therefore the six portfolios
constructed are of following types. S and B stands for the
size of the firm. S is used for small cap firms which are small
in size (market capitalization) and B is used for the large cap
firms having big size (market capitalization). L, M and H
stands for Low Book to Market ratio, Medium Book to
Market ratio and High Book to Market ratio.
SL: Portfolio having small cap stocks with low book to market
ratio
SM: Portfolio having small cap stocks with medium book to
market ratio
SH: Portfolio having small cap stocks with high book to
market ratio
BL: Portfolio having large cap stocks with low book to market
ratio
BM: Portfolio having large cap stocks with medium book to
market ratio
BH: Portfolio having large cap stocks with high book to market
ratio
Every year the portfolios were revised in the month of July.
Fama used these two factors to construct value and growth

94

portfolios and used excess returns of market proxy to model

the stock return behaviour. He found that the three factors,
market, size, and book to market ratio as the three main factors
to determine the insidious risk in the returns. The following
section has discussed the findings of the Fama and French
three factor model in detail.
Empirical Results
During the first phase of the study under consideration, it
was interesting to see that there was huge difference in the
mean returns of small and big sized portfolios. The mean
returns reported by SL portfolio were highest and BH
portfolio reported lowest mean returns with negative value.
Even the BM portfolio also reported near to zero mean
returns. No doubt the portfolios reporting the higher mean
returns also showed higher level of volatility in terms of their
standard deviation coefficients. SL & BL portfolios reported
negatively skewed distribution of their returns. In addition to
this, the kurtosis coefficient also reported leptokurtic curve
for SH portfolio, rest all five portfolios reported platykurtic
distribution curve. But the skewness and kurtosis coefficients
have not resulted into a biased distribution of returns as the
Jarque-Bera statistic reported normally distributed return
series of all the six portfolios. Further, during the first phase
the market and HML portfolios reported negative mean
returns and higher level of volatility in comparison to SMB
portfolio. But all the three factor portfolios showed normally
distributed return series during the first phase (see Table 1).
Figure 1, has depicted the patterns of mean return of all the
six size and value portfolios.

Table 1
Descriptive Statistics for All Portfolios under Fama and French Three Model (First Phase)
Portfolios

Mean

Median

Std. Dev.

Skewness

SL

0.0701

0.1254

0.1549

-0.3574

1.7891

0.7415 (0.6902)

SM

0.0377

0.022

0.1198

0.1386

2.8244

0.0404 (0.98)

SH

0.0286

-0.0224

0.1318

1.3353

4.0126

3.0589 (0.2167)

BL

0.015

0.0184

0.0933

-0.1425

1.6697

0.6941 (0.70680

BM

0.0056

-0.0236

0.1109

0.961

2.7736

1.4046 (0.4955)

BH

-0.0331

-0.0372

0.0988

0.7297

2.9712

0.799 (0.6706)

EXRET

-0.0399

-0.0102

0.0863

-1.071

2.7474

1.7445 (0.418)

SMB

0.0496

0.0658

0.0442

-0.5278

2.0848

0.732 (0.6935)

HML

-0.0448

-0.067

0.0968

-0.1196

1.4421

0.9316 (0.6276)

*Significant at 1% and **Significant at 5% level of Significance.

A Quarterly Journal

Kurtosis

Jarque-Bera (prob.)

S C M S J o u r n a l o f I n d i a n M a n a g e m e n t , April - June, 2 0 1 3

95

Figure 1
Mean Returns of All Portfolios under Fama and French Three Factor Model

The findings of the second phase were robust to see. During

the bearish phase, all the portfolios except BH reported
negative mean returns and lowest mean returns were reported
by SL portfolio followed by BL. As the BH portfolio reported
the highest and positive mean returns, it also showed highest
level of volatility followed by SL, SH, SM, BM, & BL
portfolios. All the portfolios showed negatively skewed
distribution of their monthly return series. The BM and BH
portfolios showed lesser peakedness in their monthly return
distribution curve than the normal curve. But despite the

evidences of less than normally distributed return series by

the skewness and kurtosis coefficient, the Jarque-Bera
statistic reported less than normally distributed mean return
series only for SL and BL portfolios. Rest all four portfolios
followed normal distribution in the distribution of their
monthly returns during the bear phase of the market. The
three factor portfolios also reported negative mean returns
but normally distributed returns during the bearish phase (see
Table 2). The pattern of mean returns of the six portfolios
during the bearish phase can be easily observed through
Figure2.

