Market anomalies, asset pricing models,

and stock returns: evidence from the

Indian stock market
Saumya Ranjan Dash and Jitendra Mahakud

Saumya Ranjan Dash is Abstract

based at Indian Institute Purpose – This paper aims to investigate whether the use of conditional and unconditional Fama and
of Management, Indore, French (1993) three-factor and Carhart (1997) four-factor asset pricing models (APMs) captures the role
India. Jitendra Mahakud of asset pricing anomalies in the context of emerging stock market like India.
is based at Indian Design/methodology/approach – The first step time series regression approach has been used to
Institute of Technology, drive the risk-adjusted returns of individual securities. For examining the predictability of firm
characteristics or asset pricing anomalies on the risk-adjusted returns of individual securities, the panel
Kharagpur, India.
data estimation technique has been used.
Findings – Fama and French (1993) three-factor and Carhart (1997) four-factor model in their
unconditional specifications capture the impact of book-to-market price and liquidity effects
completely. When alternative APMs in their conditional specifications are tested, the importance of
medium- and long-term momentum effects has been captured to a greater extent. The size, market
leverage and short-term momentum effects still persist even in the case of alternative unconditional and
conditional APMs.
Research limitations/implications – The empirical analysis does not extend for different market
scenarios like high and low volatile market or good and bad macroeconomic environment. Because of
the constraint of data availability, the authors could not include certain important anomalies like net
operating assets, change in gross profit margin, external equity and debt financing and idiosyncratic
risk.
Practical implications – Although the active investment approach in stock market shares a common
ground of semi-strong form of market efficiency hypothesis which also supports the presence of asset
pricing anomalies, less empirical evidence has been explored in this regard to support or repute such
belief of practitioners. Our empirical findings make an attempt in this regard to suggest certain
anomaly-based trading strategy that can be followed for active portfolio management.
Originality/value – From an emerging market perspective, this paper provides out-of-sample
empirical evidence toward the use of conditional Fama and French three-factor and Carhart four-factor
APMs for the complete explanation of market anomalies. This approach retains its importance with
respect to the comprehensiveness of analysis considering alternative APMs for capturing unique
effects of market anomalies.
Keywords Market anomaly, Emerging market, Stock return, Momentum, Asset pricing model,
Risk factor
Paper type Research paper

1. Introduction
Identification and empirical validation of factors which explain the cross-sectional variation
of expected stock returns has been one of the key issues in the investment management
research. Over the past two decades, there has been considerable research to support
that the predictable component of stock returns can be associated with several firm
characteristics or commonly perceived market anomalies. The existence of anomaly effect
as cross-sectional determinants of stock returns negates the cross-sectional relation
Received 20 June 2014
Revised 9 December 2014
between average stock returns and systematic risk measured by alternative asset pricing
Accepted 20 March 2015 models (APMs hereafter). Interpretation of such empirical evidences across developed and

PAGE 306 JOURNAL OF ASIA BUSINESS STUDIES VOL. 9 NO. 3, 2015, pp. 306-328, © Emerald Group Publishing Limited, ISSN 1558-7894 DOI 10.1108/JABS-06-2014-0040
emerging markets around the world is, of course, strongly debated. The present paper
attempts to shed more light on this debate by investigating whether cross-section of
expected stock returns is better explained by risk factors suggested by alternative APMs,
or market anomalies. More importantly, we seek to examine the use of Fama and French
(1993) three-factor and Carhart (1997) four-factor APMs in their alternative specification for
the complete explanation of market anomalies.
The role of firm characteristics or financial market anomalies for the determination of
cross-section of stock returns behavior (Amihud, 2002; Banz, 1981; Jegadeesh and
Titman, 1993; Stattman, 1980) has long been recognized as a challenge to the central
paradigm of the traditional asset pricing literature. In common, related literature suggests
two competing arguments on the cross-sectional regularity of such anomalies. First, the
apparent role of firm characteristics is the resultant of market inefficiency or profit
opportunities, and they are mere chance results that often seem to disappear after
documented in the academic literature (Dimson and Marsh, 1999). Second, the
documented firm characteristics serve as proxies for the riskiness of firms and, therefore,
serve as the better determinant of cross-section of stock returns behavior (Fama and
French, 1992; Chan and Chen, 1991). However, the debate between the two competing
arguments is still open and has drawn considerable attention in the investment
management research.
Following the risk-based argument of market anomalies, the available literature conjecture
that the identification and empirical validation of financial market anomalies have been
indeed a joint hypothesis test of informational efficiency of the market and predictability of
APMs (Schwert, 2003; MacKinlay, 1995). Rejection of the tests suggest that either the
market examined is not efficient, i.e. market anomalies are mere chance results (Schwert,
2003) appeared due to methodological illusions (Fama, 1998), or the equilibrium APMs fail
to properly describe stock returns behavior due to its inappropriateness, i.e. bad model
problem (Fama, 1998; MacKinlay, 1995). However, if the joint hypothesis is rejected, it still
fails to give a conclusive stand to attribute such rejection to either any one of the
aforementioned hypotheses. For instance, testing a three-factor model (Fama and French,
1993) for explaining the cross-sectional regularity of firm size (MC) and book-to-market
price (BP) ratio may lead to a situation where the proposed three-factor model may not be
able to capture the MC or BP effects, even though the proposed three-factor model
contains risk factors (small minus big, i.e. SMB, and high minus low, i.e. HML) that are
empirically motivated from the risk-based argument of MC and BP. In such a scenario, it will
be unconventional to argue that persistence of MC and BP effects are mere chance results
or methodological illusions (Schwert, 2003; Fama, 1998), and premature to discard the
claim that the systematic risk factors involved in the three-factor model (Fama and French,
1993) are inappropriate and subject to the bad model problem (Fama, 1998). Similar line
of argument can also be advocated with respect to the four-factor model (Carhart, 1997),
which augments the three-factor model with the momentum (MOM) risk factor. Given the
case of an opposite scenario, i.e. if the factor models are able to explain the market
anomalies, then also it will be premature to discard the propositions, e.g. stock market is
efficient (anomalies are resultant of market inefficiency) or anomalies are mere chance
results with no systematic risk attributes. We mention such an attempt as premature
because this will discard the theoretical argument of systematic risk-based explanation for
such firm characteristics, which motivates the development of alternative multifactor APMs
(Carhart, 1997; Fama and French, 1993; Pastor and Stambaugh, 2003).
To circumvent such issues, in recent finance literature, attention has been gradually shifted
from mere identification of market anomalies and testing their empirical cross-sectional
regularities to a more fundamental question, i.e. what explains expected returns, risk
factors or firm characteristics that are commonly perceived to be market anomalies. It has
been argued that if the empirically motivated risk factors of alternative multifactor APMs
(Fama and French, 1993; Carhart, 1997) represent market wide systematic risk, then after

VOL. 9 NO. 3 2015 JOURNAL OF ASIA BUSINESS STUDIES PAGE 307

 making the stock raw returns risk adjusted using such systematic risk factors, the
incremental explanatory power of the market anomalies for the risk-adjusted returns should
be statistically insignificant (Avramov and Chordia, 2006 pp. 1002-1004). In recent years,
following Avramov and Chordia (2006) argument that several studies in the context of
developed markets examine the complete explanation debate (Antoniou et al., 2007; Bauer
et al., 2010; Chou et al., 2010). However, the universal validation of a particular APM is still
inconclusive. Moreover, the available literature on this issue is quite minimal in the context
of emerging stock markets. Given the far less universal dominance of any single APM,
either in its unconditional or conditional specifications for capturing the market anomalies,
it is imperative to assess the complete explanation debate in other emerging markets due
to their different market structure (Biscarri and Espinosa, 2008; Drobetz et al., 2002; Iqbal
et al., 2010; Jun et al., 2003; Lim and Brooks, 2010; Morck et al., 2000) and importance as
a special case of international portfolio diversification.
Drawing motivation on such arguments, the present paper attempts to revisit the complete
explanation debate, i.e. the applicability of alternative factor models for the complete
explanation of market anomalies, in the context of an emerging stock market like India. To
be more specific, the present paper seeks to answer two complementary research
questions. First, whether the alternative APMs in their conditional and unconditional
specifications are able to capture the unique importance of market anomalies. Second, do
the market anomalies retain their cross-sectional return predictability potential once the raw
return of individual securities are risk adjusted using the alternative unconditional and
conditional APMs.
Our approach in this regard extends the asset pricing literature in two ways. First, given the
distinctive nature of emerging stock markets, it helps to broaden our understanding in
terms of the applicability of alternative factor models for the complete explanation of market
anomalies. Over the years, asset pricing research in the context of emerging markets have
attracted considerable attention because of the growth in market size, liquidity,
liberalization of their markets and demand for international portfolio diversification.
However, to our knowledge, none of the existing literature in the contest of emerging stock
market has tested the applicability of such wide range of alternative multifactor APMs in
unconditional and conditional specifications for the complete explanation of market
anomalies. With special reference to the Indian stock market, this study retains its
importance because of the stock market’s unique nature in terms of the ownership
structure, investor participation and the decade-long institutional reform process for
enhancing the market efficiency. Second, it provides an out-of-sample evidence for the
applicability of factor models to explain the anomaly effect for stock returns behavior. Such
out-of-sample evidence from an emerging market is empirically more intuitive, as it is not
highly correlated with the data used in previous research and free from the data-snooping
biases (Lo and MacKinlay, 1990) associated with the developed markets.
Our results for alternative APMs specifications revel that, as compared to unconditional
APMs, conditional models perform better to capture the market anomalies. Results reveal
that considering several alternative APMs in their unconditional specifications, the BP and
liquidity (LIQ) effects can be explained by the market risk factors completely. When the risk
factors of alternative APMs are allowed to vary over time employing business cycle
conditioning information, the importance of medium-term and long-term momentum effect
have been captured to a greater extent. Although the active investment approach in stock
market shares a common ground of semi-strong form of market efficiency hypothesis,
which also supports the presence of market anomalies, less empirical evidence has been
explored in this regard to support or repute such belief of practitioners. Our empirical
findings make an attempt in this regard to suggest certain anomaly-based trading strategy
that can be followed for active portfolio management.
This paper is organized into seven sections. Section 2 provides a brief review of related
literature and elaborates the motivation behind this study. Section 3 outlines empirical

PAGE 308 JOURNAL OF ASIA BUSINESS STUDIES VOL. 9 NO. 3 2015

approach. Section 4 describes the data and variables. Section 5 presents empirical results.
Section 6 documents some robustness tests, and Section 7 concludes the paper.

