The Impact of Derivatives on Spot Market Volatility: A Study on S&P CNX Nifty, India

Transcription

1 31 The Impact of Derivatives on Spot Market Volatility: A Study on S&P CNX Nifty, India Rajni Sinha, Assistant Professor, Amity Business School, Amity University, Rajasthan ABSTRACT The research study analyses the impact of derivatives on Indian spot market. The data used are daily log return of S&P CNX Nifty from Jan 1991 to Dec GARCH (1, 1) model is employed to examine the impact of information flow following the introduction of derivatives (index futures, stock futures and index options) on the conditional volatility of S&P CNX Nifty. Augmented Dickey Fuller (ADF) test is conducted to ascertain whether the return series is stationary. Ljung-Box Q test is performed to test whether any group of autocorrelation in the time-series are different from zero. The research analysis provides evidence that after the introduction of derivatives, the conditional volatility of S&P CNX Nifty has increased. ARCH effect (recent information) is more prominent in index futures than in stock futures and index options, in defining the overall conditional volatility of S&P CNX Nifty. The impact of existing information (old news) carries more weight with respect to all three derivatives (index futures, stock futures and index options). Phenomenon of volatility clustering has increased over time after the introduction of derivatives. Conditional volatility of S&P 500, conditional volatility of Bank Nifty (bank index of India) and daily log returns of S&P Banking Index (BIX, banking index of USA) and Bank Nifty (bank index of India) are used to find the impact of other factors on the conditional volatility of the Indian spot market. Further analysis is done to examine the effect of S&P Banking Index (BIX) returns and Bank Nifty on the conditional volatility of S&P CNX Nifty in different time periods; based on sub-prime mortgage crisis (pre-subprime crisis period, sub-prime crisis period and post- subprime crisis period). Vector Autoregressive analysis and Granger Causality Wald tests are employed to verify the interdependencies of daily log returns of banking indices (both BIX and BANKNifty) time-series and conditional volatility of S&P CNX Nifty. The research analysis confirms that the conditional volatility of S&P CNX Nifty caused by the conditional volatility of S&P 500 has marginally increased and GARCH effect (existing information) dominates over ARCH (recent information) in the post derivative period. During the period of Pre- Sub-Prime crisis, the combined effect of BANKNifty and BIX returns influences the conditional volatility of S&P CNX Nifty and follows the same pattern as that by the S&P 500. However, during the period of Sub-Prime crisis, the persistent information effect on market volatility completely disappears and spot market volatility relies more on the recent information. Whilst during the post Sub-Prime crisis, the ARCH effect dominates. The Granger Causality Wald test proves the conditional volatility of S&P CNX Nifty affects BANKNifty rather than BANKNifty affecting conditional volatility of S&P CNX Nifty. The key implication of the research finding is that the derivatives in Indian market have failed to alleviate risk that was much required during the time of crisis. The integration of global economy has led to the migration of risk to Indian spot market. The collapse of banking system in United States followed by global liquidity crunch seems to have increased the level of volatility in the Indian market. SEBI s (Securities and Exchange Board of India) measures to control volatility have failed during the recent times and better regulations are required to reduce the volatility in Indian spot market using better surveillance and monitoring mechanisms and curbing out price anomalies of derivatives. 1. INTRODUCTION Economic liberalization and the integration of the world economy embarked the emergence of financial engineering and risk management. The key outcome of the research and findings by the financial pundits, particularly in developed economies are Derivatives. Derivatives are the financial instruments whose values are derived from the underlying assets a currency, an interest rate, a commodity, or a stock. Financial derivatives play a key role to overcome the inertia of managing risk in crossborder transactions, in trading securities and commodities, and exposure to uncertainty arising because of interest rates volatility within boundaries. The derivatives inject the extra liquidity in the economy by attracting both domestic and foreign institutional investors and increasing the volume of trading activities. The derivative trading also attracts speculators, who seek to profit by anticipated increase or decrease in a particular market price. This in turn provides the additional capital needed to facilitate the liquidity. Also, derivatives provide additional.channel to.invest with.lower trading.cost; by facilitating the investors to extend their settlement through future contracts. Derivatives are traded either over the counter by derivative dealers or in organized exchanges. Prior to the world-wide market crash of 2007, OTC derivatives were not

2 32 standardized and proper regulations were not in place. However, after the economic meltdown and the financial catastrophe brought in by the exotic financial derivatives, ISDA (International Swap and Derivative Association) and regulatory boards of individual sovereigns have set new regulations and better stringent laws are enforced to enable more transparency and better estimates of both liquidity and volume of transactions. However, the key question still remains unanswered. Does derivative trading induce greater volatility in the market? Researchers all across globe have studied the impact of derivatives on market volatility in both developed and emerging economies and their findings have been inconclusive over the years. In emerging economy such as India, the impact of derivative trading on the stock market volatility has received considerable attention in the recent times, particularly after the market crash of 2001 because of dot com bubble and world-wide economic meltdown as an outcome of sub-prime crisis in The exchange traded derivative products such as Futures and Options have become important instruments of the price discovery, hedging risk and portfolio diversification in last one decade. One of the key reasons for the emergence and the popularity of the derivatives in the Indian market are FIIs (Foreign Institutional Investors). The global integration and liberalization of Indian economy has led to the surplus inflow of capital in Indian Bourse. During the Sub-Prime mortgage crisis (Sep Feb 2009) and post Sub-Prime crisis, India s growth has been the major attraction for the foreign investors. The majority of this inflow has been in the derivatives market, particularly Nifty50 index future. The relaxed regulation by SEBI (Securities and Exchange Board of India) by lifting the daily upper cap on the transaction amount for FIIs in derivative trading is another reason for the influx of capital in the Indian market. However, the question of whether the extra injection of liquidity, chiefly in Indian derivatives market has affected the overall volatility is still under investigation and hence is the main motivations of this research study. The other motivating factor for this research analysis is the present economic climate; that started with Sub-Prime Mortgage crisis, followed by collapse of banking systems, (particularly in USA and there after migrating to rest of world) and finally credit crunch paralyzing the global economic system. India is an integral part of global economic system, with a sustained growth of over 6.5% during the economic meltdown. No doubt economic turmoil arising because of credit crunch has affected the overall volatility in the Indian spot market but the question is to what extent. How the collapse of banking system in United States has led to the induction of volatility in Indian market and how systemic economic collapse is correlated to Indian derivative market. Has derivatives been successful in reducing the volatility in the Indian spot market or has volatility been more influenced by the overall economic climate. This research study analysis all these above questions in detail. 1.1 History of Derivatives in India The history of derivatives may be new for the developing and emerging economies but it has existed since long time in the developed economies. The history is surprisingly longer. Historically, farmers used derivative instruments to protect themselves from decline in price of their crop either because of over production or delay in rain. In Osaka, Japan, the use of derivatives dates back to 1650, when the first derivative as future contract was introduced in the Yodoya rice market. In India, the commodities derivative market dates back to 19th century with organized trading in cotton through the establishment of Cotton Trade Association in The exchange traded instruments such as futures and options are lately introduced since June 2000 at two major stock exchanges: the National Stock Exchange (NSE) and Bombay Stock Exchange (BSE). Index futures, Index options, Stock futures and Stock options are the four major derivative contracts traded in both these exchanges. 1.2 Role of Financial Derivatives Risk Management/Hedging: The actual reasoning behind the introduction of derivatives is hedging the pre-existing risk, in order to offset the potential losses in the underlying or spot market. In India, the governing body for derivative trading and its regulations is Securities and Exchange Board of India (SEBI), which along with its international counterpart ISDA frames regulations that encourages derivative trading for the purpose of risk management. Speculation: Speculation is the other prime motive of derivative trading (i.e. taking positions to profit from the anticipated price movements). In practice, it may be difficult for the policy makers like SEBI and ISDA to distinguish whether a particular trade was for the purpose of hedging or speculation. The balanced efficient market requires the participation of both hedgers and speculators. However, it is still arguable whether speculation in derivative market destabilizes the spot market and increases the market volatility. Market efficiency: Proponents of derivative trading believe, derivative trading provides information in market and hence leads to more information symmetry. The available information in the market is altered for two reasons: first, additional traders are attracted in the market either to hedge risk or speculate on anticipated market price; second, new information may be transmitted to the

3 33 derivative market more quickly than in spot market because of the low transaction costs compared to the spot market. Thus, derivative markets provide a secondary route for the transmission of information to the spot markets, thereby affecting the spot market volatility. To a large extend, the success of derivatives trading depends upon the choice of products. The derivative products traded in Indian stock exchange are futures and options. Raju and Ghosh (2004) have provided evidence for the consideration of volatility in the Indian stock market as tools of analysis of risk factors. There is certainly a concern of attention for stock price volatility due to the coupling of national markets in currency, commodity and stock with world markets and existence of common institutional investors across globe. There are two components of volatility arising in any international market: The volatility arising due to information based price changes and volatility arising due to speculative trading/ noise trading, i.e., destabilizing volatility. As a concept, volatility is simple and intuitive. L C Gupta Committee Report on Derivatives in December 1997 recommended the introduction of derivatives in the Indian capital market. The report highlighted the phase wise introduction of derivatives, first the stock index futures, followed by index options, stock options and stock futures, once after the market matures. Following the recommendations and pursuing the integration policy, futures on benchmark indices (Sensex 30 and Nifty 50) were introduced in June The policy was followed by introduction of index options on indices in June 2001, followed by options on individual stocks in July Stock futures on individual stocks were introduced in November, Classification of Derivatives traded in Indian Market By definition, derivatives are the financial instruments whose value depends upon the underlying assets, which could be stock prices/market indices, interest rates, etc. Derivatives products are specialized contracts2 which signify an agreement or an option to buy or sell the underlying asset to extend up to the maturity time in the future at a prearranged price. Futures: A futures contract is an agreement between two parties to buy or sell an asset at a certain time in the future at a certain price. In NSE, the following futures are traded: Index futures on S&P CNX NIFTY, Bank Nifty, CNX IT, Stock futures on certain specified Securities and Interest Rate Futures. All the futures contracts in NSE are cashsettled. Options: An Option is a contract which gives the right, but not an obligation, to buy or sell the underlying at a stated date and at a stated price. While a buyer of an option pays the premium and buys the right to exercise his option, the writer of an option receives the option premium and is obliged to sell/buy the asset if the buyer exercises it on him. In NSE, the following options are traded: Index options on S&P CNX NIFTY, Bank Nifty, CNX IT, and Stock options on certain specified Securities. All the option contracts in NSE are cash-settled. Derivatives have been the mainstream factor in defining the market volatility in the recent years. The theme of this thesis is calibrating the effects of derivatives and how much their presence has affected the volatility in Indian Market. The research study also highlights the other factors that affect the spot market volatility such as S&P 500 returns, Bank Nifty, S&P Banking index (BIX) returns and effects of Sub-Prime mortgage crisis of September 2007 (which lasted till February 2009) and finally makes a valuable comparison of the degree of volatility induced by the effect of introduction of derivatives and other factors. The rest of the research study is organized as follows: First, literature review of already conducted researches and their respective methodologies applied to deduce the effects of derivatives on market volatility. Thereafter the objective of study, hypothesis, methodology, data and the time period for the study are explained. The variables are identified and explained in brief, followed by Model specifications and estimations. The results are interpreted for each model followed by conclusion. 2. LITERATURE REVIEW Derivative trading and its impact on stock market volatility has been the area of interest among many researchers across globe. Numerous theories have evolved over the past two decades highlighting its impact on market volatility in both developed and emerging economies. The research studies attribute to study factors contributing volatility in stock returns. One common factor of all these studies is derivative trading, particularly index futures, and has increased the volume because of lower transactional cost relative to cash market. However the key question still remains unanswered, whether larger participation and increased volume has reduced the market volatility and information asymmetry. 2.1 Empirical research conducted on spot markets across globe Siopis and Lyroudi (2007) conducted experiments to draw the relationship between the volatility in the Greek stock market and the introduction of the future contracts on the FTSE/ASE-20 index. Various volatility forecasting approaches are used such as GARCH and EGARCH models and the GJR model using the data for a sample period of 10 years. During the analysis, the author has broken the sample period into two sub-periods, one period

4 34 before the introduction of futures trading and one after the introduction of futures trading and applied EGARCH(1,1), GARCH(1,1) and the TGARCH(1,1) models for the prefutures period and the post-futures period as well, with and without a dummy variable. The results of this study indicate that the introduction of futures leads to a significant change in the spot market volatility of the FTSE/ASE-20 index. Poshakwale and Pok (2004) examined the impact of futures trading on spot market volatility in Kuala Lumpur Stock exchange. The results obtained shows that the onset of future trading increases the spot market volatility and the flow of information to the spot market. The results also provide evidence that the underlying stocks respond more to the recent news and non-underlying stocks respond more to the old news. The lead-lag and casual relationship between the futures trading activities and the spot market volatility is also examined. GARCH (p,q) process is used in estimating the volatility and the Autoregressive Conditional Heteroscedastic model (ARCH) is used in modelling the volatility of the time series characterized by the time varying conditional variance. VAR results show that the impact of the previous day s futures trading activity on the volatility is positive but short (only a day). This is further confirmed by the Granger s causality test. Jeanneau Serge Marian Micu (2003) explained how the information based speculative transaction establishes a relationship between the volatility and derivative market. This relationship is based on whether the information is private or public and what type of asset is traded. Author claims that arrival of private information always surges the volatility of return and trading volume in both equity and equity related futures and options. Rahman Shafiqur (2001) examined the impact of Dow Jones Industrial Average index futures and options on the conditional volatility of the component stocks. The research study is much in line with the other researchers in this field of study. The conditional volatility of the intraday stock returns is estimated using the GARCH model for both pre and post periods of introduction of derivatives. The results show that the introduction of index futures and options has not produced any structural change in the conditional volatility of component stocks. Jhon, Gleb and Charles (2001) examined the hypothesis asserting the increase in future market trading increases the spot market volatility. The author has used the GARCH model and Schewert Model and results indicate the rejection of hypothesis. The dataset considered is from the UK market (FTSE 500). 2.2 Empirical research conducted on Indian market Gahlot, Datta and Kapil (2010) examined the impact of derivative trading on stock market volatility. The sample data used are closing prices of S&P CNX Nifty as well as closing prices of five derivative stocks and five non derivative stocks from April 1, 2002 to March 31, The study uses GARCH model to capture nature of volatility over time and phenomenon of volatility clustering. The evidences suggest that there is no significant change in the volatility of S&P CNX Nifty. However, results show mixed effect in case of 10 individual stocks. These results assist investors in making investment decision. It also helps to identify the need for regulation. Bandivadekar and Ghosh (2005) from Reserve Bank of India conducted research studies on the daily return volatility of both S&P CNX Nifty and BSE Sensex for a period of Jan Mar 2003 using the GARCH framework. The results of the research study concluded that the introduction of derivatives have reduced the overall volatility in the stock market. Nath (2003) also conducted research on the behavior of volatility in Indian equity market for the pre and post derivatives periods by using the conditional variance for the period of The researcher has model the conditional volatility using different methods such as GARCH(1,1) and considered data sample of 20 randomly picked NIFTY and Junior Nifty stocks as well as benchmark indices. The empirical results of the study provide evidence that for most of the stocks the overall volatility has reduced after the post derivative trading period. However, it is quite primitive to generalize at that stage that the overall respective indices S&P CNX Nifty and Nifty Juniors will follow the same trend with regards to volatility. Premalata (2003) explored the impact of the introduction of the derivative contracts such as Nifty futures and option contracts on the spot market volatility. GARCH (1,1) model is used to capture the heteroskedasticity in returns and data set comprises of the closing price between Oct 1995 and Dec 2002 for CNX Nifty, Nifty Junior and S&P 500 returns. The results obtained indicate that there is no significance impact of introduction of derivatives on the spot market volatility. However, the empirical results also indicate the shift in the nature of the GARCH process after the introduction of derivatives. Raju and Karande (2003) extended the earlier research and obtained the price discovery and the volatility in the context of the Nifty futures at the National Stock Exchange. The author has used the Co-integration and the Generalized Auto Regressive Conditional Heteroskedasticity techniques respectively on data set between Jan 1998 and Oct The empirical results suggest the volatility is reduced in the cash market after the introduction of the futures. Gupta (2002) examined the effects of introduction of the index futures on the stock market volatility by relative valuation technique. The dataset consists of daily price data (high, low, open, close) of both BSE Sensex and S&P CNX Nifty between June 1998 and June Four measures of volatility are used based on open to open price, close to close price, Parkinson s Extreme Value estimator and Garman- Klass measure volatility (GKV). The results show the overall

5 35 volatility of the stock market has decreased after the introduction of the index futures on both the indices. Ali, Rahman, and Zhong (2002) established the contrary view of earlier research and provided evidence that suggest the volatility in spot market has induced volatility in future market. The author has also established casual relationship between the volume of trades in both future markets and spot markets. The author has used EGARCH framework and Granger Causality Test. The majority of the research studies have employed standard ARCH and GARCH models to draw the relationship between impact of introduction of derivatives and the volatility in the spot market. The main limitation of all the research conducted on the volatility of the Indian market following the introduction of derivatives is: none of the research included the effect of Sub-Prime Mortgage crisis of late 2007, the coupling effects of world economy and transmission of risk across globe, the effects of banking indices on the Indian bourse and the comparative degree of influence on the overall volatility, with respect to impact of introduction of derivatives. 3. OBJECTIVE OF STUDY Spot market volatility has been a major area of concern among researchers, market makers and regulatory boards across sovereigns, particularly in today s era of highly integrated international markets. Also the role of financial engineering has reached many folds in defining the market dynamics in the recent years. Derivatives have been contemplated as the brain child of financial engineering and arguably considered as the main influencer of the cash market volatility. Researchers have always been enthusiastic to find the connection between the introduction of derivatives and the volatility in spot market. However, generalized conclusive results have never been derived. 3.1 Hypothesis In the context of the Indian market, the research study postulates the following hypothesis: Null Hypothesis (H0): The introduction of derivatives has not reduced the overall volatility of the S&P CNX Nifty. Alternate Hypothesis (H1): The introduction of derivatives has reduced the overall volatility of the S&P CNX Nifty. Test statistics is formulated to reject the null hypothesis. 3.2Descriptive Statistics Table 1: Tabular description of descriptive statistics: S&P CNX Nifty returns, S&P 500 returns, Bank Nifty returns and BIX returns. Descriptive Statistics S&P CNX Nifty Returns (Jan 1991-Dec 2011) S&P 500 Returns (Jan 1991-Dec 2011) Bank Nifty Returns (Jan 2000-Dec 2011) S&P Banking Index (BIX) Returns (June 2002-Dec 2011) Mean (%) Median (%) Maximum (%) Minimum (%) Std. Dev. (%) Skewness Kurtosis Jacque-Bera Total number of Observations METHODOLOGY 4.1 Dataset S&P CNX Nifty return is used in the analysis as the proxy for the Indian market. Data set comprising of daily closing price is collected for a period of 20 years from Jan 1991 to Dec Daily log return of S&P CNX Nifty index is calculated from the daily closing price. S&P 500 (proxy for US market) returns are calculated for the same period between Jan 1991 and Dec Bank Nifty and BIX index returns are also calculated between 2001 and 2011(after their respective introductions in Indian and US markets). The closing price of all the above indices are collected from data source, Data-stream. S&P 500 is used to find the degree of influence of highly correlated global index on the overall volatility of S&P CNX Nifty. Banking indices Bank Nifty and S&P Banking index BIX are used to find the effect of banking stock

6 36 indices on the conditional volatility of S&P CNX Nifty. Further study and analysis is conducted keeping in view Sub-Prime Mortgage crisis. Table 2: Date of Introduction of Derivative Products Derivative Products Data of Introduction Underlying Index Index Futures June 2000 S&P CNX Nifty Stock Futures Dec 2001 S&P CNX Nifty Index Options June 2001 S&P CNX Nifty 4.2 GARCH Model Generalized Autoregressive Conditional Heteoscedasticity (GARCH) model is employed to estimate the conditional volatility of the S&P CNX Nifty. GARCH model was independently developed by Bollerslev (1986) and Taylor (1986). The main advantage of GARCH model is that it captures the tendency of the volatility clustering in the financial time series data. GARCH therefore establishes connection between information and volatility. In GARCH model the conditional variance at time t is dependent on the past value of the squared error terms and the lagged conditional variance. GARCH(1,1) model is represented as follows: t = α0 + β0 Xt + ut (1) ζt2 = α1 + β1 ut-12+ β2 ζt-12...(2) Where, ζt2 = Conditional variance and ut = Error term. Equation 1 represents conditional mean equation and equation 2 represents conditional variance equation. β1, the coefficient of squared error term represents the recent information coefficient, the higher value qualifies: the recent news in the market has greater impact on the price change and market volatility. β2, the coefficient of lagged variance reflects the impact of old news in the spot market s price change. The higher value of β2 suggests high level of persistence of information effect on volatility. β1+ β2~1 indicates more integration of volatility. The greater integration of volatility qualifies lack of information and higher price inflexibility in the spot market thereby preventing immediate and continuing adjustment of price in response to demand and supply conditions. The unconditional variance is given by α1/(1- β1- β2). The higher value of unconditional variance indicates higher volatility. 4.3 Unit Root Test Augmented Dickey Fuller (ADF) test is conducted to ascertain that the return series is stationary. Non-stationary time series exhibit upward or downward trends over a sustained period of time. Since such trends are often stochastic and not deterministic, regressing a nonstationary time series can lead to the phenomenon of spurious regression. The followings are the consequences of the spurious regression: If two variables are tending over time, the regression of one on the other could have a high R2 even if the two are totally unrelated. The standard assumption of the asymptotic analysis will not be valid. t-ratios will not follow t distribution and hypothesis tests are not valid about the regression parameters. Augmented Dickey Fuller test: Yt-Yt-1=μ + (λ-1)yt-1+ βt + εt (3) T= Trend term; εt=error term Time series exhibit stationarity if λ-1 0 and β=0 4.4 Ljung-Box Q test Ljung-Box Q test is performed to test whether any group of autocorrelation in the time-series are different from zero. This test ascertains the overall randomness based on the number of lags instead of each single lag. H0= Returns and Squared Returns Series are white noise. H1= Returns and Squared Returns Series are not white noise. The test statistic is: n is the sample size, Pk is the sample autocorrelation at lag k, and h is the number of lags being tested. For significance level α, the critical region for rejection of the hypothesis of randomness is where is the α-quantile of the chi-squared distribution with h degrees of freedom. Rejection of the null hypothesis suggests that the squared returns series follow the ARCH type dependencies and hence GARCH model is appropriate for volatility estimation

7 Vector Autoregressive Analysis and Granger Causality Wald test Granger Causality test is performed to determine whether BIX (USA) index and BankNifty(India) index series is useful in forecasting S&P CNX Nifty s conditional volatility. This test is conducted during the Subprime Mortgage Crisis period (Sep 2007-Feb 2009) and post Subprime Mortgage Crisis period (Mar 2009-Dec 2011).The objective of this test is to ascertain the degree of influence of macro-economic factors such as economic meltdown (because of banking sector) on the volatility in Indian spot market. Vector Autoregressive models are developed to capture interdependencies among S&P CNX Nifty s conditional variance (as an effect of banking indices BIX and BANKNifty), BIX returns and BANKNifty series during sub-prime credit crisis and post sub-prime credit crisis. Granger Causality Wald test is further performed on these models. Variable A is said to Granger cause variable B, if the lags of A can improve a forecast for variable B. In a VAR model, under the null hypothesis that variable A does not Granger cause variable B, all the coefficients on the lags of variable A will be zero in the equation for variable B. A Wald test is commonly used to test for Granger causality. Table 3: Tabular Description of Models used in the analysis Model 1 Index Futures as Dummy; Effect of introduction of Index Futures on the conditional volatility of S&P CNX Nifty. Model 2 Conditional volatility of S&P CNX Nifty during pre introduction of Index Futures Model 3 Conditional volatility of S&P CNX Nifty during post introduction of Index Futures Model 4 Stock Futures as Dummy; Effect of introduction of Stock Futures on the conditional volatility of S&P CNX Nifty. Model 5 Conditional volatility of S&P CNX Nifty during pre introduction of Stock Futures Model 6 Conditional volatility of S&P CNX Nifty during post introduction of Stock Futures Model 7 Index Option as Dummy; Effect of introduction of Index Options on the conditional volatility of S&P CNX Nifty. Model 8 Conditional volatility of S&P CNX Nifty during pre introduction of Index Options Model 9 Conditional volatility of S&P CNX Nifty during post introduction of Index Options Model 10 Effect of conditional volatility of S&P 500 on conditional volatility of S&P CNX Nifty. Model 11 Effect of conditional volatility of BANK NIFTY on conditional volatility of S&P CNX Nifty. Model 12 Effect of both BANK NIFTY & BIX returns on conditional volatility of S&P CNX Nifty; post Introduction of Derivatives Model 13 Effect of both BANK NIFTY & BIX returns on conditional volatility of S&P CNX Nifty during Pre Subprime Crisis Period Model 14 Effect of both BANK NIFTY & BIX returns on conditional volatility of S&P CNX Nifty during Subprime Crisis Period Model 15 Effect of both BANK NIFTY & BIX returns on conditional volatility of S&P CNX Nifty during Post Subprime Crisis Period. *In each model, conditional volatility of S&P CNX Nifty is calculated using GARCH(1, 1) analysis. 5. INTERPRETATION OF RESULTS 5.1 Results of Unit Root Test Augmented Dickey Fuller test is conducted for the S&P CNX Nifty as shown in Table 4 for the entire time-series between Jan 1991 and Dec Table 4: Augmented Dickey Fuller Test NIFTYRt - NIFTYRt-1 = μ + (λ 1) NIFTYRt-1+ βt + εt Or Δ NIFTYRt = μ + (λ 1) NIFTYRt-1+ βt + εt Dickey-Fuller test for unit root Number of observations = Interpolated Dickey-Fuller Table 4: Augmented Dickey Fuller test on daily log return of S&P CNX Nifty time-series Test Statistic 1% Critical Value 5% Critical Value 10% Critical Value Z(t) MacKinnon approximate p-value for Z(t) = D.NIFTYR Coefficient Std. Err. T P> t 95% Conf. Interval L1. NIFTYR (λ 1) _trend (T) -2.00e e e e-08 _cons ( μ)

8 38 The coefficient of λ 1= at 0% significance (as p value is 0) and trend term T is almost equal to zero at 1% significance level. This proves that S&P CNX Nifty series is stationary since λ-1 0 and coefficient of trend term is quite negligible to equate to zero. 5.2 Results of Ljung-Box Q test Ljung-Box Q test is performed to test the white noise for each of the models described in Table 3. The obtained results are summarized in Table 5. The results indicate that in the Models: 1 to 13, the return and squared return series of S&P CNX Nifty follows the ARCH type dependencies and hence GARCH model is appropriate for volatility estimation. All the results are estimated at 5% significance level. However Model 14, which estimates the volatility in S&P CNX Nifty resulting by the effect of BankNifty and BIX indices during the Sub-Prime Crisis period, fails to confirm that the return and squared return series of S&P CNX Nifty follows ARCH type dependencies at 5% significance level. But the same result at 10% level of significance is significant (p-value is ). Ljung Box Q test conducted on Model 15, which estimates the volatility of S&P CNX Nifty as an effect of both BankNifty & BIX during Post Subprime Crisis Period, is not significant at even 10% level of significance. In models 14 and 15, Vector Autoregressive model is further employed and Granger Causality test is conducted to obtain the independent dependencies between each of the three time series: BankNifty, BIX returns and conditional volatility of S&P CNX Nifty resulting by the effect of both BankNifty and BIX indices. The summary of results of Ljung Box Q test for all fifteen models is tabulated as below: Models Table 5: Ljung Box Q test Q= n (n+2 ) Σ (1/n-j) ρj2 Portmanteau test for white noise Portmanteau (Q) statistic Prob > chi2(40) Model 1: Index Futures as Dummy * Model 2: Pre Introduction of Index Futures * Model 3: Post Introduction of Index Futures * Model 4: Stock Futures as Dummy * Model 5: Pre Introduction of Stock Futures * Model 6: Post Introduction of Stock Futures * Model 7: Index Option as Dummy * Model 8: Pre Introduction of Index Options * Model 9: Post Introduction of Index Options * Model 10: Effect of conditional Volatility of S&P 500 on the conditional * volatility of S&P CNX Nifty. Model 11: Effect of conditional Volatility of BANK NIFTY on the * conditional volatility of S&P CNX Nifty. Model 12:Effect of both BANK NIFTY & BIX returns on the conditional * volatility of S&P CNX Nifty; Post Introduction of Derivatives Model 13: Effect of both BANK NIFTY & BIX returns on the conditional * volatility of S&P CNX Nifty during Pre Subprime Crisis Period Model 14: Effect of both BANK NIFTY & BIX returns on the conditional volatility of S&P CNX Nifty during Subprime Crisis Period Model15:Effect of both BANK NIFTY & BIX returns on the conditional * volatility of S&P CNX Nifty during Post Subprime Crisis Period *Significant at 5% level. 5.3 Results of GARCH(1,1) estimate for S&P CNX Nifty GARCH (1, 1)estimates for all the fifteen models, tabulated in Table 2 are computed using STATA and the following results are obtained. Model 1: Index Futures as Dummy; Effect of introduction of index futures on the conditional volatility of S&P CNX Nifty 5.4 Results of Vector Autoregressive Analysis and Granger Causality Wald Test Vector Autoregressive models are developed to capture interdependencies among S&P CNX Nifty s conditional variance ζ2 (caused by the combined effect of banking indices BIX and BANKNifty), BIX returns and BANKNifty returns during the sub-prime credit crisis and

9 39 post sub-prime credit crisis. Granger Causality Wald test is further performed on these models. Table 6: Vector Autoregressive Model; Effect of both BANK NIFTY and BIX returns on the conditional volatility of S&P CNX Nifty during Subprime Crisis Period; (04/09/2007 to 27/02/2009) Dependent Independent Independent Lag Lagged Lagged Lagged Variable Variable Variable Coefficient t- Coefficien t- Coefficient of t-statistics of ζ2 statistics t of BIXR statistics NIFTYRBank ζ2 BIXR NIFTYR Bank * BIXR ζ2 NIFTYR Bank NIFTYR Bank *Significance at 5% * * * ζ2 BIXR Table 7: Granger Causality Wald tests; Effect of both BANK NIFTY and BIX returns on the conditional volatility of S&P CNX Nifty during Subprime Crisis Period; (04/09/2007 to 27/02/2009) Equation Excluded chi2 df Prob>chi2 Null Hypothesis ζ2 BIXR Accept ζ2 NIFTYRBank Accept ζ2 All Accept BIXR ζ Accept BIXR NIFTYRBank * Reject BIXR All Reject NIFTYRBank ζ Accept NIFTYRBank BIXR Accept NIFTYRBank All Accept *Significance at 5%; H0 = Variable under Excluded column doesn t Granger Cause the variable under Equation Column. HA = Variable under Excluded column Granger Cause the variable under Equation Column. The results set in Table 6 and Table 7 provides the summary of the vector autoregressive analysis and granger causality wald test during the sub-prime crisis. The BIX return at t =-3 has significant impact on the conditional volatility (ζ2). The positive sign of BIXR at t =-3 suggest conditional volatility (ζ2) rises (falls) following the rise (fall) in BIX returns. Also at t = -4, BANKNifty significantly impact current BIX returns but in inverse relation, whereas at t =-3 and t = -4, the lagged BIX returns impact the current BIX returns. Granger Causality Wald test conducted on the same data set further supports the above finding. Null hypothesis: NIFTYRBank (BANKNifty) does not Granger Cause BIXR (BIX returns) is rejected within 5% level of confidence, whereas both BIXR (BIX returns) and NIFTYRBank (BANKNifty) do not Granger Cause conditional volatility (ζ2) of S&P CNX Nifty is rejected at 10% level of confidence (since p- value is 0.061). These findings suggest BIX Returns and BANKNifty impact conditional volatility of S&P CNX Nifty, BANKNifty affects BIX returns but conditional volatility of S&P CNX Nifty doesn t affect either BANKNifty or BIX Returns at any lag during the subprime crisis period.

10 40 Table 8: Vector Autoregressive Model; Effect of both BANK NIFTY and BIX returns on the conditional volatility of S&P CNX Nifty; Post Subprime Crisis Period; (02/03/2009 to 31/12/2011) Dependent Independent Independent Lag Lagged Lagged Lagged Variable Variable Variable Coefficient t- Coefficient t- Coefficient of t-statistics of ζ2 statistics of BIXR statistics NIFTYRBank ζ2 BIXR NIFTYR Bank BIXR ζ2 NIFTYR Bank NIFTYR Bank *Significance at 5% ζ2 BIXR ** *** ** Table 9: Granger Causality Wald tests; Effect of both BANK NIFTY and BIX returns on the conditional volatility of S&P CNX Nifty; Post Subprime Crisis Period; (02/03/2009 to 31/12/2011) Equation Excluded chi2 df Prob>chi2 Null Hypothesis ζ2 BIXR Accept ζ2 NIFTYRBank Accept ζ2 All Accept BIXR ζ Accept BIXR NIFTYRBank Accept BIXR All Accept NIFTYRBank ζ * Reject NIFTYRBank BIXR * Reject NIFTYRBank All * Reject *Significance at 5%; H0 = Variable under Excluded column doesn t Granger Cause the variable under Equation Column. HA = Variable under Excluded column Granger Cause the variable under Equation Column. The results set in Table 8 and Table 9 provide the summary of the vector autoregressive analysis and granger causality Wald test post sub-prime crisis. The conditional volatility (ζ2) of S&P CNX Nifty at t =-1 has significant impact on the BANKNifty. Also BIX returns at t =-1 significantly affects BANKNifty. The positive sign of conditional volatility (ζ2) of S&P CNX Nifty and BIX returns at t =-1 suggest BANKNifty rises (falls) following the rise (fall) of both BIX returns and conditional volatility (ζ2) of S&P CNX Nifty. Granger Causality Wald test conducted on the same data set further supports the above finding. Null hypothesis: conditional volatility (ζ2) of S&P CNX Nifty does not Granger Cause NIFTYRBank (BANKNifty) is rejected within 5% level of confidence. Again the null hypothesis: BIX Returns (BIXR) does not Granger Cause NIFTYRBank (BANKNifty) is rejected within 5% level of confidence. These findings suggest that both conditional volatility (ζ2) of S&P CNX Nifty and BIX Returns affect BANKNifty at lag 1 during the post sub-prime crisis period. 6. CONCLUSION Previous studies have used the daily returns data of underlying and non underlying stocks in finding the impact of derivatives on Indian spot market. However, this research study has concentrated on examining the impact of introduction of derivatives trading on spot market volatility using index level data. The research study has established the informational effects of derivative trading on the volatility of the Indian spot market using GARCH (1, 1) model. The study also examined the impact of other factors such as conditional

11 41 volatility of S&P 500, conditional volatility of BANKNifty, S&P Banking Index BIX returns, BANKNifty, and sub-prime crisis on the volatility of the Indian spot market. The major findings obtained are summarized below: ARCH effect (recent information) is more prominent in index futures than in stock futures and index options, in defining the overall conditional volatility of S&P CNX Nifty. The impact of existing persistent information (old news) carries more weight with respect to all three market traded derivatives (index futures, stock futures and index options). Although statistically insignificant at 5% level of significance, the positive δ0 (coefficient of Dummy variable in each derivative type) confirms the overall conditional volatility has increased after the introduction of derivatives. Phenomenon of volatility clustering has increased over time after the introduction of market traded derivatives, which shows even after the introduction of derivatives, the Indian market remains in either bull state or bear state for long periods of time compared to other global markets. In case of other factors, the conditional volatility of S&P CNX Nifty has mildly increased by the S&P 500 after the introduction of derivatives in Indian market and GARCH effect (existing information) dominates over ARCH (recent information). In case of BANKNifty and BIX indices returns the conditional volatility of the S&P CNX Nifty follows the same pattern as followed by the effect of S&P 500(GARCH coefficient is high and ARCH coefficient is low), particularly during the period of Pre- Sub-Prime crisis. However during the period of Sub-Prime crisis, the persistent information effect on market volatility completely disappears and spot market volatility relies more on the recent information. Whilst, during the post Sub-Prime crisis; the ARCH effect continues to dominate over GARCH and the conditional volatility of S&P CNX Nifty affects BANKNifty rather than BANKNifty affecting conditional volatility of S&P CNX Nifty. The main implication of the research analysis is that the derivatives in Indian market have failed to alleviate the risk that was much required during the time of crisis. Integration of global markets in 21st century has led the Indian market more susceptible to Sub-Prime crisis of United States and world-wide liquidity crunch following its effect. Above findings will definitely influence the market regulators such as SEBI to bring better reforms in controlling the volatility in Indian market. The main aim of introduction of derivatives in Indian market was to reduce the market volatility and improve market efficiency by means of greater rate of information flow. It is very much difficult to decouple the Indian market from the rest of world and insulate it from crisis arising because of global integration; however there are definite regulatory measures on derivative trading that can control the spot market volatility such as by strengthening the surveillance and monitoring mechanism, implementing daily circuit filters, daily price bands and weekly price caps to curb abnormal price behaviour and volatility of derivatives. REFERENCES [1] Agarwal, Aman. Derivatives: Wave of the Future. Finance India, [2] Angelos Siopis and Katerina Lyroudi. The Effects of the Derivative Trading on the Stock Market Volatility: The Case of Athens Stock Exchange [3] Board Jhon, Sandamann Gleb and Sutcliffe Charles. The Effect of Futures Market Volume on Spot Market Volatility. Journal of business Finance and Accounting, Vol. 28, No. 7&8, October, 2001: [4] Board Jhon, Sandamann Gleb and Sutcliffe Charles. The Effect of Futures Market Volume on Spot Market Volatility. Journal of business Finance and Accounting, Vol. 28, No. 7&8, October, 2001: [5] Bollerslev, Tim. Generalized Autoregressive Conditional Heteroskedasticity. Journal of Econometrics, 1986: [6] Darat Ali, Shafiqur Rahman, and Maosen Zhong. On The Role of Futures Trading in Spot Market Fluctuations: Perpetrator of Volatility or Victim of Regret? The Journal of Financial Research, Vol. XXV, No. 3, 2002: [7] George P. Tsetsekos, Panos Varangis. The Structure of Derivatives Exchanges: Lessons from Developed and Emerging Markets. World Bank Policy Research Working Paper No. 1887, [8] Golaka C Nath. Behaviour of Stock Market Volatility after Derivatives. National Stock Exchange of India, [9] Jeanneau Serge Marian Micu. Volatility and derivatives turnover: a tenuous relationship. BIS Quarterly Review, March, [10] Lyroudi, Angelