Académique Documents
Professionnel Documents
Culture Documents
1. INTRODUCTION
Returns and volume are two important elements of every kind of economic equilibrium and these elements are
jointly determined by the same market dynamics and may contain valuable information about securities.
Therefore, it is generally believed that these two variables should have very close and straightforward
relationship. Volume analysis is an important tool in the archery of technical analysts. It tries to forecast future
price move on the basis of past volume. Two famous stock market adages say: It takes volume to make prices
move, and Volume is heavy in bull market and light in bear market. These adages imply two different kinds of
causalities between these two variables. However, conceptual and empirical findings on the relationship between
these two variables are not so straightforward. Causaliy is a kind of statistical feedback concept which is widely
used in the building of forecasting models. Causality tests can provide useful information on whether knowledge
of past stock returns improves short run forecasts of current and future trading volume.There are several evidence
for the presence of a causal relation between stock returns and trading volume in the literatures. First, Epps
(1975) gave evidence based on the asymmetric reaction of two groups of investors-bulls and bears-to the positive
information and negative information. Secondly, Clark (1973) developed mixture of distribution model and
provided evidence that the speed of information flow is a latent common factor which influence stock returns and
trading volume simultaneously. Third evidence is the sequential arrival of information model of Copeland (1976)
which indicated a positive relationship between price changes and trading volume.
This objective of this study is to empirically examine the causal relation between volume and share price for
Nepalese Stock Market using Garnger causality procedure. The study investigate the Granger causality between
stock returns and trading volumes to ascertain if the causality runs from volume to stock returns or from stock
returns to volume or in both directions. The study uses monthly time series data sets. The rest of the paper is
organized as follows: the next section discusses the conceptual relationship between volume and price and
presents a brief survey of empirical research on this issue. The third section highlights the methodology of the
present study. This is followed by discussions on the results of the study in the fourth section. Section 5 concludes
the paper.
2.
REVIEW OF LITERATURE
Granger (1969) developed a methodology to examine whether changes in one series cause changes in another. If
the current value of Y can be predicted by using the past values of X and considering other relevant information
including the past values of Y, it may be concluded that X causes Y. To determine the presence of and the
direction of a causal relation between return and volume, the Granger causality method has been chosen.
Granger causality method was utilized by several researchers to investigate the causal relation between share
price and trading volume. Some examples of such studies are as follows. Most of studies found bidirectional
causality between trading volume and stock returns volatility. Some of the studies found unidirectional causality
from volume to volatility and some from volatility to volume.
Author
Jain & Joh (1988)
Assets
S&P 500 stock index
Period
1/1979-12/1983
DIJA index
1915-1940;1941-
Causality
unidirectional; Returns
Volume
1990
Returns
Returns
unidirectional; Returns
Bidirectional;
Volume
Gwilym,
McMillan
and
1/1995-6/1995
1973-2000
Ciner (2002)
1/1992-9/2000
1/1990-6/1999
Speight (1999)
Bidirectional;
Volume
Volume
Bidirectional;
Volume
INTRA: Bidirectional;
Volume
RETURNS:
Returns
Sarwar (2003)
1-1993-10/1995
Bidirectional;
Volume
Returns
Returns
SQUARED
unidirectional;
Volume
Returns
Author
Darrat, Rahman and Zhong
Assets
DJIA Stocks
Period
4/1998-6/1998
Chen (2007)
1/1995-12/1995
HSBC stock
01/2009-05/2009
Hussain (2011)
DAX 30 Index
5/2004-9/2005
(2003)
Causality
Bidirectional;
Returns
Returns
Returns
Volume
Volume
Bidirectional;
Volume
Bidirectional;
Volume
unidirectional;
Returns
but
investigate
do
NOT
returns
Return
Return
Volume
Attari,
Rafiq
and
Awan
2000-2012
Darwish (2012)
10/2000-8/2010
Halova (2012)
1/8/2008-
(2012)
Bidirectional;
Volume
Bidirectional;
Volume
Absolute
return:
Bidirectional;
Volume
24/9/2010
returns
Conditional
variance:
unidirectional; Returns
Abdullahi,
Kouhy
and
1/2008-5/2011
Muhammad (2014)
Samman
1/2009-12/2013
Volume
Neither trading
volume
nor
and
Al-Jafari
(2015)
unidirectional; Volume
Return
In nutshell, on the basis of above-mentioned studies it can be stated that the significant efforts have been made at
the international level to evaluate stock return and trading volume, whereas in Nepal, the causal relationship
between stock price index and market capitalizations with macro economic variables has been investigated by
G.C.and Neupane (2006) using the time series data for the year 1988 to 2005. The study found the empirical
evidence that the direction of causality from real GDP to stock price index but no reverse causation from stock
price index to real GDP. The study also concluded bidirectional causality between market capitalizations. The
causal relationship between stock returns and trading volume has not been investigated in Nepalese stock
market. Therefore, the current study is an attempt to fill this gap and sheds light on the informational efficiency of
Nepalese stock market. This paper examines the causal relationship between stock returns and trading volume
context in Nepalese stock market and contributes to the literature in several respects. It deploys theGranger
causality test to investigate information flow between the variables instead of ARIMA. Moreover, the time period
considered in the study helps to evaluate the impact of new regime changes in economic situation of Nepal. Thus,
the study will enhance the understanding of market asymmetry, market efficiency and information processing in
Nepalese stock market.
