Significance of Open Interest-Based and Trading Volnme-Based

Predictors-An Empirical Study of the Option Market in India

S,vagata Seu
Professor, Department of Commerce
University of Calcutta
e-mail: [email protected]

Avijit Sikdar
Assistant Professor, Department of Commerce
Kidderpore College, Kolkata
e-mail : [email protected]

Abslract: The information content of the option market for predicting the future price index in the
underlying cash market in India has been examined in this paper by applying the method of open interest
and trading volume-based predictors. Daily data for both price as well as non-price variables for the study
period of July 2014 to December 2014 have been employed to explore the above relationship. In the study,
open interest-based predictors are found lo be significant in predicting the future underlying cash market.
But as far as volume-based indicators are concerned, it shows some mixed evidence. Both the indicators
are statistically significant at any conventional level of significance when all of them are taken together.
However, volume-based predictors turned out statistically insignificant when it estimates alone with the
current cash index, the future price of the Index in cash market. Based on the value of Adjusted R Squared
and F-Statistics in the study period we can conclude that informational role in the Indian Derivative market
has played an important role in predicting future price of underlying cash market.
Key-words : Open interest based predictor, trading volume based predictor, underlying assets.

1. Introduction
Derivative securities are considered as additional means for infonned traders to trade on their
infonnation and for others to discover that infonnation. Derivatives may not only lead the underlying
assets in imparting information, they may also provide infonnation that simply cannot be inferred
from the markets in underlying assets. This paper examines the role of non price variables of options
market like open interest and trading volume in conveying infonnation about the future movement of
the underlying asset. This study shows that the activity in the equity options market seems to contain
infonnation about future stock price that can be exploited'for trading purposes. Financial economists
have long been interested in the process of price fonnation when informed traders, uninformed
liquidity (or noise) traders and market makers interact in the asset market. Derivative instruments,
like option contracts, enhance infom1ational efficiency of the underlying's market. An option is a
contract between two parties giving the taker (buyer) the right, but not the obligation, to buy or sell
an underlying asset at a predetermined price on or before the predetennined date. It is one of the
important derivative instrument traded in derivative exchanges all over the world. It is well known
that trading in options may be more attractive than trading in underlying assets in spot (cash) market

II
 Business Studies-Vol : XXXIII & XXXIV, 2012 & 2013

due to economic incentives provided by reducing transaction cost, capital requirements and trading
restrictions, commonly seen in the equity market. The attractiveness in the equity market can be
proved by the increasing trend in the total traded value in the option market over a period of time
in a fast moving developing country like India. Option can be used for both hedging as well as
speculation. It is well documented in the research that not only the option prices, but also the non-
price variables, such as open interest, trading volume, etc., fonn the option market can affect the
stock prices in the underlying equity market.

2. Objective of the Study

The study investigates the impact of non-price variables viz. open interest and trading volume of
index option from the option market in predicting the price index; viz. Nifty index in the underlying
cash market in India. By applying the method of open interest and volume-based predictors for
both call and put options, this study empirically investigates the hypothesis (slated below) that
the above non-price variables in the option market cannot be used to predict the future price index in
the underlying cash market. This study is based on daily data for both price as well as non price
variables over the sub period June 2014 to December 2014. It reveals that the open interest based
predictors are significant in predicting the underlying spot price index in the sub period.
H0 : Non-price variables in the option market cannot be used to predict the future price index
in the underlying cash market.
H1 : Non-price variables in the option market can be used to predict the future price index
in the underlying cash market.

