Content uploaded by Laxmidhar Samal
Author content
All content in this area was uploaded by Laxmidhar Samal on Mar 07, 2023
Content may be subject to copyright.
Article
Does the Base Metal
Futures Market of
India Support the
MDH? An Empirical
Evidence
Laxmidhar Samal1,2 and Sudhansu Kumar Das3
Abstract
The study examines the mixture of distribution hypothesis (MDH) and the sequential
information arrival hypotheses (SIH) in the base metal futures market of India. We use
near-month futures daily trading data of price, volume and open interest for 7 years. It
is downloaded from the official website of Multi Commodity Exchange (MCX), India.
The study supports the MDH as it confirms the existence of a contemporaneous
correlation between the return and change in volume of all base metal futures traded
at MCX, India. The article exhibits no causality between the return and volume change
of metal futures which supports the MDH and contradicts the SIH. This indicates a
greater level of market efficiency. The study finds unidirectional causality between
the return and daily change of open interest and bidirectional causality between the
change in volume and change in open interest. This is found for all base metal futures
and this aspect is left for in-depth analysis by the futures studies.
Keywords
Base metal, futures, copper, lead, nickel
JEL Classification: G13, G14
Received 30 April 2022; revised 10 July 2022; accepted 10 July 2022
IIMS Journal of Management Science
14(1) 45–61, 2023
© The Author(s) 2022
DOI: 10.1177/0976030X221116200
journal.iimshillong.ac.in
Creative Commons Non Commercial CC BY-NC: This article is distributed under the
terms of the Creative Commons Attribution-NonCommercial 4.0 License (http://www.
creativecommons.org/licenses/by-nc/4.0/) which permits non-Commercial use, reproduction and
distribution of the work without further permission provided the original work is attributed.
1 Utkal University, Bhubaneswar, Odisha, India
2 P. G. Department of Commerce, Baba Bhairabananda Autonomous Mahavidyalaya, Chandikhole,
Jajpur, Odisha, India
3 Department of Commerce, Sadhu Goureswar College, Odisha, India
Corresponding Author:
Laxmidhar Samal, Utkal University, Bhubaneswar, Odisha 751004, India; P. G. Department of
Commerce, Baba Bhairabananda Autonomous Mahavidyalaya, Chandikhole, Jajpur, Odisha 755044,
India.
E-mail: laxmidharsamal.ckl@gmail.com
46 IIMS Journal of Management Science 14(1)
Introduction
In India, commodity trading was banned with the apprehension that it will lead to
speculation, unethical hoarding and profiteering which will push up the prices
(Biswal, 2008). In May 2003, through a notification, the Government of Indi,
revoked the prohibition on commodity futures trading (Biswal, 2008) and estab-
lished three national-level commodity exchanges. Multi Commodity Exchange
(MCX) is the leading commodity exchange which accounts for a major percent-
age of the total commodity trading in India. Since 2003, commodity futures
trading in India has been showing significant growth both in terms of volume and
value. In the financial year 2019–2020, the average daily turnover of commodity
futures in India touched `34,491 crores which is more than a 23% increase from
the previous year (Commodity Insight Year Book 2019–2020, MCX).
Manifold’s increase in the market turnover of commodity futures attracts the
researchers for in-depth analysis. The information content of the data has become
the main point of attraction for researchers, policymakers and investors. Change
in demand and supply position affects the price and with the change in price dif-
ferent traders reacts differently having different information content. Participation
of different investors with diverse national and global information brings price
equilibrium to the market. The trading volume contains important information
which is very useful for different market participants Srinivasan et al. (2016).
According to Biswas and Rajib (2011), ‘a change in demand induces a price
change and subsequently adjustment takes place in price and trading volume until
a new equilibrium is arrived at’. This process of adjustment depends upon the
information flow to the market. With the arrival of new information to the market
and attainment of equilibrium, there are two competing hypotheses, namely the
mixture of distribution hypothesis (MDH) and the sequential information arrival
hypothesis (SIH). The former states that the price change and volume have a joint
response to information (Harris, 1986; Touchen & Pitts, 1983) whereas the latter
states that information arrives in the market in a sequential random fashion
(Copeland, 1976; Jennings et al., 1981). Price volume relation is the main indica-
tor of the structure of that financial market (Karpoff, 1987). Thus by studying the
price volume dynamics it would possible to predict the future return with the help
of present trading volume. The prime aim of the article is to evaluate empirically
the price volume dynamics of the base metal futures market of India. The study
uses causality checks which will also explore the different aspects of technical
analysis.
Review of Literature
There has been extensive research concerning the price volume dynamics
of different financial assets. Several researchers attempted to identify the
contemporaneous relation between return and volume of different asset classes
traded at both equity and commodity markets. Researchers have explained it with
the help of information flow which not only affects the return but also the trading
volume. Gallant et al. (1992) emphasised the joint dynamics of return and trading
Samal and Das 47
volume. A change in trading volume indicates the change in investors’ perception
because of information flow which affects the price of the asset. Besides other
financial assets, the demand and supply model of commodity futures indicates the
existence of a strong association between price and trading volume.