3.
This section describes the following aspects of research methodology to test the relationship between stock
returns and trading volume in Nepalese stock market: (i) nature and sources of data, (ii) selection of enterprises,
(iii) the variables, (iv) methods of analysis, and (v) the limitations of the study.
3.1 NATURE AND SOURCES OF DATA
The relationship between trading volume and stock returns are examined by using secondary sources of
trading volume and return data series. Most of the data related stock returns and trading volume were
collected from annual report and official reports of Security Exchange Board of Nepal (SEBON), official
website of Nepalese Stock Exchange (NEPSE). The data set used in this study comprises monthly closing
prices and trading volume in NEPSE. The study period covers the time period of eight years, ranging from
July 2007 to February 2015 and thereby making 92 months. The monthly data set are collected from
http://www.nepalstock.com/ which is available from July 2007 onward. Both series, (monthly trading volume
and monthly closing prices), are expressed in the local currency. The monthly stock returns are computed
using monthly closing prices.
3.2 SELECTION OF ENTERPRISES
The study consists of all listed companies of Nepalese stock market at index level. The study also considers
sector wise data of Nepalese stock market.
3.3 VARIABLES SPECIFICATION
The study mainly considers monthly volume series and return series to examine the relationship between
trading volume and stock returns.
Trading volume: The total number of shares traded were used to measure of trading volume for study of
aggregate trading activities e.g., Epps and Epps, (1976), Gallant, Rossi, and Tauchen, (1992), Hiemstra and
Jones (1994) and Ying (1966). Other studies used the total number of shares traded divided by the total
number of shares outstanding as a measure of trading volume, e.g., Campbell, Grossman, Wang (1993),
and LeBaron (1992). Individual share volume was often used in the analysis of price/volume and
volatility/volume relations, e.g., Andersen (1996), Epps and Epps (1976), and Lamoureux and Lastrapes
(1990, 1994). Alternatively, even the total number of trades were used by Conrad, Hameed, and Niden,
(1994). James and Edmister, (1983) used the number of trading days per year as measures of trading
activity.
The NEPSE provides three measures of trading volume: 1) the volume traded (the sum of the number of
share traded during the period); 2) the number of trades (the sum of trade during the period); 3) the total
value traded expressed in NPR (the sum of the price of share traded times the number of share traded). This
study uses the total value traded of the shares as the measure of trading volume because it takes into
account of the relative market value of shares while other measures mentioned above do not. The level
volume series is non-stationary, thus the study use first order difference volume series. The volume series at
the time t noted as
D ( V t ) is expressed as:
V
D( t)=First order differences (Volumet )
Where
Volumet is the NPR value of the shares traded at time t, and D (.) is the first order difference
operator.
Stock returns: The study considered monthly price index change as stock returns. A monthly price index
change is calculated using the natural log of the ratio of a stocks price index (P) from the current month (t) to
the previous month (t-1) as:
Where,
Pt
Pt
Pt 1
( )
represents the closing price index for the period t; t is the time in months.
Pt1 is the
closing price index for the period of t-1; Ln (.) is the natural logarithm operator. All returns are expressed
in local currencies and are not adjusted for dividends. The study also utilize the first order difference of
stock returns to maintain same order.
D(
t =1
t =1
Rt =R + j . Rt 1+ j . V t 1 + t
(1)
t=1
t=1
V t = V + j . V t1 + j . R t1 + t
Where,,
(2)
to lagged variables of
Rt V t .