3. Review of Literature
A considerable amount of literature deals with 1he interrelationship between the derivative markets,
viz., options market, and underlying cash market. Different issues addressed in those studies include
(1) the effect of option listing on the volatility, bid-ask spread and liquidity of underlying cash
market; (2) the options expiration effect on the prices of underlying cash market at maturity~ (3) the
lead-lag relationship among the price as well as non-price variables from both options market;
(4) the role of the oplion market in discovering the price in underlying cash market etc. A brief
review of some of the past literature, relevant to this study, is presented below.
Many of the previous research studies have tried to examine the options listing effect on the
volatility, liquidity etc., of the underlying cash market all over the world, document the option
expiration effect on the prices of the underlying.cash market at the time of maturity, and report the
lead-lag relationship among the price as well as non price variables from both options and underlying
cash market. Since the main aim of the study is to examine whether the open interest and trading
volume from the option market are significant in explaining the future movement of prices in the
underlying cash market, it could not be explained by the previous studies discussed here- which
cover all the aspecls of the interrelationship between the options and cash market.
Bhuyan, Rafiqul and Chaudhury (2001) examined the role of option market's open interest to
capture information about the future movement of underlying stock and show that the trading

12
 Swaga1a Sen and Avijil Sikdar

strategies based on this predictor yields better results as compared to the buy-and-hold and passive
covered call strategics. Further, Bhuyan and Yan (2002) developed several price predictors from the
open interests and trading volumes of individual stocks from the option market and concluded that
these factors exhibit significant explanatory and predictive power for the future stock prices. The
encouraging results provided by this study are one of the driving forces for the present initiative.
There is a general agreement with the basic premise drawn by Bhuyan and others (2001), and
Bhuyan and Yan (2002) that as against the total volume considered by one of the previous studies,
net open interest of call options and put options together should provide a better indication of the
future stock price movement Srivastava (2003) examined the informational role of stock option
market in India using this approach. However, it was conducted for a small period, i.e., four months.
The present study is among one of the earlier attempts to document the impact of non-price variables
of stock option market on the underlying stocks in the Indian context with a longer sample period.
Through this study, the authors seek to provide a valuable input to the uninformed investors
using, which they can disc.over, the price of underlying asset in a more efficient as well as effective
manner. The results are encouraging and provide a basis for further examination of different
variables in option market and their interrelationships with the underlying stock price.

4. Data and Methodology

The National Stock Exchange (NSE) India has commenced trading in S&P CNX Nifty Index Options
from June 4, 2001. This study covers the study period July 2014 to December 2014. Daily data
relating to price as well as non-price variables such as spot price index in the underlying cash
market, and open interest, trading volume, different strike price, etc. in the option market, have
been collected for the sub periods. There are three types of options contracts available in the Indian
market-which gels matured in one month, two months and three months. Since the option price
on the nearest contract is characterized by high level of activity, all the data for both price and non
price variables of index option on the first nearest contract, i.e., on the next month contract, are
taken into account. Further, all the data on the expiration day have been excluded from the study in
order to avoid possible bias expected lo occurr due to expiration effect. All the relevant data have
been collected from NSE website and all the calculations have been made using MS-Excel and
SPSS (version 16.0).
Methodology
The interrelationship between the open interest and trading volume in the option market and the
prices in the underlying cash market can be measured by various techniques. But the present study,
is based on a simple methodology used by Bhuyan, Rafiqul and Chaudhury (200 I), Bhuyan and Yan
(2002) and Srivastava (2003) for investigating the significance of the net open interest and trading
volume in the Index option market. The term and notations applied in the methodology are same
as used in the above studies.
Let 'T' be the time of maturity of a set of call and put option. The current price of the cash
index is assumed to be II; while, xr and X fare assu_med to be the set of strike prices for call

13
 Business Studies-Vol : XXXIU & XXXIV, 1012 & 1013

and put options, where i = 1,2, ...... k; j = 1,2,....... ,m. Let oiJ and o)t be the net open interest at
the current time ·c for call and put option with the strike price Xf and X'j respectively. Similarly~
V{ and V f are assumed to be the trading volume for a call and put option at time t with strike price
Xf and respectively.
Now the two predictors open interest-based predictor and volume-based predictors that are
proposed to be used in predicting the price of underlying cash index, can be defined by using
the above variables.
The call option Open interest-based predictor (COP) can be defined as:

of =I.f=k~~xf
In the above equation, of
represents the call option ope~ interest-based predictor at a
time t; while k denotes number of different types of call options showing some non zero open
interest. JV(: is the weight of call options with strike price of Xf.
The put option Open interest-based predictor (POP) can be defined as:

of =I.1=k~i xf
In the above equation,of represents the put option open interest-based predictor at a time t;
while k denotes number of different types of put options showing some non zero open interest.
~f is the weight of put options with strike price of Xf .
The call option Volume-based predictor (VOP) can be defined as:

v,c =I.f=kq~Xf
In the above equation, v,C represents the call option value-based predictor at a time t; while
k denotes number of different types of call options showing some non zero open interest. qfi
is the weight of call options with strike price of xr .
The put option Volume-based predictor (VOP) can be defined as:

v/ =I.7=kq;,Xf
In the above equation, v,P
represents the put option value-based predictor at a time t; while
k denotes number of different types of put options showing some non zero open interest. is qf,
lhe weight of put options with strike price of X f.
Based on the open interest and volume-based predictors for call and put options, it is easy to find
out the relative significance of each of these predictors by using multiple regression model such that:

LnlT=a 0 + a/n(T-t) + a/nl1 + a/nO~+ a/nOJ' + a 5 JnV,' + a/nV/ +£

1

14
 Swagata Sen and Avijit Sikdar

Where Ir and 11 are lhe stock price at the dale of maturity and at the current date respectively.
(T-t) represents the actual time to maturity; Of and of
are the open interest-based predictors;
while v,c and Vlare the volume-based predictors. eI denotes error term assumed to be white noise
such that e- N (0, (j 2) and Cov ( £ 1 , e1 _ 1) = 0. The natural logarithms of lhe entire variables are
used to account for the heteroskedasticily, i.e. unequal variance among the variables. ll is to be noled
that, since the variable (T-t) is unable to improve the overall explanatory power (as shown by
adjusted R2) hence, it is removed from the above equation. Now the revised equation is

LnlT=ao + a/n/1 +a/no;+ a/110{'+ a/nV,C + a/nV/ + , 1.... (Eq. l)

In order to find out the relative significance of open interest-based predictors and volume-based
prediclors separately in the matter of price fixation in the underlying cash market, the following
regression equations have been estimated excluding one set of parameter (either open interest or
Lrading volume):

LnlT=ao + a/n/ +a/no;+ a/110{'+ , .... (Eq. 2)

1 1
LnlT=a + a/n/ + a/nV,C + a/nV/+ , .... (Eq. 3)
1
0 1
All the above regression equations have been solved for the study period July 2014 to December
2014. To remove the influence of expiration of option contracls at the date of maturity, all the
equations have been estimated by excluding the date of expiration of the contract in one case and
by excluding the previous five days up to the expiration in the other.

5. Results
The hypothesis that the non price variables viz., open interest and trading volume in the option
market, do not have the significant explanatory power in predicting the future price of the underlying
cash index is tested through the multiple regression equations (eq 1 to 3). Though the price index
at the time of maturity and at the current date can be directly observable, but the value of other
independent variables have been calculated as in the same way as mentioned in the 'data and
methodology' section.
The meaning of notations used in regression equation in the following table are as follows.
Lnlr Natural log of closing index at maturity
LCI(lnl 1) Natural log ofcurrent index
LOIBPCO (In of) Natural log of open interest-based predictor based on call option

LVBPCO (In v c) Natural log of value-based predictor based on call option

1
LOIBPPO (In O{') Natural log of open interest-based predictor based on put option

LVBPPO (In Vl) Natural log of value-based predictor based on call option

JS
 Business Studies-Vol : XXXlll & XXXIV, 2012 & 2013

Table 1 : Regression Result for the Study Period

Panel A : Excluding only the expiration date

Variable Coefficienl S.E t-Slalislics Prob.
Intercept -3.216 0.731 -4.397 0.000
LCl(lnl,) 0.427 0.195 2.183 0.031