According to Biswas and Rajib (2011), ‘a change in demand induces a price
change and subsequently adjustment takes place in price and trading volume until
a new equilibrium is arrived at’. This process of adjustment depends upon the
information flow to the market. With the arrival of new information to the market
and attainment, of equilibrium, there are two competing hypotheses, namely the
MDH and the SIH. The former states that the price change and volume have a joint
response to information (Harris, 1986; Touchen & Pitts, 1983) whereas the latter
states that information arrives in the market in a sequential random fashion
(Copeland, 1976; Jennings et al., 1981). The MDH proposes that when new infor-
mation arrives in the market both trading volume and price responds synchro-
nously and information attains equilibrium immediately without any intermediate
equilibrium.
According to Biswas and Rajib (2011), the SIH proposes that ‘the information
arrives at the market in a sequential random fashion and there are series of inter-
temporal equilibria before the final complete equilibrium is reached’. It states that
because of the sequential flow of information the past values of trading volume
are able to predict the future return and vice versa. Thus it implies the existence of
causality between the change in trading volume and return.
Several researchers use the Granger causality test for examining the dynamics
between the prices and trading volume of a financial asset. Cornell (1981) and
Grammatikos and Saunders (1986) examined the currency as well as commodity
futures and concluded that there exists a positive correlation between the volume
and return. By developing a noise trading model De long et al. (1990) found a
positive causal relationship between price and volume. Bessembinder and Seguin
(1993) studied both commodity and financial futures markets and supported that
unexpected positive volume shocks have a larger effect on volatility than negative
volume shocks. Wang (1994) found a positive correlation between trading volume
and absolute price change and concluded that information asymmetry increases
the association between the two. In their, study Hiemstra and Jones (1994) consid-
ered Dow Jones stock return and New York Stock Exchange trading volume and
found the existence of unidirectional linear causality from the stock return to
trading volume and bidirectional non-linear causality between these two. For
examining the existence of non-linear causality they considered Baek and Brock’s
(1992) approach. By studying crude oil futures Foster (1995) postulated that the
lagged volume of crude oil futures forecasts its absolute value of the return.
Fujihara and Mougoue (1997) considered petroleum futures and supported the
existence of bidirectional nonlinear causality between the price and volume.
By extending Hiemstra and Jones’s (1994) work Silvapulle and Choi (1999) exam-
ined linear and non-linear causality between the stock return and volume traded in the
Korean Stock Market and the study exhibits the existence of both bi-directional linear
and non-linear causality between the volume and stock return. Moosa and Silvapulle
(2000) studied the linear and non-linear causality between the return and trading
48 IIMS Journal of Management Science 14(1)
volume of crude oil futures and confirmed the existence of unidirectional linear cau-
sality from trading volume to return and bidirectional non-linear causality between the
price and volume. They used a bivariate vector autoregressive (VAR) framework for
testing linear causality and considered Baek and Brock’s (1992) approach for testing
non-linear causality. Further, they concluded that volume can be used for predicting
prices. Ciner (2002) studied the return and trading volume of gold, platinum and
rubber futures contracts traded at the Tokyo commodity exchange. The study exam-
ined the impact of the information content of volume on price change. He found no
linear causality from volume to price change but argued for the existence of bidirec-
tional non-linear causality between change in trading volume and return.
Mcmillan and Speight (2002) investigated the MDH and the SIH by using
London Financial Futures and Options Exchange futures data. They examined the
dynamic relationship between the trading volume and return by employing the
Granger causality test with the VAR framework. The study exhibits causality from
the return to trading volume and they argued that there is no apparent causality
from the return to volume. Chen et al. (2004) examined the return volume dynam-
ics of the China futures market by using correlation and Granger causality test and
found no contemporaneous correlation between return and trading volume but
supported the existence of bi-directional causality. Lokman and Abdulnasser
(2005) took the stock price and trading volume of different countries, namely the
Czech Republic, Russia, Turkey, Hungary and Poland used the linear Granger
causality test by following Toda and Yamamoto’s (1995) approach and found no
causality between stock price and volume of the Czech Republic. The study
reveals unidirectional causality from stock price to volume for Russia and Turkey
whereas bidirectional causality is observed for Hungary and Poland.
By considering electricity futures traded at the New York Mercantile Exchange
Hadsell (2006) explored the relationship between the trading volume and price and
found that the information was slow to be incorporated into the price in two markets
and the traders react asymmetrically to the arrival of new information in three
markets. Nevin et al. (2006) tested causality using the VAR model proposed by
Granger and the nonlinear causality suggested by Peguin-Feissolle and Terasvirta
(1999) between the stock price and trading volume of Turkey. The study exhibits
unidirectional Granger causality from volume to return. They found strong predic-
tive power of volume. Girard and Biswas (2007) studied 49 equity indices of devel-
oped and developing markets and their study supports the MDH. Puri and Philippatos
(2008) explored the volume return dynamics of interest rate and currency futures
and finds an asymmetric relationship between volume and return.