Table 2
Descriptive Statistics for All Portfolios under Fama and French Three Model (Second Phase)
Portfolios

Mean

Median

Std. Dev.

Skewness

SL

-0.0297

-0.0087

0.1104

-0.924

3.9438

6.4588** (0.0396)

SM

-0.0015

0.0019

0.0943

-0.0476

3.5802

0.5185 (0.7716)

SH

-0.0021

0.0053

0.1038

-0.0176

4.0976

1.8088 (0.4048)

BL

-0.0195

-0.013

0.0708

-1.4292

6.0377

26.097* (0.000)

BM

-0.0039

-0.0062

0.0862

-0.1671

2.3945

0.7175 (0.6986)

BH

0.0048

0.0188

0.119

-0.4613

2.7875

1.3446 (0.5105)

EXRET

-0.0893

-0.0767

0.088

-0.8328

3.252

4.2564 (0.1191)

SMB

-0.0049

0.0009

0.0345

0.05691

2.5194

0.3659 (0.8328)

HML

0.02599

0.02416

0.0701

-0.0435

3.7056

0.7582 (0.6845)

*Significant at 1% and **Significant at 5% level of Significance.

A Quarterly Journal

Kurtosis

Jarque-Bera (prob.)

S C M S J o u r n a l o f I n d i a n M a n a g e m e n t , April - June, 2 0 1 3.

96

Figure 2
Mean Returns of All Portfolios under Fama and French Three Factor Model

During the bull phase, some new evidences were depicted

through the distribution of mean returns for various portfolios
(Table 3). Although all the three small sized portfolios
performed better than all big-sized portfolios, the highest
mean returns were reported by SH portfolio instead of SL
portfolio in case first phase of rising market. All the six
portfolios reported negatively skewed distribution of their
monthly returns during the third phase. The peakedness of
the distribution curve of monthly returns was more than
normal curve in case of all the portfolios except for SH

(platykurtic). But the Jarque-Bera statistic reported less than

normally distributed returns for SL, BL, & BM portfolios.
The mean returns distribution of all the six size and value
sorted portfolios can be seen through Figure 3. In addition to
this, during the third phase, the market factor portfolio
continued to show negative mean returns with highest
volatility and SMB factor portfolio reported highest mean
returns with lowest volatility. Only the market factor
portfolio depicted less than normal distribution of monthly
returns during the bull phase.

Table 3
Descriptive Statistics for All Portfolios under Fama and French Three Model (Third Phase)
Portfolios

Mean

Median

Std. Dev.

Skewness

SL

0.053

0.0693

0.0846

-1.0692

4.2054

14.311* (0.0008)

SM

0.0544

0.0667

0.087

-0.7278

3.391

5.3948 (0.0674)

SH

0.0605

0.0736

0.1039

-0.203

2.7123

0.5882 (0.7452)

BL

0.033

0.0456

0.0615

-0.9266

4.3256

12.331* (0.0021)

BM

0.0328

0.0443

0.0747

-0.6438

4.0915

6.7673** (0.0339)

BH

0.0371

0.0433

0.0879

-0.0427

3.366

0.3355 (0.8456)

EXRET

-0.0207

-0.0133

0.0697

-0.7813

4.6021

11.896* (0.0026)

SMB

0.0217

0.0204

0.0305

0.20702

3.116

0.4391 (0.8029)

HML

0.0059

4.8E-05

0.0482

0.4491

3.3268

2.1697 (0.338)

*Significant at 1% and **Significant at 5% level of Significance.

A Quarterly Journal

Kurtosis

Jarque-Bera (prob.)

S C M S J o u r n a l o f I n d i a n M a n a g e m e n t , April - June, 2 0 1 3

97

Figure 3
Mean Returns of All Portfolios under Fama and French Three Factor Model

Implication of Fama and French Three Factor Model

during First Phase
Table 4 has reported the outputs of all the variants of Fama
and French model during the first phase of the study under
consideration. The findings were really attention-grabbing.
The market factor has not explained the variations in the
return of all six size and value sorted portfolios. For big sized
portfolios, the adj. R 2 was found even less than 10% when
regressed on the EXRET. The highest adj. R 2 was found in
case of BL portfolio explaining 43.53% variability in the
monthly returns. The second factor alone, i.e., SMB showed

comparatively better power to explain the risk involved in

the monthly returns of all portfolios. The SMB factor was
found a better parameter to explain the return behaviour of
portfolios with low book to market ratio (SL & BL). But for
rest of four portfolios it showed more power than the EXRET
factor but still having adj. R 2 with lower value. When the
other variant of the model having HML factor alone was
observed, it showed even more fascinating power to explain
the return behaviour of various portfolios.
The magnitude of adj. R 2 was found very low and even with
some negative signs (see Table) indicating that this factor

Table 4 Fama and French Three Factor Model (First Phase)

Explanatory
Variable

Dependent
Variable

SMB

HML

Adj R2

EXRET

S/L
Std.Error
S/M
Std.Error
S/H
Std.Error
B/L
Std.Error
B/M
Std.Error
B/H
Std.Error

0.0277
0.0436
-0.015
0.0331
-0.0253
0.0409
-0.0466
0.0259
-0.0581
0.0367
-0.1056
0.0354

1.2455
0.4815
0.9903
0.3655
0.9598
0.4511
0.7637
0.2853
0.7136
0.4045
0.4919
0.391

0.4156

0.4421

0.3059

0.4353

0.2089

0.0679

S/L
Std.Error
S/M
Std.Error
S/H
Std.Error

-0.1811
0.0329
-0.1613
0.0399
-0.165
0.0534

3.2072
0.5085
2.1553
0.6159
2.046
0.824

0.829

0.5844

0.3923

SMB

A Quarterly Journal

S C M S J o u r n a l o f I n d i a n M a n a g e m e n t , April - June, 2 0 1 3.

Explanatory Dependent
Variable
Variable

HML

EXRET-SMB

EXRET-HML

SMB-HML

B/L
Std.Error
B/M
Std.Error
B/H
Std.Error
S/L
Std.Error
S/M
Std.Error
S/H
Std.Error
B/L
Std.Error
B/M
Std.Error
B/H
Std.Error
S/L
Std.Error
S/M
Std.Error
S/H
Std.Error
B/L
Std.Error
B/M
Std.Error
B/H
Std.Error
S/L
Std.Error
S/M
Std.Error
S/H
Std.Error
B/L
Std.Error
B/M
Std.Error
B/H
Std.Error
S/L
Std.Error
S/M
Std.Error

b
-0.1649
0.0275
-0.1638
0.0482
-0.1787
0.048
-0.0597
0.0518
-0.0601
0.0472
-0.0498
0.0512
-0.095
0.0334
-0.0925
0.0438
-0.1049
0.0352
-0.2022
0.0644
-0.1261
0.0771
-0.1252
0.1039
-0.1521
0.054
-0.139
0.0947
-0.1625
0.0946
-0.006
0.0465
-0.0094
0.0409
0.0057
0.0439
-0.0604
0.03
-0.0566
0.0455
-0.0721
0.0343
-0.1808
0.0327
-0.1617
0.0389

98

SMB

1.7709
0.4239
1.5578
0.7443
1.0798
0.7401
-

-0.1803
0.4609
0.3007
0.5519
0.3398
0.7436
0.1097
0.3865
0.2115
0.6778
0.139
0.677
1.0776
0.4608
1.0177
0.4051
1.1141
0.4346
0.6951
0.2975
0.7208
0.4509
0.6587
0.3393
-

3.488
0.8998
1.6868
1.0773
1.5167
1.4516
1.6
0.7545
1.2283
1.3231
0.8632
1.3216
-

A Quarterly Journal

2.9577
0.5563
2.4819
0.6631

HML

Adj R2

0.6729

0.2971

0.1236

-0.8411
0.5095
-0.1272
0.4641
0.3064
0.5031
-0.4002
0.3288
-0.134
0.4306
0.4523
0.3462
-

0.1774
-0.1307
-0.0854
0.0567
-0.1273
0.0811
0.8055

0.538

0.3149

0.6234

0.193

-0.0153

-0.6022
0.411
0.0985
0.3613
0.5534
0.3876
-0.2461
0.2654
0.0258
0.4021
0.5983
0.3026
-0.2693
0.2542
0.3526
0.303