Rt V t . Both equations
t , are uncorrelated. The null hypothesis for equation (1) and (2)
Rt V t and no causation from V t Rt respectively.
y t = 0+ y t 1+ i y t i + t
i =1
y t = 0+ y t 1+ 1 t + i y ti + t (4)
i=1
(3)
Johansen's multivariate maximum likelihood method using this co-integrated process to test those variables
have existed long-run equilibrium relationship (Johansen 1988). Co-integration test is to identify existence of
any co-integrating relationship between trading volume and stock returns in Nepalese stock market.
trace =T
i=r +1
T is sample size,
ln ( 1 i )
i
(5)
(6)
The study started investigation with same basic descriptive statistics of time series of stock returns and
trading volume from monthly data of NEPSE index for the period of 2007.07 to 2015.02. It would provide
insight to the average size and deviation from mean value of the variable. The descriptive statistics for the
variables are shown in annex 1. The analysis shows that the mean value of monthly stock returns over the
period is -0.0039 with standard deviation of 0.3313.The maximum and minimum returns observed are 1.3438
and 1.4224 respectively. Applying Jarque-Bera test for normality, Significant Jarque-Bera statistics clearly
rejects the hypothesis, which implies that pattern of monthly trading volume and monthly stock stock returns
series does not conform to normal distribution, which is the precondition for any market to be efficient in the
weak form (Fama (1965), Stevenson and Bear (1970), Reddy (1997), Kamath (2008) and Mahajan and
Singh (2008). Further, skewness and excess kurtosis preserve the evidence of the nature of departure from
normality because the skewness and kurtosis value of stock returns are -0.3949 and 13.8980 respectively.
The skewness coefficient in excess of unity is generally taken to be fairly extreme. Kurtosis generally either
much higher or lower indicates extreme leptokurtic or extreme platykurtic (Parkinson, 1980). Thus, monthly
stock returns and trading volume series are not normally distributed.
Similarly, the descriptive statistics of sector wise data set shows that the highest mean value of return series
is 2.316% with standard deviation of 7.723% was found in hotel sector among the sectors. The highest mean
value of trading volume is NPR 1654.896 with standard deviation of 3069.45 in commercial bank sector. The
monthly stock returns and trading volume series are not normally distributed in all sectors except return
series of commercial bank and other sector. It implies that pattern of almost all variables does not conform to
normal distribution.
4.2 UNIT ROOT TEST
The time series variables - stock returns and trading volume - should be stationary for the analysis of time
series statistics. The unit root test helps to test whether the times series variables are stationary or not.
The unit root test results are shown in annex 2. The statistics are inferior to the McKinnon critical value at
10% level. This indicates that the null hypothesis of unit root is rejected for the level form of stock returns and
first order difference of stock returns series. But, volume variable have unit root in level form and no unit root
in first order difference of trading volume series in ADF test.
The PP test also rejected the null hypothesis of unit root for stock return, first order differnces and first order
differenced trading volume variable as of ADF test. Philips-Perron test have also shown that level form of
volume variable have unit root in level forms. Hence, the study comes to conclusion that the times series
variables- trading volume and stock returns- of the study should be changed with first order difference to get
these variables are stationary. Having determined the non-stationary of time series in their levels, and they
are also of some order of integration I (1), co-integration test has been applied to ascertain whether trading
volume and stock returns are co-integrated or not.
4.3 JOHANSEN'S CO-INTEGRATION TEST
Johansen's co-integration is an econometric property of time series variables. If two or more series are
themselves non stationary, but a linear combination of time is stationary, then the series are said to be cointegrated. A vector of time series variable is co-integrated if each element of order one, I (1), but there exists
a nonzero vector (called co-integrating vector) such that is integrated of order zero, I (0). Economic theory
often suggests that certain subset of variable should be linked by a long-run equilibrium relationship. The
Johansen's co-integration tests statistics are shown in annex 3. The result of the Johansen's co-integration
tests shows that the null hypothesis of no co-integration between trading volume and stock returns is
rejected since both Trace and Max-Eigen statistics are larger than the critical values at 1 percent significance
level. In other words, for co-integrating regression, trading volume = f (stock returns), one can reject the null
hypothesis r = 0 against the alternative hypothesis r = 1 since both Trace and Max-Eigen statistics are larger
than the critical values at 1 percent significance level but cannot reject the null r1 against the alternative r = 2
since both Trace and Max-Eigen statistics are less than the critical values even at 1 percent level. The fact
that the presence of co-integration between trading volume and stock returns suggest (i) that there is a longrun equilibrium relationship between the two time series, and (ii) the existence of causality in at least one
direction.