2. Related literature and motivation

Following the seminal work of Brennan et al. (1998), attention in the recent finance literature
has been gradually shifted from mere identification of market anomalies and testing their
cross-sectional regularity, toward the applicability of existing factor models for capturing
the effects of such anomalies. Put it differently, whether the factor models are sufficient
enough to subsume the return predictability of asset market anomalies, or still the
anomalies retain their unique importance for the cross-sectional return predictability
(Avramov and Chordia, 2006). In this section, we attempt to review a set of related literature
that investigates the complete explanation debate in several developed and emerging
stock markets. While presenting the respective studies in this area, we tried to emphasize
in terms of the use of alternative APMs and their significance to capture the anomaly effect.
Brennan et al. (1998) lead the way in the US stock market by using unconditional Fama and
French (1993) three-factor model find evidence that MC and BP effects are attenuated,
while the MOM and LIQ effects persist. Further evidence to the debate on the application
of factor models to capture the effect of market anomalies is provided by Avramov and
Chordia (2006). In essence, Avramov and Chordia (2006) using US individual stock return
data for the sample period July 1964 through December 2001 and employing six alternative
APM attempt to find best APM for this debate. Avramov and Chordia (2006) find evidence
that capital asset pricing model (CAPM) (Lintner, 1965; Sharpe, 1964), Fama and French
(1993) three-factor model and Carhart (1997) four-factor model are not able to capture the
MC, BP, MOM and LIQ effects in their unconditional specifications. In the European stock
market, Bauer et al. (2010) find that unconditional three-factor model (Fama and French,
1993) fails to eliminate the size, value, liquidity and momentum effect completely.
Motivated with the success of conditional APMs, which accounts for time-varying nature of
systematic risk premia (Jagannathan and Wang, 1996), recent literature also explores the
complete explanation hypothesis using alternative conditional APMs. Using alternative
conditional APMs tests, Avramov and Chordia (2006) show that conditional specification of
Fama and French (1993) three-factor model is able to capture the MOM effect; however, it
fails to capture the LIQ effect. Ho and Hung (2009) and Ho (2012) using US stock market
data for the sample period July 1964 to December 2005 observe that conditional model
specifications outperform the unconditional models in terms of capturing anomalies.
Specifically, Fama and French three-factor model in its conditional specification is able to
capture the impact of MC and BP effects on returns. Ho and Hung (2009) also find
evidence to support the applicability of momentum augmented four-factor model to capture
the effect of MOM and LIQ effect. Chou et al. (2010) for the US stock market find evidence
that MC and BP effects are captured with the conditional Fama and French (1993)
three-factor model, whereas MOM and LIQ effects are still persistent. In contrast to
Avramov and Chordia’s (2006) findings in the US stock market, Antoniou et al. (2007) and
Bauer et al. (2010) for the European stock market observe strong MOM effect even after
using the conditional multifactor APMs for deriving risk-adjusted returns of individual
securities. To briefly sum up, the related literature across different developed markets
suggest that the effects of market anomalies are not consistently captured through
alternative APMs either in their unconditional or conditional specifications. Furthermore, the
unconditional specification of APMs are less effective to capture the anomaly effect as
compared to their conditional specifications (Antoniou et al., 2007; Avramov and Chordia,
2006; Bauer et al., 2010; Chou et al., 2010; Ho and Hung, 2009; Ho, 2012). Most of the
empirical research concerning this complete explanation debate has been focused on the
developed markets, and limited attention has been paid to the emerging stock markets.
Extending the available literature for the Japan stock market, Chang et al. (2010) observed
strong liquidity effect even after making the raw return risk adjusted using alternative

VOL. 9 NO. 3 2015 JOURNAL OF ASIA BUSINESS STUDIES PAGE 309

 multifactor models. Narayan and Zheng (2010) in the context of Chinese stock market find
that five-factor model or Carhart (1997) four-factor model augmented with the liquidity
factor is able to explain the LIQ, MC and BP effects but not the MOM effect. Using similar
argument, Dash and Mahakud (2013) in the context of Indian stock market observe that
five-factor model, i.e. Carhart (1997) four-factor model augmented with the liquidity factor
helps to explain BM and LIQ effects, but not the SZ and MOM effect. Apart from the
aforementioned two studies, the existing literature remains silent about the applicability of
such research hypothesis in the context of any other emerging markets. Table I
summarizes the related literature in this area and presents the anomaly explanation
evidence across different markets.
In this paper, we attempt to revisit the complete explanation debate in the context of Indian
stock market employing unconditional and conditional specification of CAPM, Fama and
French three-factor model and Carhart four-factor model. Apart from the limited attention
caveat in the context of emerging stock markets to test the complete explanation debate,
another significant limitation of the existing literature is the use of only unconditional multifactor
APMs in the context of emerging stock market. The existing study that closely resembles ours
is Dash and Mahakud (2013). Our study employs same set of anomalies and conditioning
information variable from the Indian stock market, as in Dash and Mahakud (2013), and
extends the sample period till March 2012. However, the unique distinction of our paper from
Dash and Mahakud (2013) is its focus on several alternative APMs. In this paper, we explore
Fama and French (1993) three-factor and Carhart (1997) four-factor APMs which are never
been considered by Dash and Mahakud (2013). In essence, our study provides a more
microscopic approach to this debate by focusing each APM as a unique representation of
systematic risk factors. As there is no empirical consensus with respect to any particular APM
that is having a universal dominance, so it is perceptible to test them all together. To find a
comparative benchmark, our motivation to examine several APMs in alternative specifications
follows Jagannathan and Wang (1996 p. 36), who suggest that:
[. . .] we have to keep in mind that any APMs is only an approximation of reality. Hence, it would
be rather surprising if it turns out to be 100 per cent accurate. The interesting question is not
whether a particular asset-pricing model can be rejected by the data. The question is how
inaccurate is the model?

Table I Correlation matrix of conditioning information variables and risk factors

Variables MRKT SMB HML WML TS DY IR IRD

MRKT 1
SMB 0.27 1
HML ⫺0.23 0.29 1
WML ⫺0.03 0.02 ⫺0.09 1
TS 0.04 0.01 ⫺0.04 ⫺0.08 1
DY 0.26 0.14 ⫺0.02 0.16 ⫺0.27 1
IR ⫺0.30 ⫺0.21 0.11 ⫺0.02 ⫺0.21 ⫺0.28 1
IRD ⫺0.16 ⫺0.15 0.06 ⫺0.03 ⫺0.37 0.01 0.84 1
Notes: Sample period consists of 199 monthly observations from September 1995 to March 2012;
following Fama and French (1993), SMB is measured each month as the equal-weighted average of
the returns on the three small stock portfolios minus the returns on the three big stock portfolios;
similarly, HML is measured each month as the equal-weighted average of the returns on two high BP
portfolios minus returns on the two low BP portfolios; following Carhart (1997), WML is the
equal-weighted average of the returns on the two winner stock portfolios minus the returns on the two
loser stock portfolios; for the construction of size factor or SMB, firm size is measured as natural
logarithm of market capitalization (stock prices times outstanding shares) at the end of August of
year y; for value factor or HML, BP has been measured as the ratio between book price for the fiscal
year ending in calendar year y by the market value of equity at the end of August in year y; for
momentum factor WML, we measure momentum as the cumulative return of a stock in month t-12
through month t-2 preceding August of the year y; we skip one month between portfolio formation
and holding period to avoid the effects of bid-ask spread, price pressure and any lagged reaction;
conditioning variables are lagged by one month to the risk factors

PAGE 310 JOURNAL OF ASIA BUSINESS STUDIES VOL. 9 NO. 3 2015

Our approach to test alternative APMs for explaining anomaly effect follows the similar
argument. Therefore, we attempt to accommodate alternative APMs in our analysis to
observe their individual anomaly explanation capacity.
Our motivation to test the complete explanation debate in the context of an emerging
market retains its importance because of two complementary arguments. First, it has been
argued in the prior literature that the cross-sectional return predictability associated with
large number of firm characteristics or market anomalies in the context of developed
markets may not be applicable in the context of emerging stock markets (Chen et al., 2010).
It could be due to the fact that emerging markets have been characterized with high stock
price synchronicity or low price informativeness (Chana and Hameed, 2006; Morck et al.,
2000), and with such characteristics, it is reasonable and practical to anticipate that there
is little scope for the diffusion of more firm-specific information to the stock prices and, thus,
a lower chance to find a profitable trading strategy using market anomalies. Therefore,
minimal presence of market anomalies in the context of emerging markets no way validates
the claim that market is efficient; rather, it is the lower price informativeness of stock prices
that hinders the significant role of market anomalies. Second, emerging stock markets are
distinguished by certain unique characteristics like unstable macroeconomic condition
(Drobetz et al., 2002; Iqbal et al., 2010), gradual liberalization and volatile market, thin
trading and low liquidity (Jun et al., 2003) and accounting information heterogeneity
because of accounting regulation effects (Biscarri and Espinosa, 2008). Pertinent to such
unique characteristics, it has also been argued that “emerging markets have long posed a
challenge for finance. Standard models are often ill suited to deal with the specific
circumstances arising in these markets” (Bekaert et al., 2003). Therefore, the factor models
that account for the explanation of certain market anomalies in the context of developed
market can never be generalized in the context of emerging markets because of its
distinctive nature. If market structures of emerging stock markets hinder the price
informativeness negatively and so also the early detection of anomalies, and the special
characteristics of emerging market provide a notion that the performance of APMs in such
markets cannot be generalized with the developed markets, then, which APMs performs
better in such kind of market scenario to capture the effects of market anomalies becomes
an important empirical issue to investigate.

3. Empirical approach
Our methodology closely follows the empirical approach developed by Brennan et al.
(1998) and Avramov and Chordia (2006). Assuming that returns of security j at time t are
generated by an L-factor approximate factor model:
L

R̃jt ⫽ Et⫺1(R̃jt) ⫹ 兺␤
l⫽1
f̃ ⫹ ˜
jlt⫺1 jt ␧jt, ∀j, t ⬎ 0, (1)

where: Et⫺1 is the conditional expectation operator; R̃jt is the return on security j at time t;
␤jlt⫺1 is the factor loading of the security’s return on factor l; f̃lt is the return on factor l at time
t; and ˜ ␧jt is the error term. To derive the risk-adjusted return of each security j for the month
t, we follow the first step time series regression approach of Brennan et al. (1998) (for a
detail discussion, see also Avramov and Chordia, 2006) with the following specifications:
L

R*jt ⬅ Rjt ⫺ RFt ⫺ 兺 ␤ˆ F ,

l⫽1
jl lt (2)

where: Flt ⬅ flt ⫹ ␭lt is the sum of the factor realization and its corresponding risk premium.
␤ˆ jl is the beta estimated by a first pass time series regression over the entire sample period,
and R*jt is the estimated risk-adjusted return on stock j at time t from the first pass time series
regression. The suggested risk adjustment procedure imposes the assumption that the
zero beta equals the risk-free rate, and that the L-factor premium is equal to the excess
return on the factor (Avramov and Chordia, 2006).