0.4979
0.3571
0.3956
0.4238
0.0777
0.3416
0.832
0.6044

S C M S J o u r n a l o f I n d i a n M a n a g e m e n t , April - June, 2 0 1 3

Explanatory
Variable

EXRET-SMB
-HML

99

Dependent
Variable

SMB

HML

Adj R2

S/H
Std.Error
B/L
Std.Error
B/M
Std.Error
B/H
Std.Error
S/L
Std.Error
S/M
Std.Error
S/H
Std.Error
B/L
Std.Error
B/M
Std.Error
B/H
Std.Error

-0.166
0.0363
-0.1648
0.0294
-0.164
0.051
-0.1796
0.03
-0.1936
0.0659
-0.1372
0.0779
-0.1533
0.0732
-0.1493
0.0591
-0.1453
0.1029
-0.1895
0.0606

2.8379
0.6174
1.7057
0.5013
1.7464
0.8684
1.8255
0.5106
3.1376
0.9874
2.1369
1.1669
2.6603
1.097
1.4872
0.8859
1.4831
1.5422
1.9645
0.9076

0.855
0.2821
-0.0705
0.2291
0.2036
0.3967
0.8052
0.2333
-0.2588
0.2807
0.3324
0.3317
0.8446
0.3119
-0.0833
0.2519
0.1882
0.4385
0.8133
0.258

0.7199

-0.1092
0.4731
0.2094
0.5592
0.1078
0.5257
0.1326
0.4245
0.1598
0.739
-0.0844
0.4349

could not be relied during the first phase of rising market to

study the variability in the return of the various portfolios.
Thus the above discussion concluded that none of the three
factors stated above, when considered individually, could
explain the variability in the mean return behaviour of all the
six portfolios. The following paragraph will discuss the
changes in the power of coefficient of determination when
two factors were considered jointly to measure the level of
risk in the mean returns of various portfolios.
The explanatory power of EXRET increased when SMB factor
was combined with the regression model to study the variation
in the returns of various portfolios. As depicted in the table,
both these factors explained 80.55% variability for SL
portfolio, but still it has not explained the cause of variability
in big sized portfolios significantly. The power of EXRET
and HML factor was found comparatively better in explaining
the variability in the returns of small-sized portfolios (adj.
R 2 ranged from 35.71% to 49.79%) rather than for big-sized
portfolios (adj. R 2 ranged from 7.77% to 42.38%). When two
factors were jointly taken to determine the return variability
of all the portfolios, then SMB & SML factors jointly showed
greater power than other two factors stated above.
The SMB & HML factors explained 21.44% to 83.2%
variability in the monthly returns of the all size and value

0.6243
0.2144
0.6575
0.8005
0.5383
0.6667
0.5578
0.066
0.5921

sorted portfolios. Further, these two factors were better in

explaining the return behaviour of small sized portfolios in
comparison to large ones.
In addition to this, the combination of all the three factors
(EXRET, SMB, & HML) jointly produced better results but
less significant thos the explained by SMB & HML factors
jointly. All the three factors jointly were found very poor in
explaining the variability in the returns of BM portfolio. The
highest value of adjusted R 2 was reported in case of SL
portfolio. Hence it can be concluded that during the first phase
of the market, the size and book to market ratio jointly
explained the risk involved in the return series of all the size
and value sorted portfolios in a better way than other variants
of the Fama and French Model.
Implication of Fama and French Three Factor Model
during Second Phase
During the bear phase of the market, all the variants of the
Fama and French three factor model showed different results
than the first phase. All these findings are depicted in Table
5. As reported in the Table, the EXRET factor alone showed
good explanatory power to explain the cause of variability in
the monthly returns of various portfolios. The adjusted R 2
co-efficient ranged from 57.86% to 82.91%.

A Quarterly Journal

S C M S J o u r n a l o f I n d i a n M a n a g e m e n t , April - June, 2 0 1 3.