The findings of the study is supported to the study of Abdullahi, Kouhy and
Muhammad (2014) in West Texas Intermediate(WTI) and Brent Crude Oil futures Markets. The study results is
contradictory with the study of Heimstra and Jones(1994) in Dow Jones stock returns and New York stock
exchange, Ciner (2002) in Tokyo commodity futures market and Mahajan and Singh (2009) in Indian stock
market and Meshkin, Gargaz and Abbasi (2014) in Tehran stock exchange that is indicative of noise trading
model of return-volume interaction in this market. The results presented in this study helps to increase our
understanding regarding the relationship between trading volume and stock returns, particularly for the
NEPSE. These findings also help to explain the behaviour of returns and a better understanding of market
movement. It also indicated that any study of trading volume and returns is necessarily relating to information
flow and possibly to identify a better proxy for information flow. From investment perspective, the relationship
between stock returns and trading volume is of great importance to individuals who invest and
trace a
perspective as trading volume reflects information about market expectations and its relationship with price,
having important implications on trading, speculation, forecasting and finally on hedging activities.
These results are useful for regulators, market participants, and the efficiency of the share market sector. The
results are useful for regulators when they consider such measures as limits on price movements and
positions. For market participants, the results are useful since they imply that volume can be used to predict
prices, lending support to technical analysis. Finally, the results imply market inefficiency that may be due to
the fact that traders may condition their prices on previous volumes.
The findings have a number of important implications for future research in this area. First, the results
contribute to the empirical literature on Nepalese stock market by highlighting detectable Granger linear
causalities. In addition, the findings of significant unidirectional Granger causality from trading volume to stock
returns illustrate the importance of these predictor variables in the forecasting of stock prices.
BIBLIOGRAPHY
Abdullahi, S., Kouhy, R., & Muhammad, Z. (2014). Trading volume and return relationship
in the crude oil futures markets. Studies in Economics and Finance, 31(4), 426-438.
Andersen, T. G. (1996). Return Volatility and Trading Volume: An Information Flow Interpretation of Stochastic
Volatility. Journal of Finance, 51(1), 169-204.
Attari, J. M., Rafiq, S. A., & Awan, H. M. (2012). The dynamic relationship between stock
volatility and trading volume. Asian Economic and Financial Review, 2(8), 10851097.
Campbell, J., Grossman, S., & Wang, J. (1993). Trading Volume and Serial Correlation in Stock Returns.
Quarterly Journal of Economics, 108, 905-939.
Chen, G., Firth, M., & Rui, O. (2001). The Dynamic Relation Between Stock Returns, Trading Volume and
Volatility. The Financial Review, 38, 153-174.
Ciner, C. (2002). Information content of volume: An investigation of Tokyo commodity futures markets. Pacific
Basin Financial Jounral, 10, 201-215.
Clark, P. (1973). A subordinated stochastic process model with finite variance for
speculative prices. Econometrica, 41, 135-155.
Conrad, J., Hameed, A., & & Niden, C. (1994). Volume and Auto covariance in Short Horizon Individual Security
Returns. The Journal of Finance, 49(4), 305-1329.
Dickey, D., & Fuller, W. (1981). Likelihood Ratio Statistics for Autoregressive Time series with Unit Root.
Econometrica, 49, 057-1072.
Epps, T. W. (1975). Security price changes and transaction volumes: Theory and Evidence.
American Economic Review, 65, 586-597.
Epps, T., & Epps, M. (1976). The stochastic dependence of security price changes and transaction volumes:
Implications for mixture of distribution hypothesis. Econometrica, 44, 305-321.
Fama, E. (1965). The Behavior of Stock-Market Prices. The Journal of Business, 38(1), 34-105.
G.C., S., & Neupane, S. (2006, December). Stock Market and Economic Development: A causality test. The
Journal of Nepalese Business Studies, 3(1), 36-44.
Gallant, A., Rossi, P., & Tauchen, G. (1992). Stock prices and volume. Review of Financial Studies, 5, 199-242.
Granger, C. (1969). Investigating causal relations by econometric models and cross-spectral methods.