Biswas and Rajib (2011) considered gold, silver and crude the oil futures of the
MCX of India and their study supports the SIH for all three commodity futures.
Zwergel and Heiden (2012) studied the German stock market and supported the
existence of contemporaneous relation between the return and volume. Srinivasan
et al. (2016) studied the Indian stock futures market by taking 25 stock futures
contracts and found a positive relationship between return and volume of stock
futures. They opined that return volatility influences the trading volume. Darolles
et al. (2017) analysed the mixture of distribution hypotheses in the context of
Samal and Das 49
liquidity friction and they distinguished liquidity friction into two parts, namely
short-term and long-term liquidity frictions. They found the former impact the
stock return and affect the traded volume and the latter is responsible for the
dynamics of daily return.
By following Mcmillan and Speight (2002), Chen et al. (2004) and Biswas and
Rajib (2011) the present study is an attempt to answer the following question. Does
the MDH hold for the Indian base metal futures market? Does the market support
the SIH? In addition, to explore new insights the study also examines the relation-
ship between return and change in open interest and the relationship between the
change of volume and open interest. The next section of the study deals with the data
and methodology and the subsequent section discusses the results. The fifth section
interprets the results and the last section contains the conclusion of the study.
Data and Methodology
Data
The study is based on the secondary data which is downloaded from the official
website of MCX, India for 7 years ranging from 1st April 2013 to 31st March 2020.
During this period, the base metal market gained significant momentum and this
period is free from any abnormal market fluctuations. The data relating to price,
the volume traded and open interest of five base metal futures, namely aluminium,
copper, lead, nickel and zinc are downloaded. The study uses the daily log return
series of price, log of daily changes in volume and open interest.
Variable Specification
Price: Refers to the log return series of futures price of five base metal futures
traded at MCX, India.
Volume: Refers to the log of daily changes in the volume of five base metal
futures traded at MCX, India.
Open interest: Refers to the log of daily changes in open interest of five base
metal futures traded at MCX, India.
Descriptive Statistics
The basic statistical characteristics of log return series of price, log of daily
changes in volume and open interest of five metal futures are analysed by using
descriptive statistics such as Mean (X), Standard deviation (v), Skewness (S) and
Jarque–Bera statistics. The Jarque–Bera statistics which is employed to test the
normality of the series implies that:
JB
nk
sk
64
13
22
=
-
+-
__
ii
<F
(1)
50 IIMS Journal of Management Science 14(1)
Where ‘S’ and ‘K’ indicate the coefficient of skewness and kurtosis respectively.
‘n’ refers to the number of observations whereas ‘k’ refers to the number of
estimated co-efficient used to create the series.
Unit Root Test
The unit root properties of price, volume and open interest series of different base
metals are analysed by performing the Augmented Dickey–Fuller (1979) test as
well as the Phillips–Perron (1988) test. The stationarity level of different series of
five base metals traded at MCX, India has been checked with the ADF test by
fitting a regression equation. The random walk-based regression equation with a
drift can be specified as follows:
yy yu
tt tj t1
2|{iDD=+ ++
--
(2)
The hypothesis specified as H0: ^ = 0, H1: ^ < 0.
Granger Causality
Causality methodology is developed by Granger (1969) to study whether the
change in one series causes another or not. In a set of two series X and Y, if we are
able to predict the current values of X, by considering the past values of Y and past
values of X including the other relevant information then it is said that Y causes X
and vice versa. The above concept of causality is expressed as follows:
YYXU
titi itit0||
aa b=+
++
--
(3)
XXYU
titi itit0||
aa b=+
++
--
(4)
Wheret indicates time and ‘i’ = 1 to m.
According to Ciner (2002), nonlinear causality from volume to return arises
because of the volatility effect. He stated that when lagged volume captures per-
sistence of information flow it will lead to spurious causality. Mcmillan and
Speight (2002) studied the contemporaneous relationship between the volume and
return and the existence of causality is examined by using the Granger causality
test with the VAR framework. Lokman and Abdulnasser (2005) examined causal-
ity by using the Granger causality approach with the procedure developed by Toda
and Yamamoto (1995). By using OLS regression in the causality test Chen et al.
(2004) verified the Granger Causality between the volume and return. They also
investigated the contemporaneous relationship between the volume and return by
employing Pearson correlation and Spearman rank correlation.
We use both the Augmented Dickey–Fuller (1979) test as well as the
Phillips–Perron (1988) test for testing the unit root properties of the time series.