100

5 Fama and French Three Factor Model (Second Phase)

Explanatory
Variable

Dependent
Variable

EXRET

S/L
Std.Error
S/M
Std.Error
S/H
Std.Error
B/L
Std.Error
B/M
Std.Error
B/H
Std.Error
S/L
Std.Error
S/M
Std.Error
S/H
Std.Error
B/L
Std.Error
B/M
Std.Error
B/H
Std.Error
S/L
Std.Error
S/M
Std.Error
S/H
Std.Error
B/L
Std.Error
B/M
Std.Error
B/H
Std.Error
S/L
Std.Error
S/M
Std.Error
S/H
Std.Error
B/L
Std.Error
B/M
Std.Error
B/H
Std.Error

0.026701
0.01787
0.01134
0.011974
0.007439
0.01648
-0.02266
0.007477
0.001515
0.010609
0.026701
0.01787
-0.06864
0.020372
-0.06956
0.015565
-0.0715
0.01754
-0.08989
0.012457
-0.07602
0.01515
-0.06864
0.020372
-0.05969
0.017739
-0.08565
0.016926
-0.09524
0.016556
-0.09376
0.013658
-0.09072
0.01469
-0.09354
0.017739
0.0319
0.0165
0.0093
0.0118
0.0069
0.0168
-0.0225
0.0076
0.0041
0.0101
0.0319
0.0165

SMB

HML

EXRET-SMB

SMB

HML

Adj R2

0.607298

0.73989

0.578686

0.829179

0.751981

0.607298

-0.02927

0.113571

0.03743

0.043687

-0.02013

-0.02927

-0.04001
0.593208
1.061388
0.453227
0.784806
0.510757
0.584755
0.362727
0.245418
0.44115
-0.04001
0.593208
-

0.3021

1.1763
0.1386
0.9228
0.0994
0.9177
0.1407
0.7872
0.0638
0.9374
0.0847
1.1763
0.1386

-0.9402
0.3538
0.3551
0.2538
0.0824
0.3592
-0.0177
0.1630
-0.4720
0.2163
-0.9402
0.3538

0.965664
0.240289
0.418669
0.229282
0.765128
0.224262
0.038644
0.185009
0.519168
0.198985
0.965664
0.240289
-

1.065866
0.143557
0.964567
0.096188
0.927418
0.132389
0.78517
0.060062
0.882032
0.085222
1.065866
0.143557
-

A Quarterly Journal

0.062524
0.233131
-0.02809
0.14231
0.3021
0.6667

0.7470

0.5666

0.8240

0.7766

0.6667

S C M S J o u r n a l o f I n d i a n M a n a g e m e n t , April - June, 2 0 1 3

Explanatory
Variable
EXRET-HML

Dependent
Variable

S/L
Std.Error
S/M
Std.Error
S/H
Std.Error
B/L
Std.Error
B/M
Std.Error
B/H
Std.Error
SMB-HML
S/L
Std.Error
S/M
Std.Error
S/H
Std.Error
B/L
Std.Error
B/M
Std.Error
B/H
Std.Error
S/L
Std.Error
S/M
Std.Error
S/H
Std.Error
EXRET-SMB- B/L
HML
Std.Error
B/M
Std.Error
B/H
Std.Error

101

B
0.0013
0.0079
0.0004
0.0102
-0.0126
0.0108
-0.0234
0.0078
-0.0120
0.0064
0.0013
0.0079
-0.0942
0.0170
-0.0869
0.01390
-0.0965
0.0132
-0.0943
0.0131
-0.0913
0.0140
-0.0942
0.0171
0.0012
0.0083
-0.0104
0.0077
-0.0246
0.0079
-0.0236
0.0082
-0.0124
0.0067
0.0012
0.0083

SMB
-

1.0588
0.0613
0.9615
0.0796
0.9219
0.0843
0.7849
0.0608
0.8783
0.0500
1.0588
0.0613
-

1.0171
0.5225
1.7769
0.4260
1.8171
0.4063
0.7668
0.4033
0.8774
0.4306
1.0171
0.5225
0.0001
0.1879
0.9635
0.1747
1.0519
0.1796
0.0145
0.1864
0.0381
0.1531
0.0001
0.1879

1.0588
0.0661
0.8468
0.0615
0.7966
0.0632
0.7832
0.0656
0.8737
0.0539
1.0588
0.0661

Among the big sized portfolios, it significantly explained the

variability in the returns of BL and BM portfolios. During
the bear market the investors have not relied on SMB and
HML factor as much as they relied on the market factor. The
SMB & HML factors individually showed very lesser
explanatory power to determine the variability in the returns
of all size and value sorted portfolios. The SMB factor had
less than 5% explanatory power in case of five out of six