Econometrica, 37, 424-438.
Gwilym, McMillan, & Speight. (1999). The Intraday Relationship Between Volume and
Volatility in LIFFE Future Markets. Applied Financial Economics, 9(6), 593-604.
Halova, M. W. (2012). The Intraday Volatility-Volume Relationship in Oil and Gas Futures.
Manuscript. Retrieved from http://dx.doi.org/10.1080/096031099332041
Hatrick, K., So, M. K., Chung, S. W., & Deng, R. (2011). Dynamic Relationship among
Intraday Realized Volatility, Volume and Number of Trades. Asia-Pacific Financial
Markets, 18(3), 291-317.
Hiemstra, C., & Jones, J. (1994). Testing for linear and nonlinear Granger causality in the stock price-volume
relation. Journal of Finance, 49, 1693-1664.
Hussain, S. (2011). The Intraday Behaviour of Bid-Ask Spreads, Trading Volume and Return
Volatility: Evidence from DAX30. International Journal of Economics and Finance,
3(1), 23-34.
Jain, P., & Joh, G. (1988). The dependence between hourly prices and trading volume. Journal of Financial and
Quantitative Analysis, 23, 269-283.
James, C., & Edmister, R. (1983). The Relation between Common Stock Returns Trading Activities and Market
Value. Journal of Finance, 38, 10751086.
Johansen, S. (1988). Statistical Analysis of Co-integration vectors. Journal of Economic Dynamics and Control,
12, 231-254.
Kamath, R. (2008). Return Distribution and the Day-of-the-Week Effects in the Stock Exchange of Thailand.
Journal of Economics and Finance, 22(2-3), 97-106.
Karpoff, J. (1987). The relation between price changes and trading volume: A survey. Journal of Financial and
Quantitative Analysis, 22, 109-126.
Lamoureux, C., & Lastrapes, W. (1994). Endogenous trading volume and momentum in stock return volatility.
Journal of Business and Economic Statistics, 12(2), 253-260.
Lamourex, C., & Lastrapes, W. (1990). Heteroskedasticity in Stock Return Data:Volume Versus GARCH Effects.
The Journal of Finance, 45(1), 21-29.
LeBaron, B. (1992). Some Relations between Volatility and Serial Correlation in Stock Market Returns. Journal of
Business, 65, 199-219.
Luu, J., & Martens, M. (2003). Testing the Mixture-of-Distributions Hypothesis Using
"Realized" Volatility. 23, pp. 661-679.
Mahajan, S., & Singh, B. (2008). An Empirical Analysis of Stock Price-Volume Relationship in Indian Stock
Market. Vision, 12(3), 1-13.
Mahajan, S., & Singh, B. (2009). The empirical investigation of relationship between return, volume, and volatility
dynamics in Indian stock market. Eurasian Journal of Business and Economics, 2, 113-137.
Meshkin, S., Gargaz, M., & Abbasi, E. (2014). Investigation of the Relationship between Trading Volume, Volatility
and Stock Return in Tehran Stock Exchange. Journal of Applied Science and Agriculture, 9(9), 145-153.
Parkinson, M. (1980). The extreme value method for estimating the variance of the rate of return. Journal of
Business, 53, 61-65.
Reddy, S. (1997). Efficiency of Indian Stock Markets: An Empirical Analysis of Weak-Form EMH of the BSE. UTI
Indian Capital Market Conference, 91-115.
Samman, H., & Al-Jafari, M. (2015). Trading Volume and Stock Returns Volatility: Evidence
from Industrial Firms of Oman. Asian Social Science, 11(24), 139-146.
Sarwar, G. (2003). The Interrelation of Price Volatiltiy and Trading of Currency Options. The
Journal of Futures Markets, 23, 681-700.
Stevenson, A., & Bear, M. (1970). Commodity Futures: Trends or Random Walks? The Journal of Finance, 25(1),
65-81.
Ying, C. (1966). Stock market prices and volumes of sales. Econometrica, 34, 676-685.
ANNEX 1
The data for this table are collected and compiled from the listed at Nepal Stock Exchange Limited for the period
2064.01 to 2071.11(July 2007 to February 2015) with 92 months .The table contained monthly stock returns and trading
volume at index level. The monthly trading volume is calculated by multiplying number of shares traded times closing
Vt
share price. The monthly trading volume has been used for volume (
by natural log of current closing stock index divided by lagged closing stock index, i.e. monthly stock returns
( t )=ln
Pt
Pt1
( )
. In the table, N represents number of observations and mean, median, maximum and minimum
for average, maximum and minimum returns respectively for the period. SD stands for standard deviation of the returns.