LOIBPCO (In of) 1.244 0.194 6.409 0.000

LVBPCO (In v,c) 0.025 0.123 0.203 0.839

LOIBPPO (In of) 0.298 0.216 1.379 0.171

LVBPPO (/nV/) -0.636 0.17 -3.738 0.000

Adjusted R square = 0.823; F = 103.39
Panel B: Excluding last five days up to the expiration date
Variable Coe.fficienl S.E t-Statistics Prob.
Intercept -4.051 0.959 -4.223 0.000
LCl(lnl,) 0.494 0.221 2.241 0.Q28
LOlBPCO (In on 1.558 0.287 5.42 0.000
LVBPCO (lnv c) -0.083 0.171 -0.487 0.628
1
LOIBPPO (In Of) -0.003 0.334 -0.01 0.992
LVBPPO (lnVP) -0.515 0.241 -2.133 0.036
Adjusted R square = 0.803 ; F = 71.32
The findings have been broadly categorised into two parts. The first calegory includes all the resulls
derived from regression equation 1 where both open interest-based predictors and volume-based
predictors for both call option and put option are taken into consideration to observe their relative
significance in predicting the future price index in the underlying cash market. The results are
reported in table 1. While the second category is derived from equation 2 and equation 3 where
both the predictors for call and put options arc considered separately (through different equations)
and are reported in table 2 and table 3.
The results in all the tables (1 to 3) are divided into 2 panels. Panel A in each table reports 1he
results which are based on the data set where all the data only on the date of expiration of option
contract are excluded. While the panel B reports the results based on the data set excluding all the
data of previous five days up to the date of expiration of option contract.
Table I reports the results based on regression equation 1 where both open interest-based
predictors and volume-based predictors for both call option and put option are taken into

16
 Swagata Sen and Avijit Sikdar

consideration to observe their relative significance in predicting the future price index in lhe
underlying cash market during the study period July to December 2014 i.e. just after the end of
coalition government UPA 2 in India. The table shows that there is not much difference in the result
reported in Panel A and Panel B. Both panels reveal that open interest-based predictors based on
call option and volume-based predictors based on put option are found to be significant in explaining
future movement of the underlying price movement at conventional I% or 5% level of significance. It
is noted that result of open interest-based predictors during the study period found significant positive
coefficient and negative coefficient with volume based put option predictors. It indicates that the
investors expecting the price of underlying cash index to increase and they prefer to buy the call
option at higher strike price and prefer to sell put option. In other words, the relationship among
the open interest-based predictor and future price is positive for call option and negative among
the value-based predictors and future price in case of put option.
From the table 2 it can be observed that there is not much difference in the results if an open
interest-based indicator is considered separately in estimating the equation 2. This table confirms
that open interest-based predictor during the study period of Panel A and Panel B shows significant
explanatory power in predicting the future price index in the underlying cash market. From table 3
it can be observed that there is much difference in the results if value-based indicators are considered
in predicting the future price of underlying cash market. These indicators fail to explain statistically
the future price index at conventional level I% and 5% level of significance.
Table-2 : Regression Result for Open Interest-based Predictor during the Study Period
Panel A : Excluding only the expiralion date
Variable Coefficient S.E t-Statistics Prob.
Intercept -3.'147 0.966 -3.250 0.002
LC! (ln/1) 0.121 0.12 0.008 0.994

LOIBPCO (/nO,C) 1.72 0.256 6.435 0.000

LOII3PPO (lnOf') -0.425 0.245 -0.237 0.813

Adjusted R squared - 0.802; F - 149.44
Panel B : Excluding last five days up to the expiration date
Variable Coefficient S.E t-Statistics Prob.
Intercept -3.747 0.966 -3.877 0.000
LC! (Inf,) 0.121 0.12 1.002 0.319

LO!BPCO (/nO,C) 1.72 0.256 6.714 0.000

LOIBPPO (In of) -0.425 0.245 -1.733 0.087

Adjusted R squared - 0.795; F - 112.1

11
 Business Studies-Vol : XXXlll & XXXIV, 1012 & 2013

Table 3 · Regression Result for Volume-based Predictor during the Study Period
Panel A : Excluding only the expiration date