HML
0.9534
0.0769
0.4075
0.0999
0.7545
0.1059
0.0295
0.0763
0.5090
0.0628
0.9534
0.0769
1.1852
0.2570
0.8022
0.2095
1.1573
0.1998
0.2041
0.1983
0.7085
0.2118
1.1852
0.2570
0.9534
0.0882
0.6168
0.0820
0.983
0.0843
0.0327
0.0874
0.5173
0.0718
0.9534
0.0882

Adj R2
0.9283
0.8217
0.8289
0.8248
0.9145
0.9283
0.3549
0.3675
0.5080
0.0453
0.2150
0.3549
0.9261
0.9057
0.9148
0.8193
0.9120
0.9261

portfolios. On the other hand, the HML factor alone has not
explained more than 30% variability in the returns.
But a significant improvement could be noticed in the
explanatory power of the regression models when two factors
jointly were considered to study the risk in the monthly
returns of various portfolios. As depicted in Table, when
EXRET and SMB factors were taken together, there was an
increase in the value of adjusted R2 for all the portfolios. But

A Quarterly Journal

S C M S J o u r n a l o f I n d i a n M a n a g e m e n t , April - June, 2 0 1 3.

this increase was very high when the joint effect of EXRET
and HML was studied. The explanatory power of the model
increased with a huge difference. The R 2 was found more than
80% for three portfolios and more than 90% for other three
portfolios. It showed that the market factor and book to market
ratio jointly explained the variability in the monthly returns
of all the portfolios in a more powerful manner.
However the findings obtained through joint effect of SMB
and HML factors were not found significant enough to explain
the risk involved in the cross section of mean returns. Both
these factors individually and jointly had poor power to
explain the deviations in the return series of various
portfolios. But the findings obtained through the Fama and
French three factor variant was very impressive. In exception
with BL portfolio (81.93%), it explained more than 90%
variation in rest all the portfolios. The three factors taken

102

together in the regression model definitely increased its power

to determine the deviations in the monthly returns of all the
six size and value sorted portfolios. Therefore it can be stated
that during the bear phase also the three factors taken together
in the regression model will definitely prove better than other
variants of the mode to explain the variations in the returns
of various portfolios.
Implication of Fama and French Three Factor Model
during Third Phase
Table 6 has reported the findings of all variants of Fama and
French three factor model for the third phase of the study
under consideration. When all the factors were regressed
individually, the market factor (EXRET) reported better
explanatory power than other two factors (SMB & HML).
As reported in the Table, when EXRET factor was taken
alone, the adjusted R 2 ranged from 56.58% to 89.68%.

Table 6 Fama and French Three Factor Model (Third Phase)

Explanatory
Variable

Dependent
Variable

EXRET

S/L
Std.Error
S/M
Std.Error
S/H
Std.Error
B/L
Std.Error
B/M
Std.Error
B/H
Std.Error
S/L
Std.Error
S/M
Std.Error
S/H
Std.Error
B/L
Std.Error
B/M
Std.Error
B/H
Std.Error

SMB

B
0.002
0.0067
0.0201
0.0061
0.0271
0.0096
-0.007
0.0028
-0.004
0.0042
0.0267
0.0179
-0.046
0.0132
-0.042
0.0114
-0.047
0.013
-0.041
0.0095
-0.045
0.0116
-0.069
0.0204

SMB

1.0641
0.0928
1.1095
0.0844
1.1478
0.1335
0.8438
0.0382
1.0053
0.0578
1.0659
0.1436
-

HML

Adj R2

0.6997

0.7542

0.5658

0.8968

0.8433

0.6073

1.2101
0.3552
1.8103
0.3079
2.3086
0.3487
0.7613
0.2546
0.9554
0.3115
-0.04
0.5932

0.1593

0.3748

0.4334

0.1242

0.1306

-0.029

A Quarterly Journal

S C M S J o u r n a l o f I n d i a n M a n a g e m e n t , April - June, 2 0 1 3