Samal and Das 51
By following Mcmillan and Speight (2002), Chen et al. (2004) and Biswas and
Rajib (2011), the Granger causality test is used to examine the causality between
the trading volume and return of base metal futures traded at MCX, India. The
following VAR framework of causality test is applied only when the data series
are stationary.
RRVU
titi jtjxt0||
aa b=+
++
--
(5)
VRVU
titi jtjyt0||
ba b=+
++
--
(6)
Where ‘i’ = 1 to m, lag length for the daily return; and ‘j’ = 1 to n, lag length for
the daily trading volume.
According to Gujarati (1995), the Granger causality test is sensitive to the
number of lags selected for the analysis therefore we have used Akaike informa-
tion criteria for selecting the lag length of the variable. In the above-specified
equations and R indicates return whereas V indicates volume change and both are
stationary variables. The null and alternative hypotheses of the above equation can
be specified as follows:
H0: β1 = β2 = … = 0.
And the alternative hypothesis can be specified as:
H1: βj ≠ 0 for at least one j.
The study will support the MDH if a significant contemporaneous correlation is
found between the return and trading volume of different base metal futures. On
the other hand, the study will support the SIH if there exists causality between the
return and trading volume. Moreover, the study also examines the causality
between the trading volume and open interest and between the return and open
interest of base metal futures for exploring any newer dimension. For examining
the causality between the return and open interest the following equations are
specified:
RROU
titi jtjxt0||
aa b=+
++
--
(7)
OROU
titi jtjyt0||ba b=+ ++
--
(8)
Where ‘i’ = 1 to m, lag length for the daily return; and ‘j’ = 1 to n, lag length for
the daily open interest. Similarly for examining the causality between the volume
and open interest the following equations are specified:
OOVU
titi jtjxt0||aa b=+ ++
--
(9)
52 IIMS Journal of Management Science 14(1)
VOVU
titi jtjyt0||ba b=+ ++
--
(10)
Where ‘i’ = 1 to m, lag length for the open interest; and ‘j’ = 1 to n, lag length for
the daily trading volume.
Empirical Findings
The study considers the daily trading data of all the five base metal futures traded
at MCX, India. In order to create a continuous data series volume and open
interest, data have been added for all outstanding contracts. The study considers
the return series of futures price of five base metal futures, that is, aluminium,
copper, lead, nickel and zinc. For this purpose, it is considered that:
Rt = Ln (Pt / Pt–1)
Where Pt is the daily closing price and Pt–1 is the closing price of the previous day.
Similarly, the log of daily changes in volume and open interest of five base metal
futures are used for analysis. For this purpose, it is considered that:
Vt = Ln (Vt / Vt–1)
Where Vt is the daily traded volume.
Table 1 indicates basic statistical characteristics of log return series of price,
log of daily changes in volume and open interest of five metal futures. The mean
daily returns of aluminium, lead and zinc futures are positive whereas it is nega-
tive for copper and nickel. The mean log of daily changes in volume and open
interest of five base metal futures are negative. Except for copper, the median
daily returns of all base metals are zero. The maximum daily return of aluminium,
copper, lead, nickel and zinc is 10.25%, 6.14%, 17.72%, 7.26% and 9.36%,
respectively. Nickel and aluminium returns register the highest and lowest daily
fluctuations respectively.
Unlike nickel, all four base metals return series are positively skewed. The
daily changes in volume and open interest series of five base metal futures are
found to be negatively skewed. The fluctuation in daily changes in the volume of
base metal futures is higher than the fluctuation in daily changes in open interest.
The fluctuation in daily changes of volume is highest in aluminium whereas it is
lowest in the case of lead futures. Similarly, the fluctuation in daily changes of
open interest is found to be highest in nickel futures and lowest in copper futures.
The kurtosis value is higher than 3 for all daily return series and daily changes in
volume and open interest series of all base metal futures, which indicates that all
series are leptokurtic, that is, fat tailed. As Jarque–Bera’s p-value is zero for all
base metal series hence the null hypothesis of the presence of normality is rejected.
We use the Augmented Dickey–Fuller test (Dickey & Fuller, 1979) and
the Phillips–Perron (1988) (Phillips & Perron, 1988) test for examining the
Table 1. Descriptive Statistics of Return Series.
Descriptive Stat.
Aluminium Copper
Futures Volume Open Interest Futures Volume Open Interest
Mean 0.02 –0.06 –0.21 –0.01 –0.09 –0.14
Median 0.00 0.00 0.00 –0.01 –0.56 –0.08
Maximum 10.25 414.46 64.40 6.14 438.89 45.68
Minimum –9.41 –458.23 –118.45 –6.67 –406.57 –69.00
Std Dev. 1.13 70.51 13.24 1.16 68.12 9.28
Skewness 0.66 –0.30 –1.24 0.02 –0.18 –0.02
Kurtosis 11.56 15.17 12.30 5.76 17.72 7.35
Jarque–Bera 5560.55 11014.35 6882.34 591.19 16737.97 1461.98
Probability 0.00 0.00 0.00 0.00 0.00 0.00
Sum 25.94 –114.71 –367.43 –7.75 –174.82 –261.69
Sum Sq. dev. 2289.53 8849373 311966 2494.53 8586206 159222
Observations 1781 1781 1781 1851 1851 1851
Descriptive Stat.
