Vol. 8(8), pp.

127-141, October, 2016

DOI: 10.5897/JEIF2015.0715
Article Number: 4A0F08C61306
ISSN 2006-9812 Journal of Economics and International Finance
Copyright 2016
Author(s) retain the copyright of this article
http://www.academicjournals.org/JEIF

Full Length Research Paper

The pricing of illiquidity risk on emerging stock

exchange markets: A portfolio panel data analysis
Hikouatcha Kenfack*1, Prince Dubois1, Kamdem David1, Bidias Menik Hans Patrick1 and
Adeyeye Patrick Olufemi2
1
University of Dschang, Cameroun.
2
Graduate School of Business & Leadership, University of KwaZulu-Natal, Westville, Durban, 4000, South Africa.
Received 3 September, 2015; Accepted 23 December, 2015

The main purpose of this paper is to examine the main role of liquidity in stock pricing on African
emerging stock markets. The study applies portfolios panel data analysis to modify and adapt the
existing estimation process. Using three different procedures, six portfolios have been constructed
base on the 32 most active stocks on the so called BRVM; the measures of liquidity considered are the
turnover and the illiquidity ratios. To reach our objectives, we first of all verify if liquidity is taken into
consideration in the explanation of expected excess return. Secondly, we verify whether liquidity risk is
correctly priced on BRVM. The results indicate that from 1998 to 2008, whereas liquidity is correctly
taken into account in equity pricing, there is no significant evidence that liquidity risk is priced on the
BRVM. These conclusions remain stable even when various tests of robustness are undertaken and
they are not consistent with results obtained by the authors on developed stock markets. These results
may be explained by the microstructure of the BRVM.

Key words: Liquidity, liquidity risk, expected return, emerging stock exchange market.

INTRODUCTION

The risk problem remains pervasive in the field of price valuation. This approach consequently shows that
Finance. Although Knight (1921) is one of the first to the only risk to be taken into account in the explanation of
integrate this concept into investment decisions, the the expected returns is the one related to the market with
modern formulation in the heart of the financial literature the underlying assumptions that the market is efficient
is generally ascribed to Markowitz (1952), who suggested and that there are no transaction costs. This position is
that standard deviation (sigma) should be an adequate not without criticisms, as one can note that the market is
measurement of total risk or business risk. not the only factor to be taken into consideration; and that
However, Sharpe (1964), through his famous Capital there is no perfect market in which transaction costs are
Asset Pricing Model (CAPM), proposed beta as a completely non-existent. It is in response to this criticism
measure of systematic or non-diversifiable risk of that Ross (1976), through the Arbitrage Pricing Theory
financial assets, thus turning the problem into the stock (APT), proposed that there are several factors of different

*Corresponding author. E-mail: duprin@yahoo.fr.

Authors agree that this article remain permanently open access under the terms of the Creative Commons Attribution
License 4.0 International License
128 J. Econ. Int. Finance

nature, which must be considered in the explanation of explanatory factors of the shares expected returns have a
the prediction of stock prices. Fama and French (1993) base on the standard pricing models of financial assets.
suggested that one should not neglect the factors Thus, the two factor model of Amihud and Mendelson
suitable for the size and the value of listed companies. (1986) and Liquidity Capital Assets Pricing Model
Other scholars (Amihud and Mendelson, 1986; Pedersen (LCAPM) of Archarya and Pedersen (2005) acted as a
and Acharya, 2005; Brennan et al., 1998, 2001) proposed base on the Capital Assets Pricing Model (CAPM) of
the need to take into consideration certain variables Sharpe (1964), while the four factor model used for the
suitable for the microstructure of financial markets and first time by Chordia et al. (1998) before spreading itself
more precisely, liquidity in the explanation of the largely with this field of research has a theoretical base of
expected returns of financial assets. Fama and French (1993) model.
Consequently, the investment decisions should not only
consider the risk and return, but also the liquidity of the
assets concerned. Liquidity, contrary to accounting and The two-factor model (The clientele theory)
banking interpretation, can be seen as the capacity, the
ability, or the skillfulness to proceed to significant Amihud and Mendelson (1986) suggested liquidity as an
transaction on shares at lower costs, without any essential element of stock price formation process.
significant influence on the prices of the assets According to them, a patient investor must profit from an
concerned (facilitated transaction). About it could also extra return (illiquidity premium) in remuneration of his
mean the facility with which an operator in the market patience, liquidity being evaluated by transaction costs.
finds a counterpart for his operations. The basic idea is as follows: An investor (indifferent
Many studies have been carried out on the American from risk), who buys a share and considers possible
stock exchange market where liquidity and risk are well transaction costs must take them into account at the time
taken into account in the formation of stock prices but the he wants to resell. On the basis of this idea, the
results are very divergent in other developed and proposition shows that the required return (rate of return),
emerging markets. The latter however, presents for an investment in the illiquid assets (I), is given by the
particular characteristics (they are said to be very narrow, output of the investment of perfectly liquid assets (RL)
illiquid and quotations are far from frequent there) which added to a premium of lack of liquidity represented by the
offer a good verification opportunity because, according ratio of the cost of lack of liquidity ( Ci ) by the price of the
to Hichan (2007), no investor worthy of his name could
action ( Pi ).
offer the luxury to be unaware of them. Liquidity is one of
the principal factors that should attract investors to the
1
financial market considering the clientele theory (Amihud Ci
and Mendelson, 1986).
Ri RL
Pi (1)
Studies in this field has almost left out the African stock
market, which however presents rather relevant
characteristics, since it is famous for being far from liquid. Equation 1 is obtained under the assumption that
It is in sense that this study seeks to examine the investors have a single-period investment horizon. By
contribution of liquidity and the associated risk in the supposing a multi-period situation, one can generalize
explanation of stock price valuation process, in order to this result based on the fact that the price of an asset,
draw a set of lessons from it and to compare the results which indefinitely offers constant dividends (Di) is given
with those found in the context of the American stock by the present value of these dividends minus costs
exchange market and other developed financial markets. expected for lack of liquidity:
Moreover, in addition to this contextual and theoretical
Di Ci C
2
Di
involvement, one major contribution of this study is Pi Ri RL i
RL Pi Pi (2)
essentially methodological. Indeed, it is shown here that
the choice stock portfolio panel analysis, makes it
possible to reduce the heaviness of the work which is is the probability of liquidating the concerned asset.
carried out with the standard form of the Fama and
Macbeth method like Rahim et al. (2006). More According to Dalgaard (2009), this relation is purely
specifically, it makes it possible to cancel the third stage intuitive insofar as it is known that the return obtained on
which generally is most tiresome, reducing the number of an asset must take into account the expected "liquidity
stage of this method to two either to three. premium". This relation presumes that the investor
behaves as the periods of investments were
homogeneous. By taking the case of heterogeneity for
THE THEORETICAL FRAMEWORK the duration into consideration, this can be condensed
into what is described as "clientele theory", according to
All theories developed in this approach of the liquidity as which a long-term investor will have a weak frequency of
 Kenfack et al. 129

transaction, which will consequently reduce transaction is that of Fama and French (1992), taking into account
costs (costs of illiquidity) and then allow him to realize a the adjustments suggested by Litzenberger and
higher output after taking into account the costs relating Ramaswang (1979). The results obtained show overall
to it. This simply means that the least liquid assets must that the expected return is a decreasing function of
be held by the investors over a long investment horizon. liquidity. More precisely, it is positively and significantly
Algebraically: related to turnover contrary to the indicator of illiquidity
which must be negatively related to return with respect to
the Amihud and Mendelsons (1986) clientele theory.
ERi RL RJ RL j 3
Ci
Pi Chordia et al. (2001) study the effect of variation of
(3)
liquidity on the output of shares quoted on the NYSE and
the AMEX from January 1966 to December 1995.
RJ , represents the return required by the investor J. This According to them, insofar as the relation makes a
two-factor theory is the basis for other extensions. consensus in the literature, investors should protect
themselves from the risk related to its variation, since
they are supposed to be risk averse. The empirical
The four-factor model method used is that of Brennan et al. (1998), which binds
the expected output to volatility or liquidity. The results
Chordia et al. (1998) following the work of Fama and show that the effect of liquidity measured by turnover and
French (1993), proposed a model which makes it the ratio of lack of liquidity on the output is significant and
possible to take into account the existence of liquidity in persistent, which means that return is a decreasing
the explanation of returns in excess of the risk-free rate function of liquidity in the American financial market.
for a portfolio of assets. They showed that this exceeds Chan and Faff (2005) highlight the role of liquidity in the
the two-factor model in terms of significance, which financial valuation of assets by increasing an explanatory
explained return more than the CAPM. It is represented variable relating to liquidity. The technique used is that
in the following equation: of Fama and French (1993) and the monthly data used
relates to the shares quoted on the Australian market
ERit Rft it it ERit Rft sit SMBit hit HMLit lit LIQit it 4 over the period 1989 to 1998.
(4) The empirical analysis shows that liquid assets ratio is
significant just like on the American market. This means
lit is the coefficient which measures the sensitivity return that, the turnover tends to have an effect on the expected
compared to the liquidity. More details on this will be returns of shares. Moreover, they show that the least
provided in the methodology. liquid shares tend to have significantly positive beta while
the most liquid portfolios tend to have negative betas.
These results are robust and are consistent with financial
Liquidity, illiquidity risks and rate of return required literature on the American market.
by the investors: a review of empirical results The object of the work of Soosung and Lu (2005) is to
give an insight on the role of the liquidity risk in the
Amihud and Mendelson (1986) propose the role of explanation of the expected return of shares on the
liquidity in financial assets pricing. In their empirical British stock exchange market. Particularly, they seek to
study, portfolios are constituted on the basis of individual check if the liquidity or the value of the liquidity premium
systematic risk and value of the liquidity indicator (the is well taken into account as on the American market.
bid-ask spread) of the shares quoted on the NYSE from They use variables of the model of Fama and French
1960 to 1980. The results obtained allowed them to (1993), and the indicator of liquidity of Amihud (2002)
confirm that liquidity is a decreasing function of the over a period of study from January 1987 to December
expected returns and also, that this function is concave. 2004. The results show that irrespective of the strategy of
Archarya and Pedersen (2005) use lack of liquidity as portfolio formation, large companies are more powerful
indicator to carry out an analysis in 5 steps to show that than the small ones and the direction of relation between
liquidity and the risk of illiquidity are very significantly return and liquidity is contrary to that highlighted in the
taken into account in the explanation of the expected literature, although the latter is taken into account on this
returns of shares on NYSE and AMEX for a period from market. That is true whatever the test of robustness
1962 to 1995. Also, Datar et al. (1998) used turnover as carried out.
indicator to test the role of liquidity in the process of stock Shing-Yang (1997) examines the problem of
exchange price formation on the NYSE between July measurement of liquidity and its impact on the surplus of
1962 and December 1991. expected return of shares on the Tokyo Stock Exchange
The fundamental difference in this study with all the between 1975 and 1993. The results show that the
others is that it relates to individual shares rather than stocks which have a raised turnover tend to have a
portfolio of shares usually used. The method employed weaker expected return (adjusted with the risk). Also,
130 J. Econ. Int. Finance

Geert (1998) studied 1700 shares of a group of 20 time t ( PJt , Djt represents respectively the price and the
emerging stock exchange markets over the period of dividend of share j at time t).
1975 to 1997. The measurement of liquidity used is the The risk free rate ( R ft ) used is the interbank rate, because the
turnover and the other explanatory variables are those of
public treasury bonds rate (in one year) regularly used in the
Fama and French (1992). He finds out that the liquidity
literature is very unstable and irregular in the UEMOA zone. The
is strongly related to the expected return in excess of the use of the simple averages (arithmetic mean) contrary to the
risk free, but that it is very unlikely that this is taken into weighted averages joined the works of Amihud and Mendelson
account overall on these markets. (1986), Amihud (2002) and Chordia et al. (2001). The reason is
Furthermore, Bekeart et al. (2007) show that liquidity simple: In theory, the use of weighted averages makes it possible to
becomes one of the most significant factors in the stock take into account the size effect of large companies which could
skew the results, but that is already taken into account when using
exchange prices formation process of the 19 emerging stock exchange capitalization (book to market) as explanatory
stock exchange markets, contradicting more or less the variable.
results of Hearn and Piesse (2005), and Hearn (2007) on For the explanatory variables, we use the three variables
Africa in the South of Sahara. suggested by Fama and French (1993). It concerned the return of
These results aligns with those of Huson (2009) on the the market portfolio in excess of the risk free rate (RMt Rf). One of
Malaysian stock exchange market, Dalgaard (2009) on the problems in the determination of the market portfolio is that it
must be the most possible representative. To face this problem the
Denmark, Barend (2009) on the South-African market study used the BRVM composite index which represents the whole
and Miralles et al. (2011) on the Portuguese market, in securities quoted on the BRVM. The market portfolio return is the
that, although liquidity is a significant factor in the stock same as that in the CAPM, being the ratio of the difference between
exchange price formation process, the pricing of the the value of the market index ( I ) between two consecutive
associated risk is long in becoming a reality periods, by that of the previous period:

It It 1
METHODOLOGY RMt
It 1
Source of the data, selection of the variables and portfolios
formation The size of the company or the portfolio here makes it possible to
take into account the "the size effect", which supposes that the
The data large companies or large portfolios will have higher returns as long
as their sales are high. This effect is taken into account in this study
Data used in this study are obtained from the financial database of by the stock exchange capitalization (CB), which is the product of
BRVM. They are primarily the data on the daily transactions and the number of title j held at moment t and the stock exchange
those on the annual financial statements of the companies quoted average price of this same security over the same period.
at the BRVM. After several filters, in particular with regard to the
incompleteness and the absence of information, a sample of 32
titles is retained for one period which extends from 1998 to 2008. CB jt NAjt .CMB jt
It should be noted that it concerns particularly the 32 most traded
CBjt
1
shares on this stock market during this period. It should also be
noted that the study use the average value of information (the
CBit
n
monthly average price of share) over the period considered, which
is more representative.
CBjt , NAjt and CMBjt respectively represent stock exchange
capitalization, the number of shares held and the average price of
the share j during the month t.
The variable of the study
In line with the study of Amihud and Mendelson (1986) and
Batten et al. (2010), the value of stock exchange capitalization used
The explained variable is the return in excess of the risk free rate
is log-normalized to take into account the imperfections of the
( Rit Rft ). The return of portfolio i is a simple average of that of market.
the individual shares which constitute it. It is calculated according to The "book to market" (BM) takes into account the variations of
the method suggested by the CAPM, while taking into account the the company. It is about the relationship between two values of
net dividends produced by each share, as seen below: the company: The book value (VC) and the market value (VM) for
share j at time t:

R Rft , with n being the number of share

1 n
Rit Rft VC jt
n J 1 j BM jt
in a portfolio i.
VM jt
To obtain the value of the BM for a portfolio, it is enough to make
PJt Pjt 1 Djt the simple average of these individual values for the shares which
RJt , measures the profitability of share j at
make it up in the month considered.
The study used not only one measurement of liquidity, but also
Pjt 1 an indicator of lack of liquidity. Several studies used the "bid-ask
 Kenfack et al. 131

spread" as an indicator of lack of liquidity but on most emerging Finally, there are six (6) portfolios as presented in Figure 1 below. A
stock markets like the BRVM, one cannot have the required particular attention is also given to the period of study because it
information for its determination because a system of order and not can rather be special in the methodology of Fama and Macbeth
of price quotation is still being used. It is for this simple reason that (1973), in the sense that it gives the possibility of being broken up
other indicators like the turnover or the ratio of illiquidity are used into sub-periods with an aim of not only making the model more
here to measure liquidity. predictive but also of making the results more significant and better
Datar, Naik and Radcliffe (1998) proposed Turnover (TO) as a in explaining the phenomenon studied. Typically, with the present
measure of liquidity to mitigate the insufficiencies of the "bid-ask case and in reference to the literature on this methodological
spread". For them, it is the ratio between the volume of transaction approach, our period of study which is 10 years (1999 to 2008) is
of the title (j) during the month t by the number of title held or subdivided in two sub-periods: one estimated period (of the betas
available during the time and it is given by: for each explanatory variable), and a test period which will make it
possible to estimate the factors of sensitivity to be used to explain
Vjt the studied phenomenon (gammas).
TOjt Thus, the first four years (1999 to 2002) are used to estimate the
NTDjt betas and the last six (2003 to 2008) years to estimate the
gammas.
TOjt , Vjt , NTDjt , respectively represent turnover, volume of In addition, in order to obtain the most significant and
representative possible results of the sample, the beta of Sharpes
transaction and the number of securities held for stock j during market model which is used for the formation of the portfolios is
month t. To calculate the turnover of a portfolio (i) of security (j) calculated for each title on the whole for the period of study1.
simply involves making simple average of the turnover of shares The comprehension of all these processes and techniques of
included in the portfolio. There are several advantages to use the analysis is easier when these variables are specified in an
turnover as an indicator of liquidity. According to Xuan et Batten econometric model.
(2010), it has a very strong theoretical attraction; Amihud and
Mendelson (1986) showed that at equilibrium, the return is
correlated with the transaction frequency. The econometric specifications
Thus, if one cannot directly observe the liquidity by the bid-ask
spread, that can also be done in an effective way with turnover. In It will be necessary to distinguish the econometric model, which
second place, the turnover is one of the most easily calculable makes it possible to detect the role of the liquidity from what makes
measurements, if one takes into account the system of functioning it possible to take into account the risk of liquidity in the explanation
of emerging stock exchange markets in general, and the BRVM in of the expected return of stock portfolios.
particular. It is supposed to be a decreasing function of the
expected output. Due to the various difficulties that arise in the
evaluation of liquidity, Amihud (2002) proposes an indicator, which The role of liquidity in the stock pricing process
can readily be obtained at the emerging markets and this has
gained a great importance in the literature. It is presented in the To detect the role of liquidity in the explanation of expected return,
form of a relationship between daily return of asset j in the month t we go through the following procedure:
of year n by the volume of corresponding transaction (in value). Firstly, we regress with ordinary least square for each portfolio i,
Nevertheless, this indicator is slightly modified (without changing its each month t, using the following model:
Rit Rft i it RMt Rft sit CBt hit BMt it
direction and its significance) in that instead of working with daily
data as, we rather used monthly values. Algebraically one can
write: t= 1, , 49
For each portfolio i one estimates each month as from January
2003, the parameters it ,
Rjtn sit and hit Each one of these parameters
il jtn
Vjtn is obtained by making the regression of the preceding model on
information from the previous 49 months. The estimated value of a
il jtn , Rjtn and Vjtn respectively represent the ratio of lack of parameter at January 2003 is obtained from the regression of the
liquidity, the output and the volume of transaction for each title j in preceding model using data from January 1999 to January 2003.
the month t of year n. The objective of this first regression (over period of estimate) is to
obtain the monthly value of the btas ( it , sit and hit ) of each
Each one of these variables are calculated for each individual stock
before being aggregated on the level of the portfolios. The portfolio company characteristic, which will be used at the second stage to
of stocks are used as against individual stocks simply because this explain the role of liquidity in the explanation of the return of the
process makes it possible to reduce skewness in the estimate of portfolios of shares.
the explanatory variables. Just like in Fama and French (1993),
we form six (6) portfolios according to a double assignment. The Secondly, the factors of sensitivity ( it , sit and hit ) obtained by the
method of formation adopted is similar to that of Amihud and
Mendelson (1986) and Dalgaard (2009) with regard to the ordinary least square are used simultaneously with the liquidity in a
technique and the proxies of formation of stock portfolios. Initially, panel of six portfolios. The model at this second phase becomes:
the 32 BRVMs shares are classified according to the decreasing
value of their liquidity then divided into three groups by considering Rit Rft o 1it 2sit 3hit 4 LIQit i
the thresholds of 30% and 70%. In the second time, each one of
these three groups is segmented into two equal partitions
compared to the value of beta of each stock constituents of the
portfolio. Beta used is obtained from the regression of market
model of Sharpe (1964).
1
Rjt Rf t j j RMt Rft jt ,.................t 1,...,120
132 J. Econ. Int. Finance

Figure 1. The derivation procedure of the panel of six Portfolios.

t= 1, , 72 et i= 1,, 6 (Model 1) Now, when we want to check if the associated risk (risk of liquidity)
LIQ represents the liquidity variable which is measured either by is priced, this one is introduced at the first stage, such that a beta
turnover or by the ratio of illiquidity. value is also calculated for the liquidity variable. Just like previously,
At this level, because instead of individually making the the analysis is done on portfolios of shares instead of on individual
regression for each portfolio before aggregating thereafter the shares.
results at the level of the overall market, we choose a form of The first regression which makes it possible to obtain the factors of
regression 2 which makes it possible to make both in one, while sensitivity is made for each portfolio i at the moment t, according to
reducing the risks of omission and calculation which can affect the the following model:
stage of aggregation of the results in the procedure. This second
regression (which makes it possible to really explain the studied Rit Rft it it RMt Rft sit CBt hit BMt Ilit LIQit it
relation) is made on all of the test period, which extends from t = 1, , 49
January 2003 to December 2008. The indicator of liquidity As previously mentioned, the model from the first stage already
(turnover or ratio of lack of liquidity) used for each portfolio is that of integrates liquidity which is consequently considered as a risk and
the preceding period (month); the use of the indicator of liquidity of no more in terms of degree (level of liquidity). Thus, the sensitivity
the previous period aims to make the model more predictive. of the return to the changes in liquidity is the time taken into
account at the first stage of the regression (Fama and Macbeth,
The first gamma ( 0 ) does not represent anything theoretically and 1973).

, 2 and 3
Then, the whole of the parameters of the portfolios obtained at the
first stage (factors of sensitivity it , si , hi ,
must however be null. The 3 gammas following 1 Ilit ) are used in the
represent the risk premiums and consequently must be different
regression as follows:
from zero according to the relation which exists between liquidity
and expected output. The last gamma ( 4 ) represents the effect of
Rit Rft o 1it 2sit 3hit 4lit it
liquidity in the explanation of the excess expected return, the
awaited sign of this relation is a function of the indicator used. If it
is about a measurement of liquidity like the turnover, then the t = 1, , 72 and i = 1,, 6 (model 2)
relation should be negative since the less liquid a portfolio is, the The regression parameters are the risk premiums, which explain
higher is the required return. If on the other hand, the indicator the disclosure of the risk of each portfolio. So the portfolio with a
used for the liquidity is an indicator of lack of liquidity then the high sensitivity on one or the other of the four factors, must expect
awaited sign becomes positive. the highest return in theory to remunerate the additional risk to
which it is exposed. This is verified and is simply perceptible in the
CAPM where portfolios of shares which are most sensitive to the
The illiquidity risk and the stock price formation process: the variation of the market variable (shares with high beta) are
model supposed to have a higher return.
In short, the method used here is that of multiple regressions in
In this work, in addition to the direct pricing effect of liquidity on the the first phase and panel analysis in the second phase (that
return of individual shares, we analyse the relationship between enables us to check the sense of the relation between liquidity and
liquidity risk and return of portfolios of shares on the BRVM. To return). More clearly and simply, one first estimates by the ordinary
evaluate the impact of liquidity in the explanation of expected least square method the betas or factors of sensitivity for each
return, its value is introduced at the second stage of methodology. variable which is supposed to explain the expected return of the
portfolios of shares and in the second time, one estimates the
gammas by the panel regressions.
2
Panel datas analysis Pre-test and post-estimation tests were carried out to ensure the
 Kenfack et al. 133

Table 1. Descriptive statistics (average).

3X2 TO-beta portfolios

Portfolios P1 P2 P3 P4 P5 P6
Return 0.017409 0.108725 0.048234 0.016418 0.033185 0.094929
Turnover 0.003761 0.004620 0.001205 0.001528 0.000387 0.000474

3X2 it-beta portfolios

Return 0.055633 0.200257 0.040934 0.021533 0.016204 -0.011480
Amihud (2002) 0.030975 0.013964 0.003752 0,004022 0.000853 0.00125
illiquidity

06 beta portfolios
Return 0.059268 0.024487 0.031382 0.009876 00.036214 0.16058
Turnover 0.001453 0.001899 0.003325 0.001334 0.002309 0.001318
Amihud (2002) 0.002467 0.021315 0.005250 0.003169 0.010277 0.0111134
illiquidity
The liquidity tends to be a decreasing function of expected return in excess of the risk free rate on the BRVM (Source: the
author starting from the data of the BRVM and the exits of Eviews software).

robustness of the estimates. These include, among others, Breitung between 1,6418% for portfolio 4 (P4) to 10,8725% for
(2000) unit root tests to account for stationarity, Breush-Pagan test portfolio 2 (P2). These values confirm the assumption of
to account for heteroscedasticity and Hausman test to account for
fixed and random effect of the various panels respectively.
a strong variation and very high value of the returns on
At first, estimate of the factors of sensitivity is focused on the 32 the African emerging stock exchange markets globally,
liquidity-beta portfolios but this can appe0r skewed and less and on the BRVM in particular. Moreover, it is equally
specified. For this reason, to check the significance of the results, noted that Sharpes theory of risk premium tends to be
same work is restarted for portfolios formed starting from beta of respected, because the first three portfolios, which are
shares only. Still aimed at checking if the liquidity and the the riskiest are also most profitable. Also, portfolio 2 that
associated risk are significant factors in the explanation of the
expected return, the variables of Fama and French are removed
has the highest liquidity value measured by the turnover
from the model and, on the other hand, in line with the Mooradian (TO = 0,004620) has the highest output (0,108725). But,
(2010) model, the two indicators of liquidity are simultaneously used portfolio 5 which has one of the three lowest returns is
in the same model. The objective is not to make the model more the least liquid on average. This leads s that return and
or less explanatory but, simply to check if the relation between liquidity (turnover) move in the same direction. The pace
liquidity and return remains unchanged when the model loses or
of this relation is confirmed even more when it is noted
gains power. Again, in order to take into account the January effect,
information on the first month of the year is eliminated for each that the three most liquid portfolios are also the most
portfolio of the study. profitable. In addition, it is clear in terms of the average
values that liquidity is an increasing function of the output
of stock portfolios quoted on the BRVM.
RESULTS Table 2 shows that the result almost does not change
when the indicator of lack of liquidity is used as proxy.
The descriptive statistics More specifically, the two least liquid portfolios on
average (P1 = 0,030975 and P2 = 0, 013964) are most
Here, we analyse the relationship between return and profitable (P1 = 0,055633 and P2 = 0, 200257).
liquidity using the average of the observations according Additionally, starting from the second portfolio, the
to each portfolio formation method. The same process average output decreases gradually until the sixth.
is repeated with correlation coefficients between the However, the method of formation of the portfolios shows
various variables of the study and according to the that lack of liquidity decreases from P1 to P6. Thus, P6
various methods of portfolio formation. Table 1 brings out would be most liquid; this is confirmed owing to the fact
the average of each liquidity variable and return in excess that the last two portfolios (P5 and P6) hold the lowest
of the risk-free rate. From this table (3X2 TO-beta average values of the ratio of illiquidity and consequently
portfolios), one can note that while the average value of are most liquid. And also, by the fact that the more the
the market return adjusted with the associated risk isle portfolio is profitable, the more the average value of the
negative (approximately -4,5%), the return adjusted to the ratio of lack of liquidity is decreasing.
risk of all the portfolios is completely positive and varies For the third method of formation of the portfolios
134 J. Econ. Int. Finance

Table 2. Turnover as liquidity indicator.

Obtained results
Panels Parameters Sign waited
Coefficient (%) P< |z| (%)
0 none 61.62* 0

1 + 2 86.4

3X2 TO-beta portfolios 2 - -11.1** 3

3 + 20.05** 3

4 - -143.28 10.6

0 none 10.53** 1.1

1 + -7.63*** 8

06 Beta portfolios (TO) 2 - 3.77* 0.2

3 + -4.51 30.3

4 - 12.88 79.7

0 none 8.09* 0

Without Fama and French (1992) factors (TO) 1 + -3.37 23.1

4 - -126.64 23.4

0 none 7.02* 0

1 + -0.69 44.1

January effect (TO) 2 - -11.20* 0.6

3 + 17.92** 2.3

4 - -154.28*** 9.5

(Table 3), the results are almost identical, but since the Sharpe principle in particular, with regards to the direction
two variables of liquidity are at stake here, it is noted that of the relation.
the more the average value of turnover (TO) increases, 2. The relation between the expected output of the
the more that of the ratio of lack of liquidity (IT) decreases portfolio and the variable which takes into account the
for all the portfolios. This proves that these two variables size of the company in the model (stock exchange
move in opposite directions. capitalization (CB)), is positive overall irrespective of the
From the foregoing, it is clear that liquidity is well taken method of formation of the portfolios. This shows that in
care of on the BRVM and the related risk is priced. In theory, the larger the firm the more significant will be its
other words, liquidity would be a decreasing function of expected return. This analysis is against the results of
the output because on average, the most liquid portfolios Fama and French (1992, 1993). According to the latter,
are the least profitable. The analysis of the correlation the shares with weak stock exchange capitalization must
coefficients between the variables shows us a set of profit from a greater risk premium related to their size. In
results: our case, this result is obtained only when the portfolios
are formed from the individual betas of shares.
1. The Sharpes variable appears to be the most 3. The Book to Market (BM) variable which takes into
significant variable in the explanation of the expected accounts the valuation of the firm whose share is quoted,
return, since it has the highest correlation coefficient is for most portfolios negatively related to the expected
(between 18% for P2 and 57% for P3), which respects the return adjusted to the risk. This consolidates the results
 Kenfack et al. 135

Table 3. Amihud (2002) illiquidity ratio.

Results obtained
Panels Parameter Sign waited
Coefficient (%) P< |z| (%)
0 None 8.87** 2.5

1 + -4.4* 0.3

3X2 il-beta portfolios 2 - -9.71 13.4

3 + 20.74 12.3

4 - 60.21*** 8.2

0 None 9.62** 1.3

1 + -8.35** 4

06 Beta portfolios (il) 2 - 3.66* 0.6

3 + -3.83 36.5

4 - 115.79** 3.5

0 None 8.29** 1.8

Without Fama and French (1992) factors (il) 1 + -2.45*** 8.7

4 - 56.32*** 7.4

0 None 8.35** 3.1

1 + -2.82** 3.9

January effect 2 - -9.07 17.4

3 + 19.25 15.5

4 - 64.33*** 7.9

0 None 9.58** 1.4

1 + -8.84** 4.1
2 - 3.67* 0.6
TO and il simultaneously
3 + -3.86 35.4
4 - 18.04 73

5 + 116.01** 3.5

of Fama and French (1992, 1993). liquidity. This clearly respects for almost all the portfolios
4. Regarding liquidity, the turnover, in spite of some whatever the method of their formation. Important
contrary tendencies, presents a negative relationship to information that is provided in this matrix concerns the
the expected return. On the contrary, the relationship to relation between the two liquidity variables. For
the liquidity measured by the ratio of lack of liquidity Mooradian (2010), the direction of the relation is negative
completely respects the literature, since the expected when they are of different nature. But in the case of
return of the portfolio is an increasing function of lack of BRVM, this relation varies in a negative way between 0,
136 J. Econ. Int. Finance

06% and 18%. (2002) finds contrary results to those of Fama and French
In conclusion, these statistics enable us to say in (1992, 1993) not only for the size of the company, but
advance that the illiquidity risk correctly appears to be also concerning the "Book to Market" which takes into
priced on the BRVM. account the value of the listed company.
Another significant parameter in the explanation of the
BRVMs expected return on stock portfolios is that which
The liquidity role in stock pricing process takes into account the risk related to the correct valuation
of the company ( 3 ). An increase by 1% of this factor is
After the various corrections recommended by the results
at the origin of an increase of approximately 20% on
of the econometric tests we seek to detect the place of
expected return in excess of the risk-free rate. Several
the liquidity factor in the explanation of the expected
return by estimating Model 1. The results are a function authors found this same result which had first been
of the variable of liquidity used. proposed by Fama and French (1992). For instance,
Miralles et al. (2011) find that contrary to stock exchange
capitalization (market value), risk related to the market
The turnover as an indicator of liquidity value is an increasing function of the share portfolios
return. Saud et al. (2013) found that the value of the
To evaluate the contribution of liquidity (measured by parameter 3 is very high compared with that of 2 . Also,
turnover) in the pricing of stocks, the regression of the Molay (2002) and Mourina et al. (2012) show that this
regression of Model 1 are presented in the following relation is rather opposite with the results of Fama and
table: One can note that all the gammas have the French (1992, 1993).
expected signs but only three are significant including the Lastly, gamma estimates ( 4 ) which is that on which
constant ( o ). This last is significant at 1%, a proof that
this study is based has a negative value. This value is
even if the model is overall significant at 1%, certain highest in absolute terms among all the gammas of the
variables must be omitted in the explanation of the model. For an increase of 1% of the level of liquidity of
expected return on the BRVM. This does not respect our the portfolios of shares, there will be a reduction of more
expectations (it was supposed that constant is null), but is than 100% of the expected return. It is the only
significantly in line with Rosss (1976) work on the phenomenon that is non-significant. Nevertheless, as
existence of a non-quantifiable number of factors to be regards the direction of the relation between liquidity and
taken into account in the shares (portfolios of shares) expected return in excess of the risk-free rate, it respects
return forecast. In addition to the constant, the other existing results in that liquidity is a decreasing function of
gammas respect the underlying assumptions. The value output.
of the coefficient 1 which explains the relation that exists Thus, the most liquid shares are the least profitable;
between the expected return of portfolio i and the return agreeing with the clientele theory such that investors who
of the market is approximately 0, 2%. This means that a prefer the most liquid shares are short-term investors.
1% variation of market return engenders a variation in the One can consequently say that even if the significance is
return of the shares for about 0, 2%. This value very not verified (we can even see that P >|z| is just slightly
weak as the parameter is not significant. But if it should higher than 10%), it appears that the BRVM stock market
be left within the meaning of the relation, it will be said takes into account liquidity when it is measured by
that it is in agreement with literature like the work of turnover. These conclusion is in line with Datar et al.
Sharpe (1964) and Saud et al. (2013) on the Iranian (1998) who found that expected return is significantly and
stock exchange market. negatively related to the liquidity measured by turnover
The coefficient ( 2 ), which takes into account the size
on American stock market. For them, this result is robust
even in the absence of Fama and Frenchs (1993)
of the company belongs to one of the significant variables.
parameters of Model 1. In fact, the value (absolute) of Batten and Xuan (2010) also find the same type of
this parameter is equal to approximately 11, 1% and is relation between liquidity and return on a set of emerging
significant at 1% (P > |z| = 0,003 < 1%). This means that stock exchange markets. Soosung and Lu (2005) found a
the risk related to the size of the company is contrary result on the British market, although the
recompensed on the BRVM. This result simply shows significance is not verified.
that the companies which have a market value tend to be These results do not change with the various tests of
the least profitable. Fama and French (1992, 1993) results robustness, except liquidity which becomes
found the same result on the American market, just as significant at 10% level when the January effect is taken
Miralles and al. (2011) on the Portuguese market and into account, an indication that BRVM is affected when
Saud et al. (2013) who showed that this relation is January information excluded. Thus one can say that
significantly negative (at 1%). On the other hand, Molay even if its significance is long in coming, the turnover is
 Kenfack et al. 137

Table 4. Turnover as liquidity risk measures.

Results obtained
Panels Parameters Sign waited
Coefficient (%) P< |z| (%)
0 None 6.25* 0

1 + -0.1 88.3

3X2 TO-beta portfolios 2 - -7.46*** 7.8

3 + 6.88 22.1

4 - 0.14* 0

0 None 11.11* O.2

1 + -9.71** 2.9

06 Beta portfolios (TO) 2 - 1.36 61.5

3 + -1.17 56.2

4 - 0.21* 0.3

0 None 6.92* 0

Without Fama and French (1992) factors (TO) 1 + -2.81* 0.8

4 - 0.23* 0

0 None 6.98* 0

1 + -0.79 10.5

January effect (TO) 2 - 7.77*** 8.6

3 + 6.06 36.9

4 - 0.1* 0

by far the most significant factor (with more than 100%) in account lack of liquidity.
explanation of the expected return of the BRVMs equities However, in spite of some disturbances, when liquidity
portfolios. is measured by turnover, it is well taken into account in
the BRVMs stock portfolios price formation process
(more than 100% with TO and nearly 60% with Il). The
The illiquidity indicator interrogation which persists consists of knowing if the risk
Amihuds (2002) illiquidity indicator which is attached there is also priced.

By replacing turnover in Model 1 with illiquidity ratio, the

results provided in Table 5 are more precise with regards The illiquidity risk in the stock exchange price
to the role of liquidity in the stock price formation process. formation process
More specifically, the constant term, the market portfolio Just like in the preceding section, the study has the
return, the company size and the book to market value results for each indicator of the liquidity.
respectively, did not change sign. Only that their
significance varies from one panel to another. On the
other hand, the importance of liquidity becomes more The turnover as measures of the of illiquidity risk
visible in the sense that its effect is at 5% significant
when the two indicators of liquidity are simultaneously Table 4 shows the results when is used in Model 2 as a
used in the same model and 10% everywhere else. The measure of risk. The constant is always significant at
January effect is also significant when one takes into 1% level in whatever panel that is considered. The market
138 J. Econ. Int. Finance

Table 4. Turnover as liquidity risk measures.

Results obtained
Panels Parameters Sign waited
Coefficient (%) P< |z| (%)
0 None 6.25* 0

1 + -0.1 88.3

3X2 TO-beta portfolios 2 - -7.46*** 7.8

3 + 6.88 22.1

4 - 0.14* 0

0 None 11.11* O.2

1 + -9.71** 2.9

06 Beta portfolios (TO) 2 - 1.36 61.5

3 + -1.17 56.2

4 - 0.21* 0.3

0 None 6.92* 0

Without Fama and French (1992) factors (TO) 1 + -2.81* 0.8

4 - 0.23* 0

0 None 6.98* 0

1 + -0.79 10.5

January effect (TO) 2 - 7.77*** 8.6

3 + 6.06 36.9

4 - 0.1* 0

variable (beta) has a negative coefficient ( 1 ) of 1, 10% BRVMs listed companies leads to a reduction in its
return by 7, 46% per month, which corresponds to an
and is not significant. This simply implies that if it varies 3
annual remuneration of 1,371 per unit of risk.
by 1% the expected return of the portfolios of shares will Consequently, the risk related to the size of the
vary in the opposite direction by 1, 10. Thus the most
company is well rewarded by the threshold of 10% on the
profitable stocks are the least risky, which means that the
BRVM. These conclusions contradict the work of Molay
systematic risk (market return) is not priced on the BRVM
(2002) on the French market and of Lu et al. (2005)
within the Sharpe framework. This result is normal given
which show that large companies are more powerful than
the characteristics of the emerging stock exchange smaller ones.
markets where returns are notable to be strongly volatile
The risk related to the value of the company ( 3 ) is
and most listed companies are very weak
para-public and public companies. positive (6.88%) but not significant since its p-value is
The risk related to the company size ( 2 ) has a
higher than 10% threshold. This positive sign is the

negative value (7, 46%) and significant at 10% (P > |z| =

0,078 < 10%). An increase by 1% in the size of the 3
1,371= (1+0,0746)12-1
 Kenfack et al. 139

Table 5. Illiquidity ratio indicator as a risk.

Results obtained
Panels Parameters Sign waited
Coefficient (%) P< |z| (%)
0 None 6.25 10.5

1 + 1.19 70.5

3X2 il-beta portfolios 2 - -15.68** 3

3 + 26.55 10.5

4 - 0.03 88

0 None 10.56** 1.7

1 + -9.20*** 6.9

06 Beta portfolios (il) 2 - 4.36** 2.3

3 + -4.78*** 6.4

4 - 0.14 63.1

0 None 8.91** 3.2

Without Fama and French (1992) factors (il) 1 + -2.39** 4.8

4 - -0.027 83.6

0 None 6.15 10
1 + 1.94 49.6

January effect 2 - -16.07** 3.8

3 + 27.01 11.2

4 - 0.03 88.6

0 None 7.89** 2.9

1 + 0.20 94.3

2 - -13.26** 4.1
TO and il simultaneously
3 + 19.47 16.9

4 - -0.19 1.2
5 + 0.012 93.27

*Significant at 1%; **Significant at 5%; ***Significant at 1%.

expected sign since it supposes the risk premium results are consistent with the analysis of Fama and
associated with the companys value is consistent. The French (1992, 1993), of Chordia et al. (2001) on the
increase in the value of the company (either by reduction American market and that of Beert et al. (1998) on a
of the number of shares without there being as much group of 20 emerging stock exchange markets.Financial
reduction in the book value, or by raising of prices of the literature (Chordia et al., 2001 and Datar et al., 1998)
shares) is priced by an increase of 6,88% in a monthly suggest that for a liquidity risk (evaluated by turnover) to
expected return in excess of the risk-free rate. These be priced in a market, it would be necessary that the
140 J. Econ. Int. Finance

relation between it and the expected return should be Mendelson (1986) in America. On the emerging markets,
negative. The results in Table 4 rather show the opposite. Dalgaard (2009) found a monthly premium of 5.9% in
In fact, the value of the premium ( 4 ) is positive (0, 14%) Denmark. This shows the weak liquidity of the African
financial markets. Thus, in spite of the multiple
and significant at 1%. This is a proof that increase of a
tendencies which seek to contradict the basic results in
shares liquidity risk (of 1%) will rather lead to a reduction particular when turnover is used as indicator of liquidity,
in the expected return approximately by 0,14%. Thus, we do not have enough proof for saying that liquidity risk
the relation between liquidity measured by turnover (TO)
is priced on the BRVM. Only the risk related to the size
and the yield are positive: risk of liquidity is not
seems to be.
recompensed on the BRVM. The investors who prefer
to hold long-term shares do not receive remuneration on
the additional risk which they support; on the contrary, CONCLUSION
they lose their surpluses of profitability. This contradicts
the "buy and hold" strategy recommended to investors on
In conclusion, the use of portfolio panels in the estimation
the emerging stock exchange markets. This result is
method enabled us to conclude that although liquidity is
also contrary to the work of Datar et al. (1998) and that of
one of the most significant factors (or most significant
Chordia et al. (2001) on the American market. But is in
according to our results), the risk associated is not priced
agreement with those of Nahandi et al. (2012). Miralles et on the BRVM. That means that apart from risk and
al. (2011) also find that the most liquid shares in Portugal profitability, liquidity must be given due consideration in
are most profitable. Illiquidity risk is thus not priced with
the stock markets investment decisions. A possible
turnover even when Fama and Frenchs (1992) factors
explanation is directly linked to the principal characteristic
and January information are excluded, or even when the
of the African stock exchange markets with propensity for
two liquidity indicators are simultaneously used. inefficiency. Another conclusion from our comparative
analysis is that each financial market, except those of
American market, has its own specificities with regards to
The illiquidity ratio as a risk measure
the concept of liquidity. It then becomes imperative to set
up policies that will support efficiency of the emerging
From Table 5, risk premium which is related to the market
stock exchange markets or those that will reduce its
portfolio ( 1 ) has become positive such that at a inefficiency.
non-significant level (P > |z| = 0,705 > 10%), an increase We document a strong instability of turnover as
in the market risk is priced by an increase in the indicator of liquidity. This casts a doubt on its explanatory
portfolios return considered by about 1.19% per month, power. A new research track would be to seek to know
4
for an annual premium of 15.25% . This is an indication which indicators of liquidity could be adapted as suitable
that illiquidity is considered better on the BRVM. This to emerging financial markets overall and Africa in
may be considered normal since African markets are particular; or, if it is possible, to create an indicator that is
famous for being largely less liquid. The systematic risk specific to the African continent.
related to the company size is better remunerated on the In addition, of all the traversed literature, none treats of
BRVM in that the value almost doubled compared to the the applicability of Archaya and Pedersens (2005) model
case where the TO is used (15, 68% against 7, 46%). of liquidity on African stock exchange markets. This new
The pricing ( 2 ) of "risk-size" is also significant at 5% question could also be subject to a new study, aimed at
checking if the African markets have specificities
against 10% previously. This once more shows the compared to this theory.
effectiveness of the ratio of lack of liquidity compared with Finally, we note that when the portfolios are formed on
turnover on the BRVM. As regards the risk premium the basis of individual share betas, the results become
attached to the company value ( 3 ), it is significant and very unstable.
its value doubled considerably with the same conclusions.
With regards to liquidity risk, certainly lack of liquidity is
slightly an increasing function of expected return in Conflict of Interests
excess of the risk free rate (0.031%) but not significant
contrary to expectations. A reduction of liquidity by 1% The authors have not declared any conflict of interests.
leads to an increase in expected return of about 0.031%
per month or 0.37% per annum. This result in sharp
contrast to the value of liquidity premium (4.6%) found by REFERENCES
Acharya and Pedersen (2005) and 1% by Amihud and
Acharya V, Pedersen L (2005). Asset pricing with liquidity risk. J.
Financ. Econ. pp. 375-410.
Amihud A, Mendelson H (1991). Assets prices and financial policy.
4
15,25= (1+0,0119)12-1 Financ. Anal. J. 47(6):56-66.
 Kenfack et al. 141

Amihud Y, Mendelson H (1986). Asset pricing and the bid-ask spread. Minovic J, Zivkovic B (2012). The impact of liquidity and size prenium
J. Financ. Econ. pp. 223-249. on equity price formation in Serbia. Economic Anals, 57: 195.
Amihud Y (2002). Illiquidity and stock return: cross-section and Miralles JL, Miralles M, Olivera C (2011). The role of an illiquidity factor
time-series effects. J. Financ. Mark. 5:31-56. in the Portuguese Stock Market. XII Iberian-Italian Congress Of
Amihud Y, Mendelson H.(1991). Liquidity, asset prices and financial Financial and Actuarial Mathematics (Lisbon).
policy. Financ. Anal. J. 47(6):56-66. Molay E (2002). Une analyse transversale de la rentabilit espre des
Bekaert G, Campbell R, Christian L (2007). Liquidity and expected actions sur le march franais. WP CEROG, EDHEC.
returns: lessons from emerging markets. Rev. Financ. Stud. Mooradian MR (2010). Illiquidity and stock returns. Rev. Appl. Econ.
20:1783-1831. 6(1-2):41-59.
Brennan M, Tarum C, Subrahmanyam A (1998). Alternative factors Mourina AB, Habib H (2012). Asset pricing and liquidity risk
specification, security characteristics, and cross-section of expected interrelation: an empirical investigation of Tunisian Stock Market. J.
stock returns. J. Financ. Eco. 49:345-373. Res. Int. Bus. Manage. 2(12):312-322.
Brennan MJ, Tarum C, Avanidhar S, Qing T (2011). sell-order liquidity Nahandi YB, Zeynali M, Maleki A (2012). Survey the influence of stock
and the cross-section of expected stock returns. working paper liquidity on the stock return of corporations listed at the Tehran Stock
(SSRN). Exchange. J. Basic Appl. Sci. Res. 2(5):4956-4960.
Chan HW, Faff RW (2005). Asset pricing and the illiquidity primium. Rahim AR, Shaari AM (2006). a Comparison Between Fama and
Financ. Rev. 40(4):429-458. French model and liquidity-based three-factor models in predicting
Chordia T, Sahn-Wook HS (2005). The cross-section of expected the porfolio returns. Asian Acad. Manage. J. Account. Financ.
trading activity. University of Atlanta (USA), Brock University 2(2):43-60.
(Canada), University of California (USA). Rune D (2009). Liquidity and stock returns: evidence of Denmark.
Chordia T, Subrahmanyam QT (2013). Trend in the cross-section of Copenhagen Business School, Master thesis.
expected stock return. SSRN, Anderson Business School, UCLA, Sharpe WF (1964). Capital asset prices: a theory of market equilibrum
Singapore Management, University. under conditions of risk. J. Financ. 19:425-442.
Chordia T, Roll R, subrahmanyam (2001). Market liquidity and traiding Sing-yang H (1997). Trading turnover and expected stock returns: the
activity. J. Financ. 56:501-530. trading frequency hypothesis and evidence from the Tokyo Stock
Datar V, Naik Y, Radcliffe (1998). Liquidity and stock return: An Exchange. National Taiwan University And University Of Chicago,
alternative test. J. Financ. Mark. 1:205-219. Working Paper, P 28.
Fama EF, French KR (1993). Common risk factors in the returns of Soosung H, Chensheng L (2005). Cross-sectional Stock returns in the
stocks and bonds. J. Financ. Econ. 33:3-56. UK market: the role of liquidity risk. Cass Business School (City of
Fama E, Macbeth J (1973). Risk, return, and equilibrum: empirical tests. London).
J. Polit. Econ. 81:607-636. Van V, Chai D , Viet D (2012). Empirical test on the Liquidity-Adjusted
Hearn B (2010). Liquidity and valuation in east african securities Capital Assets Princing Model: Australian evidence. WP, Monash
markets. Social Science Resarch Network (King's College London). University.
Hearn B, Piesse J (2009). Pricing southern shares in the presence of Xin-Yuan X, Liu-liu K (2012). An empirical study on the relationship
illiquidity: a capital asset pricing model augmented by size and between tunover rate and stock returns in Chinise Stock Market. Adv.
liquidity premiums. Social Science Reseach Network (King's College Info. Technol. Manage. 2(1):239-245.
London). Xuan VV, Batten J (2010). An empirical investigation of liquidity and
Joher HA (2009). Trading activities and cross-sectionnal variation in stock markets during financial crisis. Munich Personal Repec Archive,
stock expected return: evidence Of Kuala Lumpur Stock Exchange. No. 29862.
Int. Bus. Res. 2:2.
Liu W (2006). A Liquidity Augmented Capital Asset Pricing Model. J.
Financ. Eco. 82:631-671.