Explanatory
Variable

Dependent
Variable

HML

S/L
Std.Error
S/M
Std.Error
S/H
Std.Error
B/L
Std.Error
B/M
Std.Error
B/H
Std.Error
S/L
Std.Error
S/M
Std.Error
S/H
Std.Error
B/L
Std.Error
B/M
Std.Error
B/H
Std.Error
S/L
Std.Error
S/M
Std.Error
S/H
Std.Error
B/L
Std.Error
B/M
Std.Error
B/H
Std.Error
S/L
Std.Error
S/M
Std.Error
S/H
Std.Error
B/L
Std.Error
B/M
Std.Error
B/H

EXRET-SMB

EXRET-HML

SMB-HML

103

B
-0.06
0.0092
-0.007
0.011
-0.005
0.0104
-0.026
0.0082
-0.028
0.0095
-0.094
0.0177
-0.01
0.0083
-0.008
0.0052
-0.013
0.009
-0.011
0.0035
-0.01
0.0052
0.032
0.0166
-0.006
0.0038
0.0167
0.0057
0.0164
0.0059
-0.007
0.0028
-0.006
0.0038
0.0013
0.0079
-0.042
0.0109
-0.04
0.0112
-0.042
0.0097
-0.04
0.0095
-0.043
0.0112
-0.094

SMB

HML

0.9956
0.0937
0.9565
0.0587
0.9246
0.1016
0.819
0.039
0.9685
0.0591
1.1763
0.1386
0.9324
0.0526
1.0524
0.0801
0.9657
0.0821
0.8452
0.0395
0.9608
0.0533
1.0589
0.0613
-

0.5088
0.2142
1.1365
0.1342
1.6573
0.2324
0.1844
0.0892
0.2733
0.1352
-0.94
0.3538
-

A Quarterly Journal

0.7272
0.3055
1.6191
0.314
1.7584
0.2719
0.699
0.2679
0.7378
0.3144
1.0172

Adj R2

1.1553
0.1919
0.7102
0.2293
1.4931
0.2165
0.264
0.17
0.5986
0.1973
0.9657
0.2403
-

0.3862
0.133
0.454
0.0246
0.1278
0.3021
0.7231

0.8925

0.7723

0.9026

0.8517

0.6667

0.8546
0.0761
0.3708
0.1159
1.1816
0.1187
-0.009
0.0572
0.2887
0.0771
0.9535
0.077
1.0173
0.1932
0.4027
0.1986
1.1592
0.172
0.1313
0.1695
0.4585
0.1989
1.1852

0.9083
0.7896
0.844
0.895
0.8733
0.9284
0.4342
0.4083
0.6866
0.1178
0.1938
0.355

S C M S J o u r n a l o f I n d i a n M a n a g e m e n t , April - June, 2 0 1 3.

Explanatory
Variable

Dependent
Variable
Std.Error
S/L
Std.Error
S/M
Std.Error
S/H
Std.Error
B/L
Std.Error
B/M
Std.Error
B/H
Std.Error

B
0.0171
-0.01
0.0047
-0.008
0.0049
-0.013
0.0038
-0.011
0.0035
-0.01
0.0048
0.0013
0.0083

104

SMB

0.9131
0.054
0.9357
0.0565
0.8262
0.0435
0.8229
0.0396
0.9423
0.0549
1.0589
0.0662

The market factor was found a better way to explain the

variability in the average returns of the large sized portfolios
in comparison to small ones. The second factor, SMB, was
found poor in explaining the deviations in the mean returns of
all the six portfolios and even poorer for large sized portfolios.
The adjusted R 2 reported by HML factor showed more power
to explain the deviations in the returns but these were much
lesser than market factor. Therefore none of three factors
individually found a powerful tool to explain the risk involved
in the cross section of average returns. The results obtained
through the two factors taken jointly were better than
individual factor based models. Amongst the three variants
of, EXRET-SMB, EXRET-HML, & SMB-HML, EXRETHML based regression model gave best explanation for the
variability in the cross section of average returns. As depicted
in the Table, the EXRET-SMB regression model showed the
value of adjusted R2 coefficient in the range of 66.67%-90.26%.
The higher range was found in case of small-sized portfolios.
On the other hand, the EXRET-HML model explained the
variations in the mean returns of all the portfolios in the
range of 78.96% - 92.84%, which was found significant enough
and could be relied on for all the size and value sorted
portfolios. But the power of SBM-HML regression model
was very fragile in comparison to above two variants as it has
not explained the variation in the mean returns by more than
68.66% for small sized portfolios and by 35.5% for big sized
portfolios.
In addition to this, the relevance and importance of Fama and
French three factors was further proved when the results of
regression model based upon these three factors jointly were