Lead Nickel
Futures Volume Open Interest Futures Volume Open Interest
Mean 0.02 –0.09 –0.15 –0.01 –0.08 –0.14
Median 0.00 –0.02 0.00 0.00 –0.21 –0.04
Maximum 17.72 394.18 54.68 7.26 402.94 73.39
Minimum –5.64 –387.75 –69.13 –7.57 –391.69 –80.16
Std Dev. 1.42 62.35 14.47 1.66 66.69 13.65
Skewness 1.26 –0.27 –0.46 –0.00 –0.31 –0.22
Kurtosis 17.54 17.45 5.77 4.47 15.79 6.92
Jarque–Bera 16154.79 15507.34 633.25 167.31 12643.72 1203.35
Probability 0.00 0.00 0.00 0.00 0.00 0.00
Sum 28.45 –159.43 –268.80 –5.53 –159.00 –258.38
Sum Sq. dev. 3580.56 6917824 372625.5 5117 8228913 344766.2
Observations 1780 1780 1780 1851 1851 1851
(Table 1 continued)
Descriptive Stat.
Zinc
Futures Volume Open Interest
Mean 0.03 –0.10 –0.12
Median 0.00 0.18 0.00
Maximum 9.36 394.22 58.45
Minimum –5.36 –413.99 –91.53
Std Dev. 1.39 65.43 12.78
Skewness 0.42 –0.25 –0.53
Kurtosis 5.81 17.43 6.87
Jarque–Bera 636.02 15433.27 1190.26
Probability 0.00 0.00 0.00
Sum 57.16 –183.44 –209.59
Sum Sq. dev. 3442.56 7599082 289471
Observations 1776 1776 1776
(Table 1 continued)
Samal and Das 55
Table 2. Unit Root Test.
Base Metal Series Test t-Statistics Prob. Decision
Aluminium Futures ADF –28.452 0.000 The aluminium futures series has
no unit root based on both test
PP –43.630 0.000
Volume ADF –22.249 0.000 The aluminium volume series has
no unit root based on both test
PP –156.108 0.000
Open
Interest
ADF –43.913 0.000 The aluminium open interest
series has no unit root based on
both test
PP –45.086 0.000
Copper Futures ADF –43.713 0.000 The copper futures series has no
unit root based on both test
PP –43.718 0.000
Volume ADF –24.152 0.000 The copper volume series has no
unit root based on both test
PP –205.978 0.000
Open
Interest
ADF –45.833 0.000 The copper open interest series
has no unit root based on both
test
PP –45.908 0.000
Lead Futures ADF –44.161 0.000 The lead futures series has no
unit root based on both test
PP –44.186 0.000
Volume ADF –24.243 0.000 The lead volume series has no
unit root based on both test
PP –173.030 0.000
Open
Interest
ADF –33.427 0.000 The lead open interest series has
no unit root based on both test
PP –53.46 0.000
Nickel Futures ADF –43.977 0.000 The nickel futures series has no
unit root based on both test
PP –44.007 0.000
Volume ADF –17.40 0.000 The nickel volume series has no
unit root based on both test
PP –169.67 0.000
Open
Interest
ADF –47.87 0.000 The nickel open interest series
has no unit root based on both
test
PP –48.42 0.000
Zinc Futures ADF –45.068 0.000 The zinc futures series has no unit
root based on both test
PP –45.064 0.000
Volume ADF –45.009 0.000 The zinc volume series has no
unit root based on both test
PP –188.076 0.000
Open
Interest
ADF –49.205 0.000 The zinc open interest series has
no unit root based on both test
PP –54.564 0.000
stationarity of each data series. The results of the stationarity test are tabulated in
Table 2. It is evident from Table 2 that all the series of aluminium, copper, lead,
nickel and zinc are stationary at level. The p-value of both the stationarity test is
zero for all series. As the study uses the Granger causality test to examine the
relationship between the various series of different base metal futures hence, the
VAR framework of the Granger causality test is applicable to only two stationary
series.
The results of both Pearson’s and Spearman’s correlation tests are presented in
Table 3. There are three sets of correlation tests have been carried out for each
base metal. The first set of correlations is carried out between the daily return and
the daily change of volume. Similarly, the second set of correlations is carried
out between the daily return and the daily change of open interest. We have also
explored the correlation between the daily change of open interest and volume. All
56 IIMS Journal of Management Science 14(1)
Table 3. Results of Correlation Test.
Base Metal Correlation Between
Pearson
Correlation
Coefficient Prob.