Variable Coefficienl S.E t-Statislics Prob.

lnten:ept 3.252 0.528 6.159 0.000

LCI (/n/,) 0.002 0.294 0.005 0.996

LVBPCO (/n v <) 0.333 0.184 l.814 0.072

1
LVBPPO (/n v,P) 0.304 0.208 l.461 0.147
Adjusted R squared = 0.58; F = 52.57
Panel B : Excluding last five days up to the expiration date
Variable Coefficient S.E t-Statistics Prob.
Intercept 3.894 0.624 6.24 0.000
LC! (In/,) -0.047 0.325 -0.143 0.887
LVIJPCO (Inv;) 0.226 0.243 0.929 0.356

LVIJPPO (/nV,P) 0.388 0.24 l.62[ 0.109

Adjusted R squared = 0.803; F = 71.32
However, there is a litlle change (excepl table 3) in the value of adjusted R square while going
from Panel A to Panel B (table I and 2) and there is a significant decrease in their F-value in all
cases (except table 3) while going from Panel A to Panel B (table 1 and 2), which confirms that the
overall explanatory power of non-price variable in option market is influenced by not only the date
of expiration but also the previous five days up to the date off expiration.

6. Conclusion
By applying daily data on both price and non-price variables from equity and option markets, an
effort has been made to examine the significance of open interest and Lrading volume from the
index option market in explaining the future price movements in the underlying cash market in [ndia.
lbis study documents significant results in explaining the price movement when all the variables
i.e. price variable and non-price variable in the recent time in the option market of [ndia but when
volume-based indicators are considered alone they fail to capture the future price movement of cash
market index. However we can conclude lhat the informational role of the option market non-
price variable in predicting the future price index played an important role in [ndian capital market.
References
Bhuyan. Rafiqul and Chaudhury, Mo. (2001). "Trading on the Infonnation Content of Open Interest:
Evidence from the US Equity Options Market", Working Paper, McGill.

1,
 Swagata Sen and Avijit Sikdar

I3huyan, R. and Yan, Y. (2002). "Informational role of Open Interests and Volumes: Evidence from option
markets", Paper presented at Twelfih Annual Asia Pacific futures Research Symposium, December.
Kumar, R., Sarin, A. and Shastri, K. ( 1988). "The impact of option trading of the market quality of underlying
security: An Empirical Analysis", Jo11rnal of Finance, 53, pp. 717-732.
Mukherjee, K.N. and Misra, R.K. (2007). "lnfonnalion Role of Non-Price Variables: An Empirical Study
of the Indian Option Market", The ICFAI Journal of Applied Finance, 13 (2).
Srivastava, S. (2003). "Information conlent of trading volume and Open Interest-An empirical Study of
Stock Option market in India", Journal of Finance, 45, pp. 191-220.

URL
www.nscindia.com

Annexurc
Steps to be followed while calculating the open interest-based and volume-based predictor.a for both
call and put option
Step I : Extract the data for the nearest month contract and also nonzero open interest.
Step 2 : Remove all the data other than current data, date of maturity, strike price and open interest.
Step 3 : Calculate the weights proposed to be used for different strike price. This step follows two sub steps.
(i) Calculate the sum of nonzero open interest for different strike prices in a given trading day which
is denoted by Lf oi
= 1 or =l Lj of,, i and j refers to the set of strike prices for call option
and put option respectively; and t is the current lime.
(ii) Divide lhe open interest for different strike price by that sum for all the days, such that
If;~= O;~ n;f= 1Oi~ or W_fi = 0)112,f=l 0)1-
Step 4 : Multiply each strike price with their corresponding weights calculated as per the above sleps.
Step 5 : Sum up the resultant strike prices for a given day as given in step 4 to get a single strike price
in a single day. The final figure is lermed as open interest (Call or Put)-based predictors.
The volume-based predictors are also calculated in the same way.

19