SK is a measure of skewness and KURT represents kurtosis. The Jarque-Bera (JB) test of normality is the test of the
joint hypothesis that SK and KURT are 0 and 3, respectively. In that case, the value of the JB statistic is expected to be
zero. JB is defined as JB=
S 2 ( k 3 )
n
+
6
24
any p-value less than 0.05 indicate that the distribution is not normal.
Panel I
Va
r
Std.
Dev.
Mean
Med
Max
Min
Rt
0.003
88
0.017
10
1.3438
0
1.4223
0
0.331
29
Skewne
ss
0.3947
3
Vt
2407.
46
1511.
62
15211.
61
386.11
2602.
15
2.41
Kurtosi
s
JarqueBera
Pvalu
e
13.899
73
457.804
60
0.00
0
9.98736
9.72
262.57
0.00
0
616000000
.00
Sum Sq.
Dev.
Panel II
Sect
or
CB
DB
FIN
HTL
va
r
Mean
Med
Max
Min
0.2155
0
Rt
0.000
84
0.0047
5
0.303
20
Vt
1502.
88
7883.
12
Rt
0.004
69
1023.5
1
0.0090
5
Vt
279.8
9
Rt
0.001
51
136.36
0.0087
5
Vt
171.5
9
90.12
2163.
41
Rt
0.022
66
0.0050
0
0.387
00
15.16
0.0758
0
31.75
3.19
393.3
1
0.00
Vt
0.462
90
1654.
97
0.211
40
117.55
0.4070
0
33.35
0.1628
0
Std.
Dev.
Skewn
ess
Kurtos
is
JarqueBera
Sum Sq.
Dev.
0.099
02
0.3269
9
3.7136
4
3.59175
0.89232
1532.
27
2.07
7.20
133.64***
214000000
.00
0.110
21
0.8215
8
7.7245
4
95.91467***
1.10525
334.0
6
2.21
7.96
169.06***
10155431.
00
0.063
93
0.6944
5
4.8961
5
21.17695***
0.37191
264.6
0
5.11
36.69
4750.57***
6371041.0
0
0.064
38
2.5326
5
13.241
77
500.44550*
**
0.37720
69.05
3.08
13.28
551.14***
433914.80
INS
MFG
OTR
Rt
0.020
95
0.0069
5
0.260
40
Vt
240.8
9
35.00
2441.
45
Rt
0.015
46
0.0011
5
0.338
60
29.75
Rt 0.000
20
1.91
788.4
6
0.0000
0
0.206
50
39.25
24.02
377.3
3
0.002
57
0.0000
0
2.80
1.36
Vt
Vt
Rt
TRD
Vt
0.1677
0
0.080
87
0.9496
5
4.1203
5
18.63975***
0.59519
446.6
8
2.66
10.73
337.23***
18156668.
00
0.063
18
0.5802
1
16.990
71
755.49840*
**
0.36330
118.6
2
5.42
32.55
3798.95***
1280413.0
0
0.061
66
0.2935
4
4.2568
9
7.37706**
0.34602
55.43
3.68
19.91
1303.69***
279549.40
0.439
70
0.00
0.1636
0
0.065
60
3.1466
1
23.606
75
1779.59700
***
0.39159
23.43
0.00
4.28
2.96
13.35
545.32***
1668.12
2.75
0.3083
0
0.00
0.1526
0
Annex 2
This table reports the results of the ADF test and PP test for unit roots. The lag length (p) is chosen based on AIC. The empirical model;
m
Yt Yt 1 j Yt i t t
where
Yt
i 1
0
different possibilities of the unit root process. If
and
, it corresponds to a random walk model and if only
modeling a random walk with drift.