Spearman
Correlation
Coefficient Prob.
Aluminium Price and Volume 0.099 0.000 –0.072 0.002
Price and Open
Interest
–0.261 0.000 –0.274 0.000
Volume and Open
Interest
0.455 0.000 0.533 0.000
Copper Price and Volume –0.156 0.000 –0.226 0.000
Price and Open
Interest
–0.413 0.000 –0.446 0.000
Volume and Open
Interest
0.487 0.000 0.536 0.000
Lead Price and Volume –0.115 0.000 –0.063 0.000
Price and Open
Interest
–0.244 0.000 –0.273 0.000
Volume and Open
Interest
0.465 0.000 0.370 0.000
Nickel Price and Volume –0.009 0.693 –0.075 0.001
Price and Open
Interest
–0.678 0.000 –0.749 0.000
Volume and Open
Interest
0.242 0.000 0.352 0.000
Zinc Price and Volume 0.380 0.000 0.371 0.000
Price and Open
Interest
0.267 0.000 0.308 0.000
Volume and Open
Interest
0.489 0.000 0.569 0.000
the coefficients of three sets of correlation of different base metals are found to be
significant at a 0.01 per cent level. This is found in both Pearson’s and Spearman’s
correlation. Exceptionally it is observed that the Pearson’s correlation coefficient
between nickel returns and daily change of volume are not significant at the 0.05 per
cent level.
Among all the base metals, the correlation coefficient between return and daily
change of volume of zinc is observed to be the highest. It is observed that there exists
a contemporaneous correlation between the volume change and return of all base
metals. The study also confirms that the correlation between the change in open inter-
est and the return of all base metals is significant. Further, it is also found that a correla-
tion exists between the change in volume and the open interest of all base metal futures.
Granger causality check has three possibilities, namely unidirectional, bidirec-
tional and independent. In a unidirectional situation, one variable influences the
other whereas in a bidirectional environment both variables influence each other.
No relation is established if it is an independent scenario. As per the Granger cau-
sality test, H0 indicates that one series does not Granger cause the other. If the
p-value is less than 0.05 H0 is rejected and concluded that one series Granger
Table 4. Results of Causality Test.
Commodity Null Hypothesis (H0) F-statistics Prob. Decision
Aluminium Price change does not Granger cause volume change 1.610 0.169 Independent
Volume change does not Granger cause price change 0.819 0.513
Price change does not Granger cause open interest change 4.334 0.013 Unidirectional
Open interest change does not Granger cause price change 1.777 0.169
Volume change does not Granger cause open interest change 2.509 0.028 Bidirectional
Open interest change does not Granger cause volume change 4.585 0.000
Copper Price change does not Granger cause volume change 0.836 0.542 Independent
Volume change does not Granger cause price change 0.726 0.628
Price change does not Granger cause open interest change 72.455 0.000 Unidirectional
Open interest change does not Granger cause price change 0.213 0.644
Volume change does not Granger cause open interest change 3.022 0.006 Bidirectional
Open interest change does not Granger cause volume change 4.747 0.000
Lead Price change does not Granger cause volume change 1.581 0.148 Independent
Volume change does not Granger cause price change 0.547 0.772
Price change does not Granger cause open interest change 2.783 0.062 Unidirectional
Open interest change does not Granger cause price change 3.899 0.020
Volume change does not Granger cause open interest change 0.900 0.493 Unidirectional
Open interest change does not Granger cause volume change 2.893 0.008
Nickel Price change does not Granger cause volume change 0.843 0.537 Independent
Volume change does not Granger cause price change 1.076 0.374
Price change does not Granger cause open interest change 94.486 0.000 Unidirectional
Open interest change does not Granger cause price change 2.298 0.100
Volume change does not Granger cause open interest change 2.266 0.034 Bidirectional
Open interest change does not Granger cause volume change 3.367 0.006
Zinc Price change does not Granger cause volume change 0.981 0.436 Independent
Volume change does not Granger cause price change 0.444 0.849
Price change does not Granger cause open interest change 0.752 0.556 Unidirectional
Open interest change does not Granger cause price change 16.086 0.000
Volume change does not Granger cause open interest change 1.875 0.082 Unidirectional
Open interest change does not Granger cause volume change 2.887 0.008
58 IIMS Journal of Management Science 14(1)
causes the other. Granger causality test results presented in Table 4 exhibit that
there is no Granger causality between the volume and return of the base metal
futures market of India. The p-values of the Granger causality between volume
and return of base metal futures are above 0.05 which depicts an independent
scenario. It is further supported by F-statistic values which are also smaller than
the critical F-values.
The study finds unidirectional causality between the return and daily change of
open interest. It is found in all base metal futures traded at MCX, India. For alu-
minium, copper and nickel futures price change, Granger causes the change of
open interest. Open interest fails to Granger cause return in all these commodities.