VOL. 9 NO. 3 2015 JOURNAL OF ASIA BUSINESS STUDIES PAGE 311

 For the deriving risk-adjusted returns of individual securities, we consider three alternative
linear factor models in their conditional and unconditional specifications: single-factor
CAPMs (Lintner, 1965; Sharpe, 1964), Fama and French (1993) three-factor model and
Carhart (1997) four-factor model. In our multifactor specifications, we restrict ourselves for
the three-factor and four-factor model (Carhart, 1997; Fama and French, 1993), following
the recent findings of Maio and Santa-Clara (2012), which suggest that the Fama and
French (1993) and Carhart (1997) multifactor models perform the best on time series and
cross-sectional behavior of state variables and factors. For all the three aforementioned
linear factor models, we also estimate their conditional specifications using four conditional
information variables. From the conditional asset pricing literature, we consider four
conditioning information variables, namely, term spread (TS), dividend yield (DY),
short-term interest rates (IRs) and interest rate differential (IRD) (detail discussion in Section 4).
The alternative unconditional and conditional APMs then used to derive risk-adjusted
returns of each security using the first step time series regression approach of equation (2).
The conditional specifications of the alternative APMs follow the scaled factor methodology
of Cochrane (1996), in which coefficients of the factors vary with the available conditioning
information set.
The derived risk-adjusted return in equation (2) becomes the ideal candidate for testing the
unique impact of market anomalies on stock returns left unexplained by the risk adjustment
process of alternative pricing models. Assuming that the APMs are sufficient enough to
describe the pervasive firm characteristics in both conditional and unconditional
specifications, the unique impact of the asset pricing anomalies to explain the risk-adjusted
returns is expected to be ruled out completely. The argument for testing the importance of
market anomalies (i.e. c ⫽1 [. . .] [. . .] C) for explaining the risk-adjusted return can be
specified as follows:
C

R*jt ⫽ ␣ ⫹ 兺␮ ␺
c⫽1
c cjt⫺1 ⫹ ␧jt, (3)

where: ␺jt⫺1 is the vector of financial market anomalies, and ␮jt is a vector of characteristics
rewards with respect to the market anomalies. Under the null of exact pricing or expected
excess return on security, j is determined solely by the loadings of the security’s return on
the L-factors, and the coefficients (␮c) of market anomalies will be equal to zero. Assuming
both firm and time effects in the data set, we extend the equation (3) to a two-way panel
data model specification as follows:
R*jt ⫽ ␣0 ⫹ ␣j ⫹ (␮c␺ct⫺1 ⫹ . . . . . . . . ⫹ ␮C␺Ct⫺1) ⫹ ⌫t ⫹ ␧jt, (4)
where, R*jt is the risk-adjusted return of regressors at time t. The regressors, namely, c [. . .]
[. . .] C are the firm-specific explanatory variables. ␣0 is the overall constant. ␣j is the
individual effect, which is assumed as constant over time and varies across the individual
cross-sectional unit (firm). ⌫t is the time-specific effect which varies across the time, but
constant across the firms. ␧jt is a stochastic error term assumed to have mean zero and
constant variance. For the estimation of equation (4), we use panel data estimation
technique to examine the predictability of firm characteristics for the risk-adjusted returns
derived from the first step. As an alternative to the Fama and MacBeth’s (1973) two-pass
cross-sectional regression approach, the use of panel data estimation helps to avoid the
errors-in-variables problem (Shanken, 1992) associated with Fama and MacBeth
approach. Moreover, panel dataset by using information on the intertemporal dynamics
and the individualities of entities helps to control in a more natural way for the effects of
missing or unobserved variables (Hsiao, 1986). As an alternative to the Fama and MacBeth
approach, the use of panel estimation for the data property that conjectures a balanced
panel with continuously traded stocks has also been highlighted by Goyal (2012). More
specifically, Goyal (2012) suggests that the popularity of Fama and MacBeth approach in
the asset pricing literature is attributable to the fact that the risk premium estimation results
not to be influenced because of the appearance and delisting of stocks in the exchange,
as it considers the available stocks at given point of time. In an active and operational stock

PAGE 312 JOURNAL OF ASIA BUSINESS STUDIES VOL. 9 NO. 3 2015

exchange, companies come, operate and continue to operate or delist in the due-course
of time; therefore, in a state of delisting of firms, the flexibility of the use of panel estimation
technique is limited, as it turns to be an unbalanced panel. As our data set comprised
continuously traded stocks, our empirical analysis is not susceptible to missing observation
and unbalanced panel.

4. Data and variables

Our sample consists of monthly returns of continuously traded non-financial common
stocks listed on the National Stock Exchange (NSE) of India from September 1995 to March
2012 (199 months). We obtain monthly data for individual securities from the PROWESS
database compiled by the Centre for Monitoring Indian Economy. To include a stock in the
sample for a given month, we select stocks based on the following criteria: its return in the
current month and in 24 of the previous 60 months had to be available in the PROWESS
database; must have sufficient data to calculate the market capitalization, the
book-to-market ratio, the price and other firm-specific information to calculate firm-specific
characteristics; to reduce delisting or backfilling biases, the stock must be a continuously
traded stocks. The selection of continuously traded individual stocks helps us to construct
a balanced panel, and to avoid missing observation, data-snooping bias and under
rejection bias that have been subject of considerable debate in prior asset pricing
literature. Considering the continuously traded stocks since the inception of NSE in 1994
and avoiding the stocks that have negative BP values, we restrict our sample for 582
stocks. To avoid excess volatility-induced return variation during the initial days of the NSE,
we consider monthly return observations since September 1995. For the risk-free rate
proxy, we use 91-days Treasury bill rate collected from the Reserve Bank of India official
Web site. The systematic market factor (MRKT) is measured as market return in excess of
the risk-free rate of interest. Our approach for the construction of these risk factors like
SMB, HML and WML (winners minus losers) closely follows the approach of Fama and
French (1993) and Carhart (1997). For the construction of SMB, HML and WML, consistent
with the related literature, we consider one-year holding period, and in September of each
year, portfolios are rebalanced to calculate value weighted returns from the beginning of
September of year y till August y⫹1.
Available literature on the performance of conditional APMs suggests that the selected
conditioning information variables should capture investors’ expectations about future
market return or business cycle conditions. Consistent with the conditional asset pricing
literature (Cochrane, 1996; Drobetz et al., 2002; Iqbal et al., 2010), the four conditioning
information variables have been measured as follows. Following Dash and Mahakud
(2013), TS is measured as the difference between 10-year government bond yields and 91
days Treasury bill rate. DY is computed by summing monthly dividends on the value
weighted S&P CNX Nifty index portfolio for the year preceding t and divided by the value
of the portfolio at t. IR is measured as the weighted average of call money rate reported by
the Reserve Bank of India on monthly basis. To assess the relevance of global risk factors,
we consider IRD measured as the difference between Mumbai Interbank Offered Rate
(MIBOR) or the Indian money market benchmark rate and London Interbank Offered Rate
(LIBOR) as fourth conditioning information variable. For IRD calculation, LIBOR and MIBOR
data have been collected from official Web sites of Bank of England and NSE.
Table I reports the correlation matrix of systematic risk factors and the conditioning
information variables. Observed negative correlation between TS and DY indicates the
apparent opposite relationship of both the variables with the business cycle and expected
returns (Fama, 1990). The positive correlation between TS and MRKT suggests that high
(low) values of TS reflects trough (peak) of business cycle fluctuation and, thus, higher
(lower) expected return (Fama and French, 1988). The positive correlation between DY and
MRKT suggests that when stock prices are low relative to dividends (high DY), the
expected returns are high (Fama, 1990). The positive correlation between the IR and IRD

VOL. 9 NO. 3 2015 JOURNAL OF ASIA BUSINESS STUDIES PAGE 313

 in Table I is because of the common IR component among the two variables. Observed
negative (positive) correlation of IR and IRD with SMB (HML) is consistent with the
argument of investor expectation hypothesis and superior size and value premium during
expansive monetary periods (Jensen and Mercer, 2002). The relationships between WML
and the conditioning information variables are consistent with the suggested systematic
link between macroeconomic effects and premiums associated with such risk factors in the
related literature (Chordia and Shivakumar, 2002).

4.1 Identification of firm characteristics or anomalies having cross-sectional regularity

For identifying firm characteristics or anomalies that are significantly associated with the
cross-section of stock return variation in the Indian stock market, we use the similar analysis
of Dash and Mahakud (2013). The only difference between the analysis of Dash and
Mahakud (2013) and ours is the extension of sample period. As suggested by
Subrahmanyam (2010), although the number of cross-sectional return predictors may be
as large as 50, to make a comparative stand, we have tried to focus on those market
anomalies which have been analyzed in earlier study in the context of Indian stock market.
Such approach is also consistent with related literature in the context of other emerging and
developed markets (Artmann et al., 2012; Chen et al., 2010; Haugen and Baker, 1996) and
minimizes the data or sample selection biases (Kothari et al., 1995; Lo and MacKinlay,
1990). Similar to the Dash and Mahakud (2013) study, we have considered the following
return predictors: MC, BP, earnings-to-price ratio (E/P), cash flow-to-price ratio (C/P),
dividend-to-price ratio (D/P), sales growth (SLG), accounting accruals (AC), 12 months
stock return momentum (MOM), past three years long-term reversal (LR), research and
development intensity (RDint), advertising intensity (AVint), stock liquidity (LIQ), book
leverage (Bliv), market leverage (Mliv), return on asset (ROA), capital expenditure (CAP),
asset growth (AG) and investment-to-assets ratio (INVT). Appendix provides a detail
description on the measurement of aforementioned firm characteristics or market
anomalies and the reference from which we derive their measurement motivation. If the
selected firm characteristics or market anomalies are able to explain the cross-sectional
variation in subsequent one-year returns, then, we must experience a statistically
significant return spread in the hedge portfolio that has been constructed using the return
difference of P10-P1 (large minus small). We calculate equal weighted portfolio returns of
each decile portfolios for making the return spread comparison tests. This approach also
helps to judge the return variation with respect to each anomaly without being influenced
by firm market capitalization.
With the extension of sample period and repeating the same test, we do not find any
significant difference in results (reported in Appendix) as compared to Dash and
Mahakud’s (2013) findings. The figures are qualitatively similar with statistical significance
associated with MC, BP, MOM, LIQ and Mliv. With the extension of sample period, this
further validates the weak anomaly effect in the Indian stock market. Similar to the
observation made by Dash and Mahakud (2013), we propose that the weak anomaly effect
may be partially because of more homogeneity of return predictors distribution with lower
dispersion of firm-level risk (Chen et al., 2010) or low price informativeness (Chen et al.,
2010; Morck et al., 2000). Our finding toward strong MC effects in the Indian stock market
is consistent with the findings of Mohanty (2002), Sehgal and Tripathy (2005), Moor and
Sercu (2013) and Dash and Mahakud (2013). For the significant MOM effect in the Indian
stock market, our finding is consistent with Sehgal and Jain (2011).
Motivated from the cross-sectional regularities observed for MC, BP, MOM, LIQ and Mliv,
we will restrict ourselves to these five risk characteristics for examining the explanatory
power of these market anomalies on the risk-adjusted returns of individual securities. Our
measurement of the five market anomalies closely follows the approach of Dash and
Mahakud (2013) and extends the sample period for further validation. In other words, if the
firm characteristics or market anomalies are found to be significantly explaining

PAGE 314 JOURNAL OF ASIA BUSINESS STUDIES VOL. 9 NO. 3 2015

the risk-adjusted returns derived from the alternative APMs, then it can be set for the
conclusion that anomalies retain their unique effects irrespective of the use of alternative
APMs. In other words, APMs fail to explain the anomalies, and the use of trading strategies
based on such anomalies still retains their unique importance.
Consistent with Brennan et al. (1998), Avramov and Chordia (2006), Dash and Mahakud
(2013) in our subsequent risk-adjusted return predictability tests, we measure the
aforementioned five anomalies differently. For each security and for each moth, firm
characteristics or market anomalies are calculated as follows: MC of the firm in month t is
measured as the natural logarithm of the market value of the equity at the end of the second
to last month. Monthly BP is the ratio of book value of equity at the financial year end in the
calendar year y to the market price of equity at the end of the month t-1 in the calendar year
y. LIQ is measured each month as the annual average of monthly turnover ratio, i.e. number
of shares traded to the number of shares outstanding. The liquidity values have been
considered from the end of second to last month. Each month Mliv is computed as the ratio
of total assets in the fiscal year ending March in the calendar year y to the market value of
equity at the end of the month t-2 in the calendar year y. To account for the short-, medium-
and long-term MOM effects, we consider SRM (the cumulative returns over the second
through third months, before the month of measurement), MRM (the cumulative returns over
the fourth through sixth months, before the month of measurement t) and LRM (is the
cumulative returns over the seventh through twelfth months before the month of
measurement). Consistent with Brennan et al. (1998), the lagged return variables were
constructed to exclude the return during the immediate prior month to avoid any spurious
association between the prior month return and the current month return caused by thin
trading or bid-ask spread effects. As the variables are measured for individual firms, the
descriptive statistics and the correlation matrix of these variables can be calculated as the
time series averages of monthly cross-sectional mean and correlations. We explicitly do not
go further to analyze any inference from these figures and, hence, not reported for the
purpose of brevity. Moreover, as the calculated figures will be time series averages of
monthly cross-sectional values, any statistical significance level cannot be associated with
such figures. In terms of the direction of correlation, we have found negative correlation
between BP with all the three momentum variables. Consistent with the findings of Asness
(1997) and Asness et al. (2009), it indicates that measures of MOM and BP values are
negatively correlated across stocks. Observed positive correlation between MC-MOM and
LIQ-MOM suggests that stock momentum increases (decreases) with the increase
(decrease) of firm market capitalization and liquidity. This indicates that in the context of
Indian stock market, momentum strategy can be better implemented with the large and
liquid stocks (Dash and Mahakud, 2013). Although available literature in the area of firm
characteristics effects on the predictability of stock return behavior focuses on the
portfolio-based approach (Artmann et al., 2012; Chen et al., 2010; Hart et al., 2003 among
others), the present study uses individual securities data. This approach helps us to control
for under rejection bias and data-snooping bias (Lo and MacKinlay, 1990) observed with
the portfolio-based approach.

5. Discussion of results
This section has been divided into four subsections. We discuss results for each APM that
has been employed for deriving the risk-adjusted returns of individual securities. The first
subsection presents results for the test of firm characteristics effects to explain the
risk-adjusted returns derived using unconditional APMs. The second, third and fourth
subsections focus on the explanatory power of firm characteristics for the risk-adjusted
return derived by the conditional specifications of CAPM, three-factor model and Carhart
four-factor model, respectively. In all the sections, we estimate panel data model specified
in equation (4), using the risk-adjusted returns derived through alternative conditional and
unconditional APM specifications as the dependent variable. For all the subsections,
reported model specification test statistics such as likelihood ratio (LR) test, Lagrange

VOL. 9 NO. 3 2015 JOURNAL OF ASIA BUSINESS STUDIES PAGE 315

 multiplier (LM) test and Hausman test statistics suggest that the firm-specific and
time-specific effects are present in the dataset, and the two-way fixed-effect model
(fixed-effect firm and time model) is more suitable for the estimation of equation (4).

5.1 Alternative unconditional APMs

Table II shows estimation results of the equation (4) for four alternative unconditional APMs.
The three different columns under each APMs indicate the estimation results with respect
to the three different momentum characteristics (RET 2-3, RET 4-6, RET7-12). Reported
results for unconditional CAPM (Columns 1, 2 and 3) fail to capture the importance of MC,
Mliv and MOM effects. MOM characteristics of three different forms have been statistically
significant in explaining the risk-adjusted return of individual stocks. Similar to the results of
unconditional CAPM, the unconditional three-factor model (Columns 4, 5 and 6) also fails
to give a complete explanation for MC, Mliv and MOM effects. The explanatory powers of
these characteristics are still persistent, given the inclusion of size (SMB) and value (HML)
factors in the three-factor model for deriving risk-adjusted returns of individual stocks in the
first step. This result reveal that even in the presence of systematic risk factors in
the unconditional specification of three-factor model, the role of market anomalies for the
predicting stock returns behavior cannot be ruled out completely. Similar results with
respect to the significant MC, Mliv and MOM effects are also evident for the risk-adjusted
return derived by using Carhart four-factor model (Columns 7, 8 and 9). These results are
intuitively more appealing with respect to the behavior of three momentum characteristics.
As the Carhart four-factor model augments the three-factor model with the momentum
factor (WML), from the theoretical stand of the multifactor models, the MOM characteristics
are expected to be insignificant in the second step. In common, the reported results in
Table II give an indication toward the special nature of MC, Mliv and MOM effects as the
three found to be indifferent toward the inclusion of additional risk factors in alternative
multifactor specifications.
Reported results in Table II for all the alternative APMs show a distinctive behavior for
explaining the LIQ effect. All the APMs even in their unconditional specifications are able
to explain the LIQ completely. This is different as compared to the related literature in the
context of developed stock market. For instance, Avramov and Chordia (2006) using US
stock market data observe strong LIQ effect irrespective of the use of alternative APMs.
However, our result in Table IV shows consistency with the findings of Narayan and Zheng
(2010) in the context of emerging stock market. Using unconditional multifactor models,
Narayan and Zheng (2010) observe that LIQ effect losses its impact in the case of Chinese
stock market. This difference in behavior of LIQ effect in the context of emerging stock
markets may be attributable to the special order-driven market structure which is different
from the quote-driven market structure of developed market structure of the developed
markets. The special nature of order-driven market supports several unique characteristics,
as it generates liquidity demand, supply schedules and price discovery that are consistent
with equilibrium under perfect condition (Brockman and Chung, 2002). Perhaps because
of such characteristics, the LIQ effect do not possess unique characteristics for explaining
the risk-adjusted returns, as the risk factors in the APMs are able to subsume its importance
completely.

5.2 Conditional capital APMs

Table III shows the estimation results of equation (4) for different conditional specifications
of CAPM. Reported results in Table III (Columns 1, 2 and 3) suggest that conditional CAPM
with TS as the conditioning information fails to rule out the importance of MC, Mliv and MOM
characteristics. Results for DY as the conditioning information (Columns 4, 5 and 6) even
fail to explain LIQ effect along with the earlier persistent MC, Mliv and MOM effects. In the
case of IR conditioning information (Columns 7, 8 and 9) and IRD conditioning information
(Columns 10, 11 and 12), the models still fail to rule out the persistent effect of MC, Mliv and
MOM characteristics. Overall, the reported results of conditional CAPM never seem to be

PAGE 316 JOURNAL OF ASIA BUSINESS STUDIES VOL. 9 NO. 3 2015

 Table II Fixed-effect firm and time model estimates for unconditional APMs
CAPMs Fama and French three-factor model Carhart four factor model
Coefficients (1) (2) (3) (4) (5) (6) (7) (8) (9)

MC ⫺0.93* (⫺21.27) ⫺0.91* (⫺21.25) ⫺1.03* (⫺21.1) ⫺0.89* (⫺19.13) ⫺0.83* (⫺18.41) ⫺0.85* (⫺18.62) ⫺0.81* (⫺18.29) ⫺0.85* (⫺18.05) ⫺0.81* (⫺18.57)
BP 0.17 (0.29) 0.21 (0.39) 0.49 (0.69) 0.22 (0.86) 0.11 (0.23) 0.48 (0.09) 0.08 (0.14) 0.11 (0.18) 0.17 (0.31)
LIQ ⫺0.21 (⫺0.79) ⫺0.21 (⫺0.71) ⫺0.01 (⫺0.16) ⫺0.11 (⫺0.69) ⫺0.17 (0.73) ⫺0.11 (⫺0.68) ⫺0.29 (⫺0.67) ⫺0.05 (⫺0.79) ⫺0.19 (⫺0.73)
SRM 0.02* (5.61) – – 0.03* (5.17) – – 0.14* (6.11) – –
MRM – 0.02* (5.37) – – 0.05* (3.47) – – 0.03* (5.86) –
LRM – – 0.16* (11.6) – – 0.21*** (1.93) – – 0.17** (2.59)
Mliv 0.03** (2.25) 0.05** (2.26) 0.01** (1.89) 0.07* (7.51) 0.09* (8.01) 0.07* (7.89) 0.05* (7.86) 0.09* (8.37) 0.07* (8.31)
LR test
[x2 (780)] 831.27 (0.00) 825.89 (0.00) 811.65 (0.00) 829.63 (0.00) 809.55 (0.00) 819.28 (0.00) 831.63 (0.00) 806.00 (0.00) 825.47 (0.00)
LM test [x2 (2)] 437.19 (0.00) 452.76 (0.00) 368.77 (0.00) 497.87 (0.00) 541.37 (0.00) 501.49 (0.00) 519.26 (0.00) 567.65 (0.00) 529.71 (0.00)
Hausman test
[x2 (5)] 831.11 (0.00) 1,301.67 (0.00) 652.19 (0.00) 1,103.01 (0.00) 1,206.53 (0.00) 1,704.84 (0.00) 2,580.55 (0.00) 2,861.51 (0.00) 1,383.73 (0.00)
R2 0.0049 0.0057 0.0063 0.0045 0.0043 0.0043 0.0047 0.0049 0.0043
D-W statistic 2.22 2.29 2.31 2.35 2.33 2.37 2.32 2.38 2.39
F-test 117.22 (0.00) 126.67 (0.00) 129.03 (0.00) 91.37 (0.00) 89.57 (0.00) 93.79 (0.00) 101.07 (0.00) 94.21 (0.00) 93.57 (0.00)
Notes: The three different columns under each APMs indicate the estimation results with respect to the three different momentum characteristics such as SRM, MRM and LRM along
with other firm characteristics; for each stock and for each month, SRM, MRM and LRM are the cumulative returns over the second through third, fourth through sixth and seventh
through twelfth months before the month of analysis; MC is the natural logarithm of the market value of the equity at the end of the second to last month; BP is the ratio of book value
of equity at the financial year end in the calendar year y to the market price of equity at the end of the month t-1 in the calendar year y; LIQ is the annual average of monthly turnover
ratio, i.e. number of shares traded to the number of shares outstanding; the liquidity values have been considered from the end of second to last month; Mliv is the ratio of total assets

VOL. 9 NO. 3 2015

in the fiscal year ending March in the calendar year y to the market value of equity at the end of the month t-2 in the calendar year y; sample period consists of monthly observations
of 582 continuously traded firms from September 1995 to March 2012; LR test (Gourieroux et al., 1982) carried out to identify the existence of individual firm-specific effects in the
data set; LM test (Breusch and Pagan, 1980) has been used to test the acceptability of panel data models over the classical regression models; the Hausman (1978) specification
test is performed on each system to determine which estimation method is most appropriate; t-statistics are in parenthesis; figures in the curly brackets represent corresponding
p-values of model specification test statistics; *; **; ***represent statistical significance at 1, 5 and 10%, respectively

JOURNAL OF ASIA BUSINESS STUDIES

PAGE 317
PAGE 318 JOURNAL OF ASIA BUSINESS STUDIES
Table III Fixed-effect firm and time model estimates for conditional CAPMs
Term spread Dividend yield Interest rate Interest rate differential
Coefficients (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)

VOL. 9 NO. 3 2015

MC ⫺1.47** (⫺3.15) ⫺1.48** (⫺3.22) ⫺1.48** (⫺3.61) ⫺1.67* (⫺9.48) ⫺1.66* (⫺9.39) ⫺1.68* (⫺9.27) ⫺0.82* (⫺7.41) ⫺0.85* (⫺7.37) ⫺0.88* (⫺7.43) ⫺1.07* (⫺9.72) ⫺0.99* (⫺9.51) ⫺1.08** (⫺2.09)
BP 0.03 (0.81) 0.02 (0.83) 0.03 (0.96) 0.04 (0.61) 0.02 (0.56) 0.02 (0.45) 0.02 (0.58) 0.02 (0.66) 0.04 (0.59) 0.02 (0.12) 0.03 (0.16) 0.02 (0.11)
SRM 0.02* (9.85) – – 0.02*** (1.89) – – 0.04* (12.08) – – 0.01* (6.88) – –
MRM – 0.11* (5.09) – – 0.03** (2.56) – – 0.02* (6.64) – – 0.02** (2.89) –
LRM – – 0.06* (5.59) – – 0.05* (3.62) – – 0.02* (8.69) – – 0.01* (6.90)
LIQ ⫺0.05 (⫺1.27) ⫺0.05 (⫺1.41) ⫺0.03 (⫺1.12) ⫺0.61* (⫺6.59) ⫺0.61* (⫺6.07) ⫺0.61* (⫺6.89) ⫺0.02 (⫺1.25) ⫺0.02 (⫺1.25) ⫺0.01 (⫺1.28) ⫺0.08 (⫺0.37) ⫺0.06 (⫺0.42) ⫺0.07 (⫺0.37)
Mliv 0.12* (9.08) 0.11* (9.88) 0.14* (4.21) 0.11* (9.39) 0.10* (9.11) 0.09* (8.91) 0.11* (8.17) 0.09* (8.17) 0.08* (8.61) 0.11* (9.53) 0.09* (9.53) 0.10* (9.53)
LR test
[x2 (780)] 888.51 (0.00) 879.36 (0.00) 819.21 (0.00) 891.36 (0.00) 871.35 (0.00) 839.78 (0.00) 856.19 (0.00) 889.25 (0.00) 853.21 (0.00) 836.29 (0.00) 821.81 (0.00) 829.45 (0.00)
LM Test [x2 (2)] 11.93 (0.00) 12.51 (0.00) 14.69 (0.00) 195.31 (0.00) 184.27 (0.00) 166.49 (0.00) 50.22 (0.00) 48.29 (0.00) 55.78 (0.00) 37.15 (0.00) 31.29 (0.00) 35.11 (0.00)
Hausman Test
[x2 (5)] 1,539.27 (0.00) 1,009.1 (0.00) 818.72 (0.00) 2,325.3 (0.00) 3,216.53 (0.00) 1,412.7 (0.00) 368.41 (0.00) 383.63 (0.00) 385.51 (0.00) 717.11 (0.00) 675.29 (0.00) 607.79 (0.00)
R2 0.0115 0.0307 0.0131 0.0483 0.0129 0.0127 0.0044 0.0037 0.0038 0.0045 0.0041 0.0045
DW statistic 2.39 2.33 2.57 2.41 2.42 2.44 2.43 2.43 2.45 2.41 2.42 2.42
F-test 201.21 (0.00) 186.59 (0.00) 191.21 (0.00) 207.33 (0.00) 209.43 (0.00) 204.18 (0.00) 89.27 (0.00) 71.36 (0.00) 72.09 (0.00) 91.22 (0.00) 85.27 (0.00) 90.81 (0.00)
***
Notes: Same as Table II; *; **; represent statistical significance at 1, 5 and 10%, respectively
encouraging for capturing the anomaly effect; moreover, in its conditional specification, the
CAPM performs similar to its unconditional specification.

5.3 Conditional Fama and French three-factor model

Table IV shows the estimation results of equation (4) for different conditional specifications
of Fama and French (1993) three-factor model. Reported results in Table V suggest that the
performance of three-factor model in the alternative conditional specifications have never
been satisfactory to capture MC and Mliv effects. The effects of MC and Mliv for the
firm-specific risk-adjusted return explanation retain their significance irrespective of the use
of different conditioning information variables. Significance of MC effect for explaining
risk-adjusted returns of individual securities derived using conditional three-factor model is
inconsistent with the findings of asset pricing literature in the context of emerging stock
markets. For instance, Schrimf et al. (2007) and Iqbal et al. (2010) support the applicability
of conditional three-factor model as the better benchmark APMs for describing the
cross-sectional variation of expected stock returns in the context of emerging markets.
Considering the argument of Schrimf et al. (2007) and Iqbal et al. (2010), it can be expected
to have less number of significant anomaly effects for explaining the risk-adjusted returns.
However, on the contrary to such propositions, our results suggest that anomalies like MC,
Mliv and SRM retain their significance for explaining stock returns behavior. Persistence of
MC effect is also inconsistent with the findings of Avramov and Chordia (2006), which
suggest that conditional three-factor model specification with business cycle conditioning
information helps to capture the cross-sectional regularity related to MC characteristic.
In the case of TS and IRD conditioning information variable, the conditional three-factor
model is able to capture MRM and LRM effects. Across all the four different conditional
specifications of three-factor model, the SRM effect is found to be persistent. To the extent
of MRM and LRM complete explanation, our findings are consistent with the related
literature that attributes the macroeconomic or business cycle conditions as the sources of
momentum strategy pay-off (Chordia and Shivakumar, 2002). In a related strand, Avramov
and Chordia (2006) argue that conditional APMs specification (i.e. making factor loadings
time varying with respect to business cycle variables) is suitable to explain profitability of
MOM strategies. The conditional three-factor model specifications with the use of DY and
IR conditioning information variables, however, are not able to capture any of the MOM
effects.

5.4 Conditional Carhart four-factor model

Table VI presents estimation results of equation (4) for different conditional specifications
of Carhart (1997) four-factor model. Reported results in Table VII suggest that the
alternative conditional Carhart four-factor model specifications fail to give a complete
explanation evidence for MC, SRM2-3, MRM4-6, LRM7-12 and Mliv effects. The MC and
Mliv effects are able to maintain their statistical significance as the determinants of
cross-sectional return variation. The coefficients are qualitatively similar to the conditional
three-factor model. For the TS and IRD conditional specifications, the relative importance
of MRM4-6 and LRM7-12 effects loses their significance. However, SRM2-3 which indicates
short-run momentum effect still retains its importance for the cross-sectional returns
variation irrespective of the conditional specification of Carhart four-factor model.
Consistent with the findings of Avramov and Chordia (2006) for the US stock market and
Dash and Mahakud (2013) for Indian stock market, our results in Table VI reveal that MRM
and LRM momentum effects in the case of TS and IRD conditional specifications suggest
that medium- and long-term momentum profits are attributable to the asset pricing
misspecification that varies with the business cycle. In common, reported results in
Tables IV and VI also give an insight that conditional three-factor model and Carhart
four-factor model offer similar set of evidences irrespective of the presence of additional
momentum factor in the Carhart four-factor model specification.

VOL. 9 NO. 3 2015 JOURNAL OF ASIA BUSINESS STUDIES PAGE 319

PAGE 320 JOURNAL OF ASIA BUSINESS STUDIES
Table IV Fixed-effect firm and time model estimates for conditional three-factor model
Term spread Dividend yield Interest rate Interest rate differential

VOL. 9 NO. 3 2015

Coefficients (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)

MC ⫺1.45* (⫺9.87) ⫺1.42* (⫺9.32) ⫺1.43* (⫺9.53) ⫺1.02* (⫺9.31) ⫺1.01* (⫺8.75) ⫺1.01* (⫺8.86) ⫺0.84* (⫺5.41) ⫺0.81* (⫺5.01) ⫺0.82* (⫺5.60) ⫺0.96* (⫺9.79) ⫺0.97* (⫺9.35) ⫺0.98* (⫺9.41)
BP 0.02 (0.96) 0.03 (0.12) 0.02 (0.14) 0.06 (0.98) 0.09 (1.19) 0.07 (⫺1.25) 0.02 (0.28) 0.06 (0.28) 0.02 (1.27) 0.04 (0.28) 0.02 (0.39) 0.05 (0.32)
SRM 0.03* (10.57) – – 0.09* (3.51) – – 0.03* (12.94) – – 0.05* (7.07) – –
MRM – 0.01 (0.86) – – 0.02* (4.02) – – 0.02* (6.61) – – 0.01 (0.99) –
LRM – – 0.05 (0.61) – – 0.01* (5.26) – – 0.07* (9.34) – – 0.05* (3.66)
LIQ ⫺0.10 (⫺0.56) ⫺0.06 (⫺0.57) ⫺0.02 (⫺0.58) ⫺0.62* (⫺8.01) ⫺0.59* (⫺7.19) ⫺0.61* (⫺7.11) ⫺0.02 (⫺1.39) ⫺0.16 (⫺1.61) ⫺0.01 (⫺1.42) ⫺0.09 (⫺0.51) ⫺0.08 (⫺0.50) ⫺0.09 (⫺0.49)
Mliv 0.11* (11.41) 0.14* (12.01) 0.11* (11.78) 0.07* (7.42) 0.08* (8.37) 0.09* (8.13) 0.09* (7.56) 0.73* (7.51) 0.08* (7.49) 0.15* (13.59) 0.12* (13.58) 0.12* (13.41)
LR test [x2 (780)] 828.31 (0.00) 821.16 (0.00) 825.39 (0.00) 804.41 (0.00) 809.02 (0.00) 821.31 (0.00) 834.58 (0.00) 827.53 (0.00) 831.38 (0.00) 871.49 (0.00) 838.26 (0.00) 835.21 (0.00)
LM test [x2 (2)] 18.79 (0.00) 24.05 (0.00) 25.29 (0.00) 179.0.85 (0.00) 203.47 (0.00) 209.53 (0.00) 65.31 (0.00) 59.57 (0.00) 65.29 (0.00) 21.75 (0.00) 26.83 (0.00) 19.53 (0.00)
Hausman Test 1963.7 (0.00) 1227.1 (0.00) 970.87 (0.00) 1314.7 (0.00) 2425.6 (0.00) 1296.1 (0.00) 301.7 (0.00) 320.6 (0.00) 325.8 (0.00) 1203.2 (0.00) 970.9 (0.00) 799.7 (0.00)
x2 (5)
R2 0.0099 0.0088 0.0088 0.0067 0.0064 0.0045 0.0038 0.0029 0.0032 0.0037 0.0047 0.0049
D-W statistic 2.61 2.62 2.63 2.47 2.43 2.47 2.43 2.47 2.47 2.43 2.48 2.45
F-test 211.38 (0.00) 181.27 (0.00) 189.85 (0.00) 107.41 (0.00) 105.21 (0.00) 109.47 (0.00) 81.25 (0.00) 59.93 (0.00) 64.81 (0.00) 101.27 (0.00) 95.67 (0.00) 94.91 (0.00)
***
Notes: Same as Table II; *; **; represent statistical significance at 1, 5 and 10%, respectively
 Table V Fama–MacBeth regression estimates for three-factor and four-factor model
Fama and French three-factor model Carhart four-factor model
Term spread Term spread
Coefficients (1) (2) (3) (4) (5) (6)

Intercept 0.61 (0.27) [0.43] 0.24 (0.34) [0.22] 0.73 (0.32) [0.17] 0.37 (0.82) [0.64] 0.39 (0.45) [0.17] 0.35 (0.17) [0.11]
MC ⫺4.83 (⫺2.17) [⫺1.94] ⫺3.27 (⫺5.70) [⫺2.01] ⫺5.20 (⫺2.58) [⫺1.84] ⫺3.55 (⫺2.01) [⫺1.82] ⫺4.10 (⫺2.62) [⫺2.08] ⫺4.61 (⫺3.29) [⫺1.93]
BP 1.55 (1.02) [0.83] 1.17 (0.63) [0.21] 0.57 (0.39) [0.16] 1.83 (1.49) [0.60] 1.40 (1.16) [0.43] 1.22 (1.37) [1.21]
LIQ ⫺0.22 (⫺0.11) [⫺0.03] ⫺1.24 (⫺0.83) [⫺0.27] ⫺1.52 (⫺1.20) [⫺0.87] ⫺2.71 (⫺2.47) [⫺1.73] ⫺1.66 (⫺0.22) [⫺0.16] ⫺1.63 (⫺0.38) [⫺0.20]
SRM 0.39 (0.12) [1.23] – – 1.28 (0.92) [1.21] – –
MRM – 1.39 (1.11) [0.72] – – 1.46 (0.61) [0.22] –
LRM – – 1.42 (1.83) [1.24] – – 0.83 (1.76) [1.33]
Mliv 5.13 (3.06) [2.78] 3.24 (2.09) [1.98] 5.29 (2.62) [2.02] 3.91 (2.06) [1.91] 3.52 (2.38) [1.82] 3.08 (2.41) [1.87]
R2 (%) 3.29 2.86 3.42 3.73 2.62 2.85
Notes: This table presents the time series averages of individual stock cross-sectional ordinary least squares (OLS) regression coefficient estimates using the Fama–MacBeth
approach; R2 is the time series average of the monthly adjusted R2; the OLS t-statistics are presented under the coefficient estimates, t-statistics in parenthesis use standard errors
as per Shanken (1990)

VOL. 9 NO. 3 2015

JOURNAL OF ASIA BUSINESS STUDIES
PAGE 321
PAGE 322 JOURNAL OF ASIA BUSINESS STUDIES
Table VI Fixed-effect firm and time model estimates for conditional Carhart four-factor model
Term spread Dividend yield Interest rate Interest rate differential

VOL. 9 NO. 3 2015

Coefficients (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)

MC ⫺1.44* (⫺9.61) ⫺1.41* (⫺8.97) ⫺1.39* (⫺8.58) ⫺1.07* (⫺9.37) ⫺1.03* (⫺9.37) ⫺1.01* (⫺19.4) ⫺0.86* (⫺5.91) ⫺0.85* (⫺5.69) ⫺0.86* (⫺6.21) ⫺0.96* (⫺9.57) ⫺0.94* (⫺9.21) ⫺0.94* (⫺9.41)
BP 0.02 (0.98) 0.02 (0.15) 0.06 (0.17) 0.07 (1.14) 0.07 (1.16) 0.07 (1.37) 0.02 (0.47) 0.04 (0.56) 0.02 (0.48) 0.03 (0.25) 0.02 (0.37) 0.02 (0.29)
LIQ ⫺0.11 (⫺0.60) ⫺0.12 (⫺0.61) ⫺0.16 (⫺0.61) ⫺0.63* (⫺7.81) ⫺0.65* (⫺7.71) ⫺0.63* (⫺7.47) ⫺0.03 (⫺1.41) ⫺0.03 (⫺1.39) ⫺0.02 (⫺1.44) ⫺0.02 (⫺0.57) ⫺0.02 (⫺0.57) ⫺0.03 (⫺0.58)
SRM 0.03* (10.7) – – 0.09* (4.1) – – 0.04* (12.8) – – 0.03* (7.2) – –
MRM – 0.01 (1.15) – – 0.08* (4.18) – – 0.02* (6.47) – – 0.02 (0.94) –
LRM – – 0.05 (0.45) – – 0.02 (1.26) – – 0.03 (1.73) – – 0.04 (1.11)
Mliv 1.12* (11.1) 0.14* (11.9) 0.12* (12.6) 0.09* (7.6) 0.06* (7.8) 0.08* (8.6) 0.08* (7.3) 0.07* (7.7) 0.07* (7.9) 0.14* (12.9) 0.15* (12.7) 0.14* (11.9)
LR test [x2 (780)] 809.21 (0.00) 862.01 (0.00) 810.26 (0.00) 814.24 (0.00) 822.11 (0.00) 819.36 (0.00) 829.34 (0.00) 827.15 (0.00) 821.19 (0.00) 828.53 (0.00) 821.37 (0.00) 819.27 (0.00)
LM test [x2 (2)] 39.01 (0.00) 38.78 (0.00) 38.57 (0.00) 175.02 (0.00) 199.61 (0.00) 201.77 (0.00) 47.33 (0.00) 45.79 (0.00) 48.27 (0.00) 17.19 (0.00) 21.29 (0.00) 17.51 (0.00)
Hausman test
[x2 (5)] 2,797.1 (0.00) 1,369.0 (0.00) 1,025.2 (0.00) 2,791.7 (0.00) 2,021.3 (0.00) 1,119.5 (0.00) 325.1 (0.00) 339.7 (0.00) 343.2 (0.00) 1,085.5 (0.00) 902.1 (0.00) 755.7 (0.00)
R2 0.0099 0.0080 0.0089 0.0062 0.0062 0.0063 0.0045 0.0028 0.0031 0.0048 0.0041 0.0044
D-W statistic 2.36 2.31 2.36 2.41 2.44 2.45 2.44 2.47 2.48 2.47 2.46 2.48
F-test 231.11 (0.00) 192.05 (0.00) 188.67 (0.00) 111.47 (0.00) 117.36 (0.00) 121.31 (0.00) 91.83 (0.00) 65.27 (0.00) 71.49 (0.00) 98.43 (0.00) 90.83 (0.00) 91.29 (0.00)
***
Notes: Same as Table II; *; **; represent statistical significance at 1, 5 and 10%, respectively
To briefly sum up, results across all the alternative APMs specifications suggest that, as
compared to unconditional APMs, conditional models perform better to capture the asset
pricing anomalies. APMs in their unconditional specification are able to explain only the LIQ
and BP effects completely. Among the conditional APMs, conditional CAPM performs
poorly to capture asset pricing anomaly effects as compared to the other conditional
multifactor models like three-factor model and Carhart four-factor model. Among the
conditional three-factor model and Carhart four-factor model, we observe that both the
models scaled with the TS conditioning information variable perform better by explaining
BP, LIQ, MRM and LRM anomaly effects. It is, however, interesting and appealing to note
that none of the APMs either in their unconditional or conditional specifications are able to
capture the MC, Mliv or SRM effects. The reported strong presence of short-term
momentum effect, i.e. SRM in the context of Indian stock market, is qualitatively similar to
the related literature in the context of emerging markets. Griffin et al. (2003) suggest that the
differential profit associated with the momentum strategy shows weak evidence in the
emerging markets as compared to the developed markets. For instance, as compared to
Rouwenhorst’s (1998) findings of significant momentum profit (on average 0.91 per cent
per month) in 11 of the 12 European markets and Jegadeesh and Titman’s (1993) average
0.95 per cent per month for the US market, Rouwenhorst (1999) found that only 6 of the 20
emerging markets exhibit momentum profits (average 0.39 per cent per month). Although
our empirical approach is very different as compared to the aforementioned studies, the
strong presence of short-term momentum, i.e. SRM2-3, and not the long-term momentum,
i.e. LRM7-12, gives an indication that the lower momentum profit observed by Rouwenhorst
(1999) using long-term momentum scenario with one-year holding period may be due to
the special nature of emerging markets. The significant cross-sectional regularity observed
with respect to size and momentum is also consistent with the findings of Rouwenhorst
(1998), Hart et al. (2003) and Lischewski and Voronkova (2012), in other emerging
international stock markets. However, inconsistent with Asness et al.’s (2009) proposition of
“value and momentum everywhere” in the context of international stock markets, we find
strong evidence of size and short run momentum effect. Over all, the results reveal that in
emerging stock markets, the applicability of APMs and the anomaly effect behavior cannot
be generalized completely with that of the developed markets because of the very different
market structure.

6. Robustness tests
Among the aforementioned alternative conditional APM specifications, we observe that
Fama and French three-factor model (Table IV) and Carhart four-factor model (Table VI)
specification with TS conditioning information is able to explain more anomalies. To be
specific, results reveal that BM, LQ, SRM and LRM anomaly effects are completely
explained by the two conditional APM. However, it may be argued that our results may be
because of the methodology that has been adopted. As our empirical approach is different,
the results warrant robustness tests. In this section, we specify a robustness test for the two
APMs using the Fama–MacBeth regression approach of Avramov and Chordia (2006). Our
interest is to reexamine our data using the Fama–MacBeth regression approach to see
whether our results show any deviation in terms of the observed anomaly effect.
Our robustness test results in Table V reveal that using alternative approach the anomaly
effect is only persistent in MC and Mliv. In both, the three-factor model and Carhart
four-factor model scaled with TS as conditioning variable is able to explain all other
anomaly effect. The use of conditional Carhart four-factor model is able to explain the
SRM2-3 effect, which was persistent in our earlier tests of Carhart four-factor model. In case
of other anomaly effects, like BM, LIQ, MRM and LRM, results are qualitatively similar to
fixed-effect firm and time model estimates.

VOL. 9 NO. 3 2015 JOURNAL OF ASIA BUSINESS STUDIES PAGE 323

 7. Conclusions
This paper investigates whether the alternative unconditional and conditional APMs
capture the role of market anomalies in the context of emerging stock market like India.
Results reveal that considering three alternative APMs in their unconditional specifications,
the BP and liquidity effects can be explained by the market risk factors completely.
However, conditional factor models’ do not outperform their unconditional counterparts
always. When the risk factors of alternative APMs are allowed to vary over time employing
business cycle conditioning information like TS, the importance of medium- and long-term
momentum effect have been captured to a greater extent. Results are robust to the use of
alternative approach. Our finding related to the presence of significant size effect is
inconsistent with asset pricing literature that argues for the disappearing size effect in the
developed stock markets because practitioners use it as an investment strategy and tried
to exploit the anomaly (van Dijk, 2011). Consistent with recent findings by Moor and Sercu
(2013) and Dash and Mahakud (2013) for strong size effect in the Indian stock market, our
findings suggest that size effect retains its importance as a profitable investment strategy.
For enhancing the performance of long-term investment portfolios, the observed pattern of
small size effect warrants close-ended mutual fund investment strategy. We recommend a
close-ended mutual fund strategy related to size effect because small size stocks are
generally attributed with illiquidity and high transaction cost characteristics. Considering a
short-term investment scenario, the investment managers can look forward to the large size
stocks. We expect that large size stocks with high liquidity may be better suitable for the
momentum strategy implementation because of the lower risk of illiquidity. However, the
high transaction costs because of frequent portfolio rebalancing in a short horizon
investment window warrants due attention under such momentum investment strategy.
The results also reveal that market leverage effect is still persistent in the Indian stock
market. This warrants further research in this direction by using the approach of Chou et al.
(2010) and Muradoglu and Sivaprasad (2013) for testing the applicability of leverage risk
factor in an augmented three-factor or four-factor model. Recent findings by Avramov et al.
(2013) also intuitive to test distress risk evidence in the context of Indian stock market and
implications of financial distress for the profitability of anomaly-based trading strategies.
The findings can be used for further research by incorporating some of the market
anomalies which have been excluded from our sample. This will bring better understanding
for the presence of market anomalies in the Indian stock market. One of the reasons of
limited anomaly effect related to stock price synchronicity also gives an intuitive opportunity
to test the relationship between stock price synchronicity and cross-section of stock returns
behavior.

References
Amihud, Y. (2002), “Illiquidity and stock returns: cross-section and time-series effects”, Journal of
Financial Markets, Vol. 5 No. 1, pp. 31-56.

Antoniou, A., Herbert, Y.T.L. and Paudyal, K. (2007), “Profitability of momentum strategies in
international markets: the role of business cycle variables and behavioural biases”, Journal of Banking
and Finance, Vol. 31 No. 3, pp. 955-972.

Artmann, S., Finter, P. and Kempf, A. (2012), “Determinants of expected stock returns: large sample
evidence from the German market”, Journal of Business Finance & Accounting, Vol. 39 Nos 5/6,
pp. 758-784.

Asness, C.S. (1997), “The interaction of value and momentum strategies”, Financial Analysts Journal,
Vol. 53 No. 2, pp. 29-36.

Asness, C.S., Moskowitz, T.J. and Pedersen, L.H. (2009), “Value and momentum everywhere”,
Chicago Booth Research Paper No. 12-53.

Avramov, D. and Chordia, T. (2006), “Asset pricing models and financial market anomalies”, Review
of Financial Studies, Vol. 19 No. 3, pp. 1001-1040.

PAGE 324 JOURNAL OF ASIA BUSINESS STUDIES VOL. 9 NO. 3 2015

Avramov, D., Chordia, T., Jostova, G. and Philipov, A. (2013), “Anomalies and financial distress”,
Journal of Financial Economics, Vol. 108 No. 1, pp. 139-159.

Banz, R.W. (1981), “The relationship between return and market value of common stocks”, Journal of
Financial Economics, Vol. 9 No. 1, pp. 3-18.

Basu, S. (1977), “Investment performance of common stocks in relation to their price-earnings ratios:
a test of the efficient market hypothesis”, Journal of Finance, Vol. 32 No. 3, pp. 663-682.

Bauer, R., Cosemans, M. and Schotman, P.C. (2010), “Conditional asset pricing and stock market
anomalies in Europe”, European Financial Management, Vol. 16 No. 2, pp. 165-190.

Bekaert, G., Harvey, C.R. and Lundblad, C. (2003), “Equity market liberalization in emerging markets”,
The Journal of Financial Research, Vol. 26 No. 3, pp. 275-299.

Biscarri, G.J. and Espinosa, G.L. (2008), “The influence of differences in accounting standards on
empirical pricing models: an application to the Fama-French model”, Journal of Multinational Financial
Management, Vol. 18 No. 4, pp. 369-388.

Brennan, M.J., Chordia, T. and Subrahmanyam, A. (1998), “Alternative factor specifications, security
characteristics, and the cross-section of expected stock returns”, Journal of Financial Economics,
Vol. 49 No. 3, pp. 345-373.

Breusch, T.S. and Pagan, A.R. (1980), “The Lagrange multiplier test and its applications to model
specification in econometrics”, Review of Economic Studies, Vol. 47 No. 1, pp. 239-253.

Brockman, P. and Chung, D.Y. (2002), “Commonality in liquidity: evidence from an order-driven market
structure”, Journal of Financial Research, Vol. 25 No. 4, pp. 521-539.

Carhart, M.M. (1997), “On persistence in mutual fund performance”, Journal of Finance, Vol. 52 No. 1,
pp. 57-82.

Chan, K.C. and Chen, N. (1991), “Structural and return characteristics of small and large firms”,
Journal of Finance, Vol. 46 No. 4, pp. 1467-1484.

Chan, L., Lakonishok, J. and Sougiannis, T. (2001), “The stock market valuation of research and
development expenditures”, Journal of Finance, Vol. 56 No. 6, pp. 2431-2456.

Chana, K. and Hameed, A. (2006), “Stock price synchronicity and analyst coverage in emerging
markets”, Journal of Financial Economics, Vol. 80 No. 1, pp. 115-147.

Chang, Y.Y., Faff, R. and Hwang, C.Y. (2010), “Liquidity and stock returns in Japan: new evidence”,
Pacific-Basin Finance Journal, Vol. 18 No. 1, pp. 90-115.

Chen, X., Kim, K.A., Yaoc, T. and Yu, T. (2010), “On the predictability of Chinese stock returns”,
Pacific-Basin Finance Journal, Vol. 18 No. 4, pp. 403-425.

Chordia, T. and Shivakumar, L. (2002), “Momentum, business cycle, and time-varying expected
returns”, Journal of Finance, Vol. 57 No. 2, pp. 985-1020.

Chou, P.H., Kob, K.C. and Lin, S.J. (2010), “Do relative leverage and relative distress really explain size
and book-to-market anomalies?” Journal of Financial Markets, Vol. 13 No. 1, pp. 77-100.

Cochrane, J.H. (1996), “A cross sectional test of an investment based APMs”, Journal of Political
Economy, Vol. 104 No. 3, pp. 572-621.

Cooper, M.J., Gulen, H. and Schill, M.J. (2008), “Asset growth and the cross-section of stock returns”,
Journal of Finance, Vol. 63 No. 4, pp. 1609-1651.

Dash, S.R. and Mahakud, J. (2013), “Conditional multifactor APMs and market anomalies”, Journal of
Indian Business Research, Vol. 5 No. 4, pp. 271-294.

De Bondt, W.F.M. and Thaler, R. (1985), “Does the stock market overreact?”, The Journal of Finance,
Vol. 40 No. 3, pp. 793-805.

Drobetz, W., Sturmer, S. and Zimmermann, H. (2002), “Conditional asset pricing in emerging stock
markets”, The Swiss Journal of Economics and Statistics, Vol. 138 No. 4, pp. 507-526.

Fama, E.F. (1990), “Stock returns, expected returns, and real activity”, Journal of Financial Economics,
Vol. 45 No. 4, pp. 1089-1108.

Fama, E.F. (1998), “Market efficiency, long-term returns, and behavioral finance”, Journal of Financial
Economics, Vol. 49 No. 3, pp. 283-306.

VOL. 9 NO. 3 2015 JOURNAL OF ASIA BUSINESS STUDIES PAGE 325

 Fama, E.F. and French, K.R. (1988), “Dividend yields and expected stock returns”, Journal of Financial
Economics, Vol. 22 No. 1, pp. 3-25.

Fama, E.F. and French, K.R. (1992), “The cross-section of expected stock returns”, Journal of Finance,
Vol. 47 No. 2, pp. 427-465.

Fama, E.F. and French, K.R. (1993), “Common risk factors in the returns on stocks and bonds”, Journal
of Financial Economics, Vol. 33 No. 1, pp. 3-56.

Fama, E. and MacBeth, J. (1973), “Risk, return and equilibrium: empirical tests”, Journal of Political
Economy, Vol. 81 No. 3, pp. 607-636.

Gourieroux, C., Holly, A. and Monfort, A. (1982), “Likelihood ratio test, Wald test, and Kuhn-Tucker test
in linear models with inequality constraints on the regression parameters”, Econometrica, Vol. 50 No. 1,
pp. 63-80.

Goyal, A. (2012), “Empirical cross-sectional asset pricing: a survey”, Financial market and Portfolio
Management, Vol. 26 No. 1, pp. 3-38.

Griffin, J., Ji, X. and Martin, S. (2003), “Momentum investing and business cycle risk: evidence from
pole to pole”, Journal of Finance, Vol. 58 No. 6, pp. 2515-2547.

Hart, J. van der, Slagter, E. and Dijk, D. van (2003), “Stock selection strategies in emerging markets”,
Journal of Empirical Finance, Vol. 10 Nos 1/2, pp. 105-132.

Haugen, R.A. and Baker, N.L. (1996), “Commonality in the determinants of expected stock returns”,
Journal of Financial Economics, Vol. 41 No. 3, pp. 401-439.

Hausman, J.A. (1978), “Specification tests in econometrics”, Econometrica, Vol. 46 No. 6,

pp. 1251-1271.

Ho, C. and Hung, C.H. (2009), “Investor sentiment as conditioning information in asset pricing”, Journal
of Banking and Finance, Vol. 33 No. 5, pp. 892-903.

Ho, C.W. (2012), “The role of investor sentiment in asset pricing”, Doctoral thesis, Durham University,
Durham.

Hsiao, C. (1986), Analysis of Panel Data, Econometric Society Monographs No. 11, Cambridge
University Press, New York, NY.

Iqbal, J., Brooks, R. and Galagedera, U.A.D. (2010), “Testing conditional APMs: an emerging market
perspective”, Journal of International Money and Finance, Vol. 29 No. 5, pp. 897-918.

Jagannathan, R. and Wang, Z. (1996), “The conditional CAPM and the cross-section of expected
returns”, Journal of Finance, Vol. 51 No. 1, pp. 3-53.

Jegadeesh, N. and Titman, S. (1993), “Returns to buying winners and selling losers: implications for
stock market efficiency”, Journal of Finance, Vol. 48 No. 1, pp. 65-91.

Jensen, G.R. and Mercer, J.M. (2002), “Monetary policy and the cross-section of expected stock
returns” The Journal of Financial Research, Vol. 25 No. 1, pp. 125-139.

Jun, S.G., Marathe, A. and Shawky, A.H. (2003), “Liquidity and stock returns in emerging equity
markets”, Emerging Markets Review, Vol. 4 No. 1, pp. 1-24.

Kothari, S., Shanken, J. and Sloan, R.G. (1995), “Another look at the cross-section of expected stock
returns”, The Journal of Finance, Vol. 50 No. 1, pp. 185-224.

Lakonishok, J., Shleifer, A. and Vishny, W.R. (1994), “Contrarian investment, extrapolation, and risk”,
Journal of Finance, Vol. 49 No. 5, pp. 1541-1578.

Lim, K. and Brooks, R. (2010), “Why do emerging stock markets experience more persistent deviations
from a random walk over time: a country level analysis”, Macroeconomic Dynamics, Vol. 14 No. 1,
pp. 3-41.

Lintner, J. (1965), “The valuation of risky assets and the selection of risky investments in stock portfolios
and capital budgets”, Review of Economics and Statistics, Vol. 47 No. 1, pp. 13-37.

Lischewski, J. and Voronkova, S. (2012), “Size, value and liquidity, do they really matter on an
emerging stock market?”, Emerging Markets Review, Vol. 13 No. 1, pp. 8-25.

Lo, A.W. and MacKinlay, A.C. (1990), “Data-snooping biases in tests of financial APMs”, Review of
Financial Studies, Vol. 3 No. 3, pp. 431-468.

PAGE 326 JOURNAL OF ASIA BUSINESS STUDIES VOL. 9 NO. 3 2015

MacKinlay, A.C. (1995), “Multifactor models do not explain deviations from the CAPM”, Journal of
Financial Economics, Vol. 38 No. 1, pp. 3-28.

Maio, P. and Santa-Clara, P. (2012), “Multifactor models and their consistency with the ICAPM”,
Journal of Financial Economics, Vol. 106 No. 3, pp. 586-613.

Mohanty, P. (2002), “Evidence of size effect on stock returns in India”, Vikalpa: The Journal for Decision
Makers. Vol. 27 No. 3, pp. 27-37.

Moor, L.D. and Sercu, P. (2013), “The smallest firm effect: an international study”, Journal of
International Money and Finance, Vol. 32, pp. 129-155.

Morck, R., Yeung, B. and Yu, W. (2000), “The information content of stock markets: why do emerging
markets have synchronous stock price movements?”, Journal of Financial Economics, Vol. 58 Nos 1/2,
pp. 215-260.

Muradoglu, Y.G. and Sivaprasad, S. (2013), “The effect of leverage mimicking portfolios in explaining
stock returns variations”, Studies in Economics and Finance, Vol. 30 No. 2, pp. 94-107.

Narayan, K.P. and Zheng, X. (2010), “Market liquidity risk factor and financial market anomalies:
evidence from the Chinese stock market”, Pacific-Basin Finance Journal, Vol. 18 No. 2, pp. 509-520.

Pastor, L. and Stambaugh, R.F. (2003), “Liquidity risk and expected stock returns”, Journal of Political
Economy, Vol. 111 No. 3, pp. 642-685.

Rouwenhorst, K.G. (1998), “International momentum strategies”, Journal of Finance, Vol. 53 No. 1,
pp. 267-284.

Rouwenhorst, K.G. (1999), “Local return factors and turnover in emerging stock markets”, Journal of
Finance, Vol. 54 No. 4, pp. 1439-1464.

Schwert, G.W. (2003), “Anomalies and market efficiency”, Handbook of the Economics of Finance,
North Holland, Amsterdam.

Sehgal, S. and Jain, S. (2011), “Short-term momentum patterns in stock and sectoral returns: evidence
from India”, Journal of Advances in Management Research, Vol. 8 No. 1, pp. 99-122.

Sehgal, S. and Tripathy, V. (2005), “Size effect in Indian stock market: some empirical evidence”,
Journal of Business Perspective, Vol. 9 No. 4, pp. 27-42.

Shanken, J. (1990), “Intertemporal asset pricing: an empirical investigation”, Journal of Econometrics,

Vol. 45 Nos 1/2, pp. 99-120.

Shanken, J. (1992), “On the estimation of beta-pricing models”, Review of Financial Studies, Vol. 5
No. 1, pp. 1-33.

Sharpe, W.F. (1964), “Capital asset prices: a theory of market equilibrium under conditions of risk”,
Journal of Finance, Vol. 19 No. 3, pp. 425-442.

Sloan, R. (1996), “Do stock prices fully reflect information in accruals and cash flows about future
earnings?”, Accounting Review, Vol. 71 No. 3, pp. 289-315.

Stattman, D. (1980), “Book values and stock returns”, The Chicago MBA: A Journal of Selected Papers,
Vol. 4 No. 4, pp. 25-45.

Subrahmanyam, A. (2010), “The cross-section of expected stock returns: what have we learnt from the
past twenty-five years of research?”, European Financial Management, Vol. 16 No. 1, pp. 27-42.

Titman, S., Wei, J. and Xie, F. (2004), “Capital investments and stock returns”, Journal of Financial and
Quantitative Analysis, Vol. 39 No. 4, pp. 677-700.

van Dijk, M.A (2011), “Is size dead? a review of the size effect in equity returns”, Journal of Banking &
Finance, Vol. 35 No. 12, pp. 3263-3274.

Further reading
Ferson, W. and Harvey, C. (1999), “Conditioning variables and the cross section of stock returns”,
Journal of Finance, Vol. 54 No. 4, pp. 1325-1360.

Schrimpf, A., Schröder, M. and Stehle, R. (2007), “Cross-sectional tests of conditional APMs: evidence
from the German stock market”, European Financial Management, Vol. 13 No. 5, pp. 880-907.

VOL. 9 NO. 3 2015 JOURNAL OF ASIA BUSINESS STUDIES PAGE 327

 Appendix

Table AI Detail description of firm characteristics or anomalies measurement and hedge portfolio returns
Symbol Firm characteristics or anomalies measurement and hedge portfolio return difference (HPRD)

MC (Banz, 1981) Natural logarithm of market capitalization (stock prices times outstanding shares) at the end of
August of year y. (HPRD ⫽ ⫺2.02)
BP (Stattman, 1980) Ratio between book price for the fiscal year ending in calendar year y by the market value of
equity at the end of August in year y. (HPRD ⫽ 3.21)
E/P (Basu, 1977) Ratio of net profit for the fiscal year end March to the market capitalization at the end of the
August of year y. (HPRD ⫽ 1.53)
C/P(Lakonishok et al., 1994) Sum of earnings before extraordinary items and depreciation over the firm’s market
capitalization at the fiscal year-end March of year y. (HPRD ⫽ 1.59)
D/P (Fama and French, 1988) Ratio of dividend paid by the firm for the fiscal year end March to the market capitalization at
the end of the August of year y. (HPRD ⫽ ⫺0.47)
SG (Lakonishok et al., 1994) Ratio of net sales revenue for the March in the year y over the sales revenue from the March of
the year y-1. (HPRD ⫽ ⫺1.72)
AC (Sloan, 1996) Change in noncash current assets less the change in current liabilities income tax payable)
less depreciation, during the fiscal year ending in year y-1, scaled by the average total assets
at the beginning and end of that fiscal year in the calendar year y. (HPRD ⫽ 0.84)
MOM (Jegadeesh and Titman, Cumulative return of a stock in month t-12 through month t-2 preceding August of the year y.
1993) We skip one month between portfolio formation and holding period to avoid the effects of bid-
ask spread, price pressure and any lagged reaction. (HPRD ⫽ 3.06)
LR (De Bondt and Thaler, 1985) LR of a stock in the month t is measured each month by sorting stocks on past returns from
month t -36 through t -7. For the LR, the portfolio return has been constructed from September
1997 to March 2011, given the 36 month lag. (HPRD ⫽1.01)
RDint (Chan et al., 2001) Ratio of research and development expenditure for the fiscal year ending in calendar year y to
market capitalization at the end of August of year y. (HPRD⫽⫺0.42)
AVint (Chan et al., 2001) Ratio of advertising expenditure for the fiscal year ending in calendar year y over market
capitalization at the end of August of calendar year y. (HPRD ⫽ 0.82)
LIQ (Amihud, 2002) Annual average of monthly turnover ratio i.e. number of shares traded to the number of shares
outstanding at the end of August of calendar year y. (HPRD ⫽ ⫺0.73)
Bliv (Artmann et al., 2012) Total assets divided by book equity of the fiscal year ending in year y. (HPRD ⫽ 1.61)
Mliv (Artmann et al., 2012) Total assets in the fiscal year ending in year y divided by market value of equity at the end of
August of the year y. (HPRD ⫽ 1.43)
ROA (Artmann et al., 2012) Net earnings divided by total assets for the fiscal year end March in the calendar year y.
(HPRD ⫽⫺0.27)
CAP (Chen et al., 2010) Capital expenditure for the fiscal year ending in calendar year y over the average total assets
at the beginning and end of that fiscal year in calendar year y-1 and y. (HPRD ⫽ ⫺0.59)
AG (Cooper et al., 2008) Percentage change in total assets from the fiscal year ending March in calendar year y-2 to the
fiscal year ending in calendar year y-1. (HPRD ⫽ ⫺1.52)
INVT (Titman et al., 2004) Annual change in gross fixed assets plus annual change in inventories in the fiscal year ending
March in year y divided by book value of total assets of the fiscal year ending in year y-1.
(HPRD ⫽ ⫺1.51)
Notes: This table reports the detail description on the measurement of the firm characteristics and the hedge portfolio return variation of
decile portfolios shorted on the firm characteristics; the measurement of firm characteristics closely follows the approach of Artmann et al.
(2012), Chen et al. (2010) and Dash and Mahakud (2013); for the purpose of brevity, we only report the HPRDs; HPRD is the difference
between tenth decile (i.e. largest) and first decile (i.e. smallest) of the portfolio; sample period except for long-term reversal (LR) is from
September 1995 to March 2012; since NSE started its operation in 1994, sample period for LR is from September 1997 to March 2012; except
for long-term reversal and momentum stocks are allocated to ten portfolios at the end of August of each year y (1995-2012); the portfolios
formed on momentum (long-term reversal) are rearranged every month; monthly equal-weighted returns on the portfolios are calculated from
September to the following August; to avoid the look ahead bias, five months gap has been provided to match the accounting data available
at the end of financial year end March in calendar year y to the stock price data at the beginning of September in the calendar year y

Corresponding author
Saumya Ranjan Dash can be contacted at: saumyadsh@gmail.com

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com

PAGE 328 JOURNAL OF ASIA BUSINESS STUDIES VOL. 9 NO. 3 2015

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