Panel I
Rt
Vt
-2.5804
-3.179*
-1.7389*
-9.8094***
-9.7386***
-9.8621***
-2.2881
-2.9563
-1.4936
11.2398**
*
11.3053**
*
11.1678**
*
Panel II
Commercial Bank
Unit Root Test
Rt
Vt
Development Bank
Rt
Vt
Finance Company
Rt
Vt
, it equals to
4.6996***
5.1579***
3.1674***
6.2629***
6.2766***
6.2896***
Constant
7.7089***
8.1264***
8.8303***
7.6558***
8.0752***
8.6407***
-7.759***
8.1734***
8.9488***
4.5576***
4.8755***
3.0505***
none
PP TEST@ level
8.6254***
Constant & Trend
8.7638***
none
8.6731***
PP TEST@ first difference
37.0699**
Constant
*
37.4395**
Constant & Trend
*
37.3937**
none
*
Constant
31.237***
30.5326**
*
30.002***
3.6657***
4.1588***
3.6031***
5.6771***
-2.5307**
-3.624***
-5.0653***
10.6678**
*
10.6176**
*
10.7118**
*
15.4786**
*
15.4022**
*
15.5662**
*
12.9489**
*
12.8771**
*
13.0221**
*
6.2921***
6.2815***
6.3174***
3.5894***
5.6451***
5.6382***
5.6783***
26.6065**
*
15.5525**
*
16.4699**
*
35.002***
26.8346**
*
-4.136***
-2.1523**
14.812***
22.9744**
*
23.5244**
*
23.1031**
*
-6.4312***
-6.5553***
-6.4331***
-6.5555***
-4.9824***
-31.388***
-31.056***
31.7826**
*
Insurance Company
Unit Root Test
ADF TEST@ level form
Rt
Constant
3.6621***
7.9641***
none
3.4655***
ADF TEST@ first difference
Constant
9.3808***
9.3278***
none
PP TEST@ level
9.4359***
Vt
Rt
Vt
-1.3837
3.8738***
-2.2635
5.1758***
4.6324***
-0.946
4.4864***
3.4386***
11.6588**
*
11.7476**
*
10.9099**
*
16.1597**
*
16.0904**
*
16.2139**
*
10.9229**
*
10.8607**
*
10.9855**
*
12.1567**
*
12.0853**
*
12.2223**
*
10.6822**
*
10.6183**
*
10.7433**
*
3.8268***
Constant
-7.549***
-2.1398
8.0167***
-3.5458**
-8.799***
4.6701***
-1.8109*
-8.134***
3.2923***
15.9528**
*
15.8888**
*
16.0002**
*
39.686***
39.2256**
*
40.0848**
*
17.6865**
*
17.6672**
*
17.4442**
*
Other Company
Unit Root Test
Rt
Rt
4.9837***
8.6268***
none
7.2569***
PP TEST@ first difference
66.9211**
Constant
*
66.2177**
Constant & Trend
*
69.1326**
none
*
Manufacturing
Company
Hotel Company
Vt
Trading Company
Rt
Vt
7.1922***
9.5878***
-5.9633***
7.4388***
9.6112***
-6.0807***
5.3555***
9.6311***
-4.6978***
13.4855**
*
16.1223**
*
11.0163**
*
60.4837**
*
59.9818**
*
62.683***
Vt
-2.5726
-2.546
-2.184**
-11.409***
-11.349***
11.4781**
*
-9.7495***
-9.6956***
-9.2515***
50.9596**
*
51.4925**
*
51.3447**
*
Constant
Constant & Trend
none
PP TEST@ level
12.7127**
*
12.6393**
*
12.7862**
*
10.6988**
Constant
*
10.7943**
Constant & Trend
*
10.7564**
none
*
PP TEST@ first difference
56.3257**
Constant
*
55.9354**
Constant & Trend
*
56.6658**
none
*
Notes: (*) Significant at the 10%;
13.8526**
*
13.7788**
*
13.9307**
*
10.3476**
*
10.2883**
*
10.4095**
*
7.1972***
9.7506***
-5.9468***
-7.35***
9.9165***
-6.0119***
5.2766***
9.7965***
-4.6437***
47.1165** 41.7914**
*
*
53.1938** 42.9035**
*
*
46.0511** 42.2278**
*
*
(**) Significant at the 5%
-9.5518***
-9.5009***
-9.6086***
26.9329**
*
28.2698**
*
27.3456**
*
and (***) Significant at the 1%.
test rejects
H0
trace =T
i=r +1
ln ( 1 i )
that means variables exist at least r+1 long term co-integrated relationship. and (b). Maximum eigenvalue test; The
max (r ,r + 1)=T (1 r+ 1)
sample firms for the period of 2064.4 to 2071.11(July 2007 to February 2015). P-values are in the parentheses
Panel I series first order difference of stock returns and trading volume due to non-stationary of vol at level form.
H
0
r=
0
r
1
H
1
r
1
r
2
Eigenval
ue
0.74727
2
0.37163
6
Trace
value
0.74727
2
0.37163
6
PValue
0.000
1
0.000
0
H
0
r=
0
r
1
H
1
r=
1
r=
2
Max-Eigen
statistic
122.4143
41.35255
PValue
0.000
1
0.000
0
Panel II Panel I series level form of stock returns and trading volume
Sector
CB
H
0
r=
0
H
1
r
1
Eigenval
ue
0.29085
2
Trace
value
50.4654
3
(0.0000)
H
0
r=
0
H
1
r=
1
Max-Eigen
statistic
30.93221
(0.0001)
Decisi
on
2 CEs
Decisi
on
2 CEs
DB
FIN
HTL
INS
MFG
OTR
TRD
r
1
r
2
0.19509
9
r=
0
r
1
0.25403
8
r
1
r
2
0.11055
r=
0
r
1
0.25083
5
r
1
r
2
0.12184
4
r=
0
r
1
0.24866
5
r
1
r
2
0.10725
4
r=
0
r
1
0.32138
9
r
1
r
2
0.02841
6
r=
0
r
1
0.46273
6
r
1
r
2
0.34837
2
r=
0
r
1
0.35619
4
r
1
r
2
0.26503
4
r=
0
r
1
0.43377
9
r
1
r
2
0.24903
7
19.5332
2
(0.0000)
36.9208
6
(0.0000)
10.5436
3
(0.0012)
37.6854
6
(0.0000)
11.6937
7
(0.0006)
35.9420
8
(0.0000)
10.2107
7
(0.0014)
37.4880
9
(0.0000)
2.59448
3
(0.1072)
94.4592
3
(0.0000)
38.5453
9
(0.0000)
67.3460
5
(0.0000)
27.7138
(0.0000)
76.9652
6
(0.0001)
25.7759
4
(0.0000)
r
1
r=
2
19.53322
(0.0000)
r=
0
r=
1
26.37722
(0.0004)
r
1
r=
2
10.54363
(0.0012)
r=
0
r=
1
25.99169
(0.0005)
r
1
r=
2
11.69377
(0.0006)
r=
0
r=
1
25.73131
(0.0005)
r
1
r=
2
10.21077
(0.0014)
r=
0
r=
1
34.8936
(0.0000)
r
1
r=
2
2.594483
(0.1072)
r=
0
r=
1
55.91383
(0.0000)
r
1
r=
2
38.54539
(0.0000)
r=
0
r=
1
39.63225
(0.0000)
r
1
r=
2
27.7138
(0.0000)
r=
0
r=
1
51.18933
(0.0001)
r
1
r=
2
25.77594
(0.0000)
2 CEs
2 CEs
2 CEs
1 CEs
2 CEs
2 CEs
2 CEs
Rt
The variable volume (Vt) is the log difference of trading volume in month t, Stock returns,
, is the monthly stock returns. F-statistics of the
Granger-causality test are shown for the period of 2064.4 to 2071.11(July 2007 to February 2015) P-values are in the parentheses
Panel I
F-Statistic
Null Hypothesis
(P-Value)
Note: *** 1% or less, ** up to 5%, *
Panel II
V
D( t)
Sector
D(
Rt
DB
does notNull
Granger
Cause D(
Hypothesis
Vt
CB
Rt
Rt
does not Granger Cause
Rt
Vt
Vt
9.0686**
(0.0034)
0.0271
(0.8697)
0.3267
(0.5690)
RET
Rt
Vt
Vt
Rt
FIN
Rt
Vt
Vt
Rt
HTL
Rt
Vt
Vt
Rt
INS
Rt
Vt
MFG
Vt
3.6854*
(0.0581)
1.8481
(0.1775)
0.0198
(0.8884)
1.769
(0.2616)
1.5632
(0.2145)
0.9941
(0.3215)
0.0760
(0.7835)
0.3501
(0.5555)
Sector
Null Hypothesis
Rt
Rt
does not Granger Cause
Vt
Vt
Rt
OTR
Rt
Vt
Vt
Rt
TRD
Rt
Vt
FStatistic
(PValue)
11.6627**
(0.0010)
0.4834
(0.4887)
0.9026
(0.3447)
0.08266
(0.7744)
0.4919
(0.4849)