Electronic copy available at: http://ssrn.com/abstract=1500327

Impact of Derivatives on Indian Stock Market1

Ravi Agarwal2 ravi.agarwal@bimtech.ac.in

Shiva Kumar3 shiva.kumar10@bimtech.ac.in

Wasif Mukhtar3 wasif.mukhtar10@bimtech.ac.in

Hemanth Abar3 hemant.abar10@bimtech.ac.in

Keywords: Derivatives, Volatility, GARCH

JEL Classifications: C25, C22, G12, N25, G15

1 Presented in Indian Finance Summit 2009, organized by BIMTECH, Greater Noida, India. 2 Associate Professor of Finance, BIMTECH, Greater Noida, India. 3 Class of 2010, PGDM program, BIMTECH, Greater Noida, India.

Electronic copy available at: http://ssrn.com/abstract=1500327

Impact of Derivatives on Indian Stock Market Introduction:

Derivatives, such as futures or options, are financial contracts which derive their value from a spot price, which is called the underlying. Derivative products like index futures, stock futures, index options and stock options have become important instruments of price discovery, portfolio diversification and risk hedging in stock markets all over the world in recent times. In the last decade, many emerging and transition economies have started introducing derivative contracts. Ever since index futures were introduced by the National Stock Exchange (NSE) in June 2000, there has been a lot of controversy regarding their usage. Derivatives are known to be a double edged sword, you can use it to kill enemies, kill yourself or for self-defense.

Introduction of derivative products, however, has not always been perceived in a positive light all over the world. It is, in fact, perceived as a market for speculators and concerns that it may have adverse impact on the volatility of the spot market. Government had earlier banned futures trading in commodities on the belief that they were over speculated and caused an inflationary situation. Now trading has been resumed on four banned commodities, after the inflation rate has moderated. There is also a hope that government may soon lift the trading ban on food grains. This offers great relief for importers and exporters to hedge their positions. Earlier, an expert committee headed by Abhijit Sen was constituted to study the impact of futures trading on agricultural commodity prices. Wheat, urad, tur and rice were among the products that were considered in the report.

This paper tries to study whether the Indian stock markets show some significant change in the volatility after the introduction of derivatives trading. This paper also tries to examine whether decline or rise in volatility can be attributed to introduction of derivatives alone or due to some other macroeconomic reasons.

Literature Review

There is a common belief that stock index futures are more volatile than underlying spot market because of their operational and institutional properties. The close relationship between the two markets makes the transfer of volatility possible from futures market to the underlying spot market. It is therefore, not surprising that the inception of futures contract relating to stock market has attracted the attention of researchers all over the world, and also led some observers to attribute stock market volatility to futures trading. Various studies have been conducted to assess the impact of derivatives trading on the underlying market mostly related to US and other developed countries markets. Very few studies attempted to know the impact of introduction of derivatives trading in emerging market economies like India.

Both theoretical and empirical studies were carried out to assess the impact of listing of futures and options on the cash market. Two main bodies of theory about the impact of derivatives trading on the spot market are prevailing in the literature and both are contradicting each other. One school of thought argues that the introduction of futures trading increases the spot market volatility and thereby, destabilizes the market. They are the proponents of Destabilizing forces hypothesis. (Lockwood and Linn, 1990)) They explain that derivatives market provides an additional channel by which information can be transmitted to the cash markets. Frequent arrival and rapid processing of information might lead to increased volatility in the underlying spot market. They also attribute increased volatility to highly speculative and levered participants.

Others argue that the introduction of futures actually reduces the spot market volatility and thereby, stabilises the market. They are the proponents of Market completion hypothesis. (Satya Swaroop Debasish, 2007). Kumar et al (1995) argued that derivatives trading helps in price discovery, improve the overall market depth, enhance market efficiency, augment market liquidity, reduce asymmetric information and thereby reduce volatility of the cash market. The impact that the derivatives market has on the underlying spot market remains an issue debated again and again with arguments both in favour and against them.

This study seeks to examine the volatility of the spot market due to the derivatives market. Whether the volatility of the spot market has increased, decreased or remained the same. If increased then, what extent it is due to futures market. We use Autoregressive framework to model returns volatility. To measure volatility in the markets, the VIX (Volatility Index) computed by the National Stock Exchange is used. To eliminate the effect of factors other than stock index futures (i.e., the macroeconomic factors) determining the changes in volatility in the post derivative period, the model is used for estimation after adjusting the stock return equation for market factors.

The studies is in the Indian context have evaluated the trends in NSE and not on the Stock Exchange, Mumbai (BSE) for the reason that the turnover in NSE captures an overwhelmingly large part of the derivatives market.

We use Nifty Junior as surrogate indices to capture and study the market wide factors contributing to the changes in spot market volatility. This gives a better idea as to whether the introduction of index futures in itself caused a decline in the volatility of spot market or the overall market wide volatility has decreased, and thus, causing a decrease in volatility of indices on which derivative products have been introduced. The volumes on NIFTY also has a large impact on the volatility, thus in the model to measure volatility volumes are also considered as a factor. We seek to compare this volatility with the volatility prevailing in the market before the index futures (i.e. Nifty futures) and check if it is statistically significant.

Data and Methodology:

We used daily data for our study obtained from different sources. Data on daily movement of NSEs indices such as S&P CNX Nifty, Nifty Futures VIX (Volatility Index) and NIFTY Volumes were collected from www.nseindia.com. We propose to construct our model on the recent data of these indices.

To also addresses the issue of whether introduction of derivatives (Futures and Options) has been the only factor responsible for the change in volatility we include returns from a surrogate index Nifty Junior into mean equation to account for the additional factors affecting the volatility of the market. It is worthy to note that Nifty Junior is not traded in futures market.

We are using two models arrive at a conclusion on our studies. The first model examines if there is any significant variances in thee movements of NIFTY and NIFTY Junior. In the second model i.e. the auto-regressive model, the volatility is given as equation of NIFTY Futures, NIFTY Junior and Volumes. We will be explaining the data and methodology of each model in detail when explaining that model.

MODEL 1- Using Variances in NIFTY and NIFTY Junior

Standard Deviation Value

NIFTY 0.016714

NIFTY JUNIOR 0.019317

Standard Deviation measures the volatility of the data and would be the most appropriate to use for comparison. When we compare the Standard Deviation, it is higher in case of NIFTY JUNIOR, we can infer that NIFTY is less volatile than NIFTY JUNIOR. We can test it using F statistic, test for variances. We get the following result from MS Excel.

Null Hypothesis:

There is no significant difference between the variances in price changes in NIFTY and Nifty Futures.

Alternate Hypothesis:

There is significant difference between the variances in price changes in NIFTY and Nifty Futures.

F-Test Two-Sample for Variances

NIFTY %age Chg

%AGE NIFTY JR

Mean 0.000427224 0.000374095 Variance 0.00027639 0.000371323 Observations 2125 2125 Df 2124 2124 F 0.744337104 P(F

MODEL 2 Using Regression Analysis

Data and Methodology

We use a regression model to capture the volatility of NIFTY. The dependent variable, we take as the Volatility Index (VIX). The volatility in the market is generally due to the Macro-Economic Variables, Futures (We are interested in checking this), and stock volumes. Thus when we regress this model and get the result using SPSS.

Yt = b0 + b1Xt-1 + b2X2 + b3X3 + b4X4 + e

where Y = Volatility Index

i = regression coefficients

Xt-1= Volatility Index

X2 = Nifty 1 Month Futures

X3 = Nifty Junior

X4 = Nifty Volumes

e = random error term

Volatility Index VIX

Volatility Index is a measure of markets expectation of volatility over the near term. Volatility is often described as the rate and magnitude of changes in prices and in finance often referred to as risk. Volatility Index is a measure, of the amount by which an underlying Index is expected to fluctuate, in the near term, (calculated as annualised volatility, denoted in percentage e.g. 20%) based on the order book of the underlying index options. Volatility Index is a good indicator of the investors perception on how volatile markets are expected to be in the near term. Usually, during periods of market volatility, market moves steeply up or down and the volatility index tends to rise. As volatility subsides, option prices tend to decline, which in turn causes volatility index to decline.

Computation methodology:

The generalized formula used in the India VIX calculation is:

Durbin-Watson Statistic

The DurbinWatson statistic is a test statistic used to detect the presence of autocorrelation in the residuals from a regression analysis. It is named after James Durbin and Geoffrey Watson. Its value always lies between 0 and 4.

A value of 2 indicates there appears to be no autocorrelation. If the DurbinWatson statistic is substantially less than 2, there is evidence of positive serial correlation. As a rough rule of thumb, if DurbinWatson is less than 1.0, there may be cause for alarm. Small values of d indicate successive error terms are, on average, close in value to one another, or positively correlated. Large values of d indicate successive error terms are, on average, much different in value to one another, or negatively correlated

Determine based on Eigenvalues.

An eigenvalue represents the amount of variance associated with the factor. Generally only factors with an Eigenvalue of >1.0 is included.

A Durbin-Watson Statistic below 1 is a cause of alarm. There is auto correlation between the variables in VIX. We go for Auto-Regressive model.

Model Summary

Model R

R Square

Adjusted R Square

Std. Error of the Estimate

Change Statistics Durbin-Watson

R Square Change

F Change df1 df2

Sig. F Change

1 .930a .866 .865 4.07893 .866 1712.227

1 266 .000

2 .932b .869 .868 4.03100 .004 7.364 1 265 .007 3 .934c .873 .872 3.97537 .004 8.468 1 264 .004 4 .936d .875 .873 3.95245 .002 4.071 1 263 .045 2.228 a. Predictors: (Constant), AUTOVIX b. Predictors: (Constant), AUTOVIX, FUT1 c. Predictors: (Constant), AUTOVIX, FUT1, NIFTYJR

Model Summary

Model R

R Square

Adjusted R Square

Std. Error of the Estimate

Change Statistics

Durbin-Watson

R Square Change

F Change df1 df2

Sig. F Change

1 .694a .481 .479 8.03216 .481 247.657

1 267 .000

2 .828b .685 .683 6.26741 .204 172.531

1 266 .000 .517

a. Predictors: (Constant), FUT1 b. Predictors: (Constant), FUT1, NIFTYJR c. Dependent Variable: VIX

d. Predictors: (Constant), AUTOVIX, FUT1, NIFTYJR, NIFTYVOL e. Dependent Variable: VIX

ANOVA

Model Sum of Squares df Mean Square F Sig.

1 Regression 28487.464 1 28487.464 1712.227 .000a

Residual 4425.618 266 16.638 Total 32913.083 267

2 Regression 28607.116 2 14303.558 880.277 .000b Residual 4305.967 265 16.249 Total 32913.083 267

3 Regression 28740.933 3 9580.311 606.211 .000c Residual 4172.149 264 15.804 Total 32913.083 267

4 Regression 28804.523 4 7201.131 460.964 .000d Residual 4108.560 263 15.622 Total 32913.083 267

a. Predictors: (Constant), AUTOVIX b. Predictors: (Constant), AUTOVIX, FUT1 c. Predictors: (Constant), AUTOVIX, FUT1, NIFTYJR d. Predictors: (Constant), AUTOVIX, FUT1, NIFTYJR, NIFTYVOL e. Dependent Variable: VIX

Coefficients

Model

Unstandardized Coefficients

Standardized Coefficients

t Sig.

Correlations Collinearity Statistics

B Std. Error Beta

Zero-order Partial Part

Tolerance VIF

1 (Constant)

2.707 .879 3.079 .002

AUTOVIX

.926 .022 .930 41.379

.000 .930 .930 .930 1.000 1.000

2 (Constant)

9.538 2.663 3.582 .000

AUTOVIX

.868 .031 .872 28.197

.000 .930 .866 .627 .516 1.937

FUT1 .000 .000 -.084 -2.714

.007 -.690 -.164 -.060 .516 1.937

3 (Constant)

22.118 5.058 4.372 .000

AUTOVIX

.793 .040 .796 19.842

.000 .930 .774 .435 .298 3.353

FUT1 -.006 .002 -.547 -3.376

.001 -.690 -.203 -.074 .018 54.707

NIFTYJR .002 .001 .419 2.910 .004 -.585 .176 .064 .023 43.163

4 (Constant)

26.452 5.469

4.837 .000

AUTOVIX

.794 .040 .797 19.983

.000 .930 .776 .435 .298 3.353

FUT1 -.007 .002 -.573 -3.545

.000 -.690 -.214 -.077 .018 55.058

NIFTYJR .002 .001 .440 3.065 .002 -.585 .186 .067 .023 43.391

NIFTYVOL

-6.777E-7

.000 -.045 -2.018

.045 .107 -.123 -.044 .969 1.032

Coefficients

Model

Unstandardized Coefficients

Standardized Coefficients

t Sig.

Correlations Collinearity Statistics

B Std. Error Beta

Zero-order Partial Part

Tolerance VIF

1 (Constant)

2.707 .879 3.079 .002

AUTOVIX

.926 .022 .930 41.379

.000 .930 .930 .930 1.000 1.000

2 (Constant)

9.538 2.663 3.582 .000

AUTOVIX

.868 .031 .872 28.197

.000 .930 .866 .627 .516 1.937

FUT1 .000 .000 -.084 -2.714

.007 -.690 -.164 -.060 .516 1.937

3 (Constant)

22.118 5.058 4.372 .000

AUTOVIX

.793 .040 .796 19.842

.000 .930 .774 .435 .298 3.353

FUT1 -.006 .002 -.547 -3.376

.001 -.690 -.203 -.074 .018 54.707

NIFTYJR .002 .001 .419 2.910 .004 -.585 .176 .064 .023 43.163

4 (Constant)

26.452 5.469

4.837 .000

AUTOVIX

.794 .040 .797 19.983

.000 .930 .776 .435 .298 3.353

FUT1 -.007 .002 -.573 -3.545

.000 -.690 -.214 -.077 .018 55.058

NIFTYJR .002 .001 .440 3.065 .002 -.585 .186 .067 .023 43.391

NIFTYVOL

-6.777E-7

.000 -.045 -2.018

.045 .107 -.123 -.044 .969 1.032

a. Dependent Variable: VIX

Excluded Variables

Model Beta In t Sig. Partial Correlation

Collinearity Statistics

Tolerance VIF Minimum Tolerance

1 NIFTYJR -.058a -2.114 .035 -.129 .654 1.528 .654

NIFTYVOL -.037a -1.624 .105 -.099 .976 1.024 .976 FUT1 -.084a -2.714 .007 -.164 .516 1.937 .516

2 NIFTYJR .419b 2.910 .004 .176 .023 43.163 .018 NIFTYVOL -.040b -1.772 .078 -.108 .974 1.026 .513

3 NIFTYVOL -.045c -2.018 .045 -.123 .969 1.032 .018 a. Predictors in the Model: (Constant), AUTOVIX b. Predictors in the Model: (Constant), AUTOVIX, FUT1 c. Predictors in the Model: (Constant), AUTOVIX, FUT1, NIFTYJR d. Dependent Variable: VIX

Collinearity Diagnostics

Model Dimension Eigenvalue

Condition Index

Variance Proportions

(Constant) AUTOVIX FUT1 NIFTYJR NIFTYVOL 1 1 1.959 1.000 .02 .02

2 .041 6.912 .98 .98 2 1 2.894 1.000 .00 .00 .00

2 .100 5.380 .00 .24 .08 3 .006 22.885 1.00 .76 .92

3 1 3.836 1.000 .00 .00 .00 .00 2 .152 5.027 .00 .09 .00 .00 3 .012 18.054 .11 .32 .00 .04 4 .000 94.962 .89 .58 1.00 .96

4 1 4.814 1.000 .00 .00 .00 .00 .00 2 .156 5.560 .00 .08 .00 .00 .00 3 .022 14.629 .00 .28 .00 .01 .40 4 .007 26.339 .16 .07 .01 .03 .58 5 .000 107.455 .84 .57 .99 .95 .02

Collinearity Diagnostics

Model Dimension Eigenvalue

Condition Index

Variance Proportions

(Constant) AUTOVIX FUT1 NIFTYJR NIFTYVOL

1 1 1.959 1.000 .02 .02 2 .041 6.912 .98 .98

2 1 2.894 1.000 .00 .00 .00 2 .100 5.380 .00 .24 .08 3 .006 22.885 1.00 .76 .92

3 1 3.836 1.000 .00 .00 .00 .00 2 .152 5.027 .00 .09 .00 .00 3 .012 18.054 .11 .32 .00 .04 4 .000 94.962 .89 .58 1.00 .96

4 1 4.814 1.000 .00 .00 .00 .00 .00 2 .156 5.560 .00 .08 .00 .00 .00 3 .022 14.629 .00 .28 .00 .01 .40 4 .007 26.339 .16 .07 .01 .03 .58 5 .000 107.455 .84 .57 .99 .95 .02

a. Dependent Variable: VIX Residuals Statistics

Minimum Maximum Mean Std. Deviation N Predicted Value 24.7768 79.5547 37.5897 10.38663 268 Residual -20.76211 21.75052 .00000 3.92274 268 Std. Predicted Value

-1.234 4.040 .000 1.000 268

Std. Residual -5.253 5.503 .000 .992 268 a. Dependent Variable: VIX

Statistical Result:

A R Square of 0.873 indicates the degree to which volatility is explained by the independent variables, with the incremental part of it being highest for auto correlated variable, the volatility index. Durbin-Watson Statistic of 2.229 indicates a low auto correlation, which is desirable.

The critical F value of 460.964 indicates the strength of the strength of our regression model. The p value is also extremely low.

We get a regression model from the data:

Yt = 26.452 + 0.794Xt-1 + .007X2 + .002X3 + 0X4 + e

where Y = Volatility Index

i = regression coefficients

Xt-1= Volatility Index

X2 = Nifty 1 Month Futures

X3 = Nifty Junior

X4 = Nifty Volumes

e = random error term

The p value is very low for b2, hence indicating the significant contribution of Nifty futures towards predicting Volatility,

The predicted volatility value for VIX will lie in the range 24.7768 to 79.5547 with a mean of 37.5897 with a standard deviation of 10.39.

Conclusion:

With the regression model we can see that the beta of futures i.e. b2 is -0.007. This beta value is significant. This implies that futures contribute towards stabilizing the market.

Final conclusion:

We have found conclusions from both Model 1 and 2 against the theory that Futures contribute towards the volatility in stock market. While Model 1 suggests that Futures do not contribute to the variances in stock market, the second model suggests that futures contribute towards stabilizing the market. Hence derivatives contribute towards stabilizing stock market.

Reference:

Impact of futures Trading activity on stock price volatility Satya Swaroop Debasish ICFAI Journal of Derivatives Market.

Impact of futures introduction on underlying index volatility Evidence from india- Kotha Kiran Kumar & Chiranjit Mukhopadhyay

Estimation of GARCH models based on open, close, high, and low prices-Peter M. Lildholdt

Basic Econometrics- Gujatati

Impact Of Futures And Options On The Underlying Market Volatility: An Empirical Study On S & P Cnx Nifty Index - Mr. Sibani Prasad Sarangi1 & Dr. K. Uma Shankar Patnaik 2