0.5226
0.1739
0.1262
1.0522
0.132
1.2577
0.1017
0.2004
0.0926
0.1669
0.1282
0.0001
0.1879

HML
0.2571
0.8278
0.0779
0.2086
0.0815
0.9878
0.0628
-0.039
0.0572
0.263
0.0792
0.9535
0.0882

Adj R2

0.9098
0.9025
0.9591
0.9017
0.8749
0.9261

observed. As depicted in the Table, it explained 95.91%

variations in the returns of SH portfolio, 92.61% for BH
portfolio. It was only in case of BM portfolio that adjusted
R 2 was found 87.49% but for rest of the portfolios it explained
more than 90% risk involved in the mean returns. Hence the
magnitude of explanatory power of three factor model was
found much more than other variants of the model.
To conclude it can be said that in case of a consistent rising
market also, instead of considering any single or two factors,
all the three factors regressed jointly will be really helpful to
the investor in studying the return behaviour of various
portfolios.
References
Berk, J. (1995), A Critique of Size Related Anomalies,
Review of Financial Studies, 1995, 8, pp.275-286.
Berk, J. (1996), An Empirical Re-examination of the Relation
Between Firm Size and Return, Revised Working
Paper, School of Business Administration, University
of Washington, October 1996.
Daniel, K., and S. Titman, (1997), Evidence on the
Characteristics of Cross-Sectional Variation in Stock
Returns. Journal of Finance, 52, pp. 1-33.
Drew E. Michall, Naughton, Tony and Veeraraghavan, Madhu
(2003), Firm Size, Book-to-market equity and
security returns: Evidence from the Shanghai Stock
Exchange, Australian Journal of Management, Vol.
28, No. 2, September, pp. 119-140.

A Quarterly Journal

S C M S J o u r n a l o f I n d i a n M a n a g e m e n t , April - June, 2 0 1 3

105

Fama, F. Eugene and French R. Kenneth, (1992), The CrossSection of Expected Stock Returns. Journal of
Finance, Vol. 47, pp. 427 465.

Section of Expected Stock Return. The Journal of

Financial and Quantitative Analysis, Vol. 32, No. 4,
December, pp. 463-489.

Fama, F. Eugene and French R. Kenneth, (1993), Common

Risk Factors in the Returns on Stocks and Bonds.
Journal of Financial Economics, 33, 3 65.

Lakonishok, J., A. Shleifer, and R.W. Vishny, (1994),

Contrarian Investment, Extrapolation, and Risk.
Journal of Finance, 49, 1541 1578.

Fama, F. Eugene and French R. Kenneth, (1995), Size and

Book-to-Market Factors in Earnings and Returns,
Journal of Finance, 50, 131 155.

Pandey, I.M. Mohan (2001), The Expected Stock Returns

Of Malaysian Firms: A Panel Data Analysis. IIMA
Working Paper No. 2001-09-01, September, Available
at SSRN: http://ssrn.com/abstract=299913 or DOI:
10.2139/ssrn.299913.

Fama, F. Eugene and French R. Kenneth, (1996), Multifactor

Interpretations of Asset Pricing Anomalies. Journal
of Finance, 51, 55 84.
Gaunt, Clive (2004), Size and Book to Market Effects and
the Fama French Three Factor Asset Pricing Model:
Evidence from the Australian Stock market .
Accounting and Finance, Vol. 44, March, pp. 27-44.
Kim, Dongcheol (1997), A Re-examination of Firm Size,
Book-to-Market and Earnings Price in the Cross-

Sehgal S and Tripathi V (2005) Size Effect in Indian Stock

Market: Some Empirical Evidence, Vision-The Journal
of Business Perspective, Vol. 9, No. 4, Oct- Dec 2005,
27-42.
Vassalou, M., and Y. Xing, (2004), Default Risk in Equity
Returns. Journal of Finance, Vol. 59, pp. 831-868.

A Quarterly Journal

Reproduced with permission of the copyright owner. Further reproduction prohibited without
permission.