On the contrary, causality exists from open interest to return in the case of lead and
zinc futures. In the case of aluminium, copper and nickel the study confirms the
existence of bidirectional Granger causality between the change in volume and
change in open interest. Lead and zinc futures exhibit unidirectional causality
from a change in open interest to volume. Interestingly, aluminium, copper and
nickel futures show similar behaviour whereas the causality behaviour of lead and
nickel futures are different from other base metal futures.
Interpretation of the Results
The study supports the MDH as it confirms the existence of a contemporaneous
correlation between the return and change in volume of all base metal futures
traded at MCX, India. Coefficient values of zinc futures are highest among other
base metals which indicates greater informational asymmetry according to Wang’s
(1994) hypothesis. The second set of correlations is carried out between the daily
return and the daily change of open interest. We have also explored the correlation
between the daily change of open interest and volume. All the coefficients of three
sets of correlation of different base metals are found to be significant at a 0.01 per
cent level. This is found in both Pearson’s and Spearman’s correlation. The study
also confirms that the correlation between the change in open interest and the
return of all base metal futures is significant. Further, it is found that a correlation
exists between the change in volume and the open interest of all base metal futures.
In-depth analysis can be carried out by future studies about the above aspects.
The study exhibits no causality between the return and volume change of alu-
minium futures which supports the MDH and contradicts the SIH. Similar results
are found for all base metal futures. This indicates a greater level of market effi-
ciency (Biswas & Rajib, 2011). As the p-values of F-statistics are greater than
0.05 it rejects the presence of causality between return and volume change of all
base metal futures traded at MCX, India.
The study finds unidirectional causality between the return and daily change of
open interest. It is found in all base metal futures traded at MCX, India. For alu-
minium, copper and nickel futures the price change Granger causes the change of
open interest. On the contrary, causality exists from open interest to return in the
case of lead and zinc futures. In the case of aluminium, copper and nickel the
study confirms the existence of bidirectional Granger causality between
Samal and Das 59
the change in volume and change in open interest. Lead and zinc futures exhibit
unidirectional causality from the change in open interest to volume. Interestingly,
aluminium, copper and nickel futures show similar behaviour whereas the causal-
ity behaviour of lead and nickel futures are different from other base metal futures.
Conclusion
After lifting the prohibition on commodity futures trading in India national level
commodity exchanges were established (Biswal, 2008). In recent years there has been
a spurt in the volume and value of commodity futures trading. The study supports the
MDH as it confirms the existence of a contemporaneous correlation between the
return and change in volume of all base metal futures traded at MCX, India.
No causality is found between the return and volume change of the base all-metal
futures contract which supports the MDH and contradicts the SIH. This indicates a
greater level of market efficiency (Biswas & Rajib, 2011). The study finds unidirec-
tional causality between the return and daily change of open interest in all-metal
futures. For aluminium, copper and nickel futures the price change Granger causes
the change of open interest. On the contrary, causality exists from open interest to
return in the case of lead and zinc futures. In the case of aluminium, copper and
nickel the study confirms the existence of bidirectional Granger causality between
the change in volume and change in open interest. Lead and zinc futures exhibit
unidirectional causality from the change in open interest to volume. The relationship
between return and daily change of open interest and between change in volume and
change in open interest is left for in-depth future study. Further nonlinear causality
can also be explored by future research. The study will help the policymakers for
framing future policies for developing the market further. It will also help the inves-
tors and other participants to make an informed decision.
Declaration of Conflicting Interests
The authors declared no potential conicts of interest with respect to the research, author-
ship and/or publication of this article.
Funding
The authors received no nancial support for the research, authorship and/or publication
of this article.
ORCID iD
Laxmidhar Samal https://orcid.org/0000-0002-4713-6584
References
Baek, E., & Brock, W. (1992). A general test for nonlinear Granger causality: Bivariate
model [Working Paper]. Iowa State University and University of Wisconsin.
Bessembinder, H., & Seguin, P. J. (1993). Price volatility, trading volume and market
depth: Evidence from futures markets. Journal of Financial and Quantitative Analysis,
28, 21–39.
60 IIMS Journal of Management Science 14(1)
Biswal, P. C. (2008). Price discovery in futures and spot commodity market in India
[TAPMI Working Paper Series No. 2008/01]. T. A. Pai Management Institute.
Biswas, S., & Rajib, P. (2011). Testing price volume relationship for Indian commodity
futures. Journal of Indian Business Research, 3(2), 117–131.
Chen, G., Mishael, F., & Yu, X. (2004). The price-volume relationship in China’s commod-
ity futures markets. The Chinese Economy, 37(3), 87.
Ciner, C. (2002). Information content of volume: An investigation of Tokyo commodity
futures markets. Pacific-Basin Finance Journal, 10, 201–215.
Copeland, T. E. (1976). A model of asset trading under the assumption of sequential infor-
mation arrival. Journal of Finance, 31, 1149–1168.
Cornell, B. (1981). The relationship between volume and price variability in futures mar-
kets. Journal of Futures Markets, 1, 303–316.
Darolles, S., Le Fol, G., & Mero, G. (2017). Mixture of distribution hypothesis: Analyz-
ing daily liquidity frictions and information flows. Journal of Econometrics, 201(2),
367–383.
De Long, J. B., Shleifer, A., Summers, L. H., & Waldmann, R. J. (1990). Positive feedback
investment strategies and destabilizing rational speculation. Journal of Finance, 45,
379–395.
Dickey, D. A., & Fuller, W. A. (1979). Distribution of the estimators for autoregressive time
series with a unit root. Journal of the American Statistical Association, 74, 427–431.
Foster, A. J. (1995). Volume-volatility relationships for crude oil futures markets. Journal
of Futures Markets, 15, 929–951.
Fujihara, R. A., & Mougoue, M. (1997). An examination of linear and nonlinear causal
relationship between price volatility and volume in petroleum futures markets. Journal
of Futures Markets, 17, 385–416.
Gallant, R., Rossi, P., & Touchen, G. (1992). Stock prices and volume. Review of Financial
Studies, 5, 199–242.
Girard, E., & Biswas, R. (2007). Trading volume and market volatility: Developed versus
emerging stock markets. The Financial Review, 42, 429–459.
Grammatikos, T., & Saunders, A. (1986). Futures price variability: A test of maturity and
volume effects. Journal of Business, 59, 319-330.
Granger, C. (1969). Investigating causal relations by econometric models and cross-spec-
tral methods. Econometrica, 37, 424–438.
Gujarati, D. N. (1995). Basic econometrics (4th ed.). McGraw-Hill.
Hadsell, L. (2006). A TARCH examination of the return volatility–volume relationship in
electricity futures. Applied Financial Economics, 16, 893–901.
Harris, L. (1986). Cross security tests of the mixture of distributions hypothesis. Journal of
Financial and Quantitative Analysis, 21, 39–46.
Hiemstra, C., & Jones, J. D. (1994). Testing of linear and nonlinear Granger causality in the
stock price-volume relation. Journal of Finance, 49, 1639–1664.
Jennings, R. H., Starks, L. T., & Fellingham, J. C. (1981). An equilibrium model of asset
trading with sequential information arrival. Journal of Finance, 36, 143–161.
Karpoff, J. M. (1987). A relation between price changes and trading volume: A survey.
Journal of Financial and Quantitative Analysis, 22, 109–126.
Lokman, G., & Abdulnasser, H. (2005). Stock price and volume relation in emerging mar-
kets. Emerging Markets Finance and Trade, 41(1), 29–44.
Mcmillan, D., & Speight, A. (2002). Return-volume dynamics in UK futures. Applied
Financial Economics, 12, 707–713.
Samal and Das 61
Moosa, I. A., & Silvapulle, P. (2000). The price-volume relationship in the crude oil futures
markets, some results based on linear and non-linear causality testing. International
Review of Economics and Finance, 9, 11–30.
Nevin, Y., Erdem, C., & Erdem, M. S. (2006). Testing for linear and nonlinear Granger
causality in the stock price-volume relation: Turkish banking firms’ evidence. Applied
Financial Economic Letters, 2, 165–171.
Peguin-Feissolle, A., & Terasvirta, T. (1999). A general framework for testing the Granger
noncausality hypothesis [Working Paper Series in Economics and Finance for Stock-
holm School of Economics No. 343]. Stockholm School of Economics.
Phillips, P. C. B., & Perron, P. (1988). Testing for a unit root in time series regression.
Biometrica, 75(2), 335–346.
Puri, T. N., & Philippatos, G. C. (2008). Asymmetric volume-return relation and concen-
trated trading in LIFFE futures. European Financial Management, 14(3), 528–563.
Silvapulle, P., & Choi, J.-S. (1999). Testing for linear and nonlinear Granger causality in
the stock price-volume relation: Korean evidence. The Quarterly Review of Economics
and Finance, 3(1), 59–76.
Srinivasan, K., Murthy, K., & Hajiri, S. A. (2016). Does the sequential information arrival
and mixture of distribution holds for stock futures market in India. International Jour-
nal of Business, Economics and Law, 9(1), 33–45.
Toda, H. Y., & Yamamoto, T. (1995). Statistical inference in vector auto regressions with
possibly integrated processes. Journal of Econometrics, 66, 225–250.
Touchen, G., & Pitts, M. (1983). The price variability-volume relationship on speculative
markets. Econometrica, 51, 485–505.
Wang, J. (1994). A model of competitive stock trading volume. Journal of Political Econ-
omy, 102, 127–168.
Zwergel, B., & Heiden, S. (2012). Intraday futures patterns and volume-volatility relation-
ship: The German evidence. Review of Managerial Science, 8(1), 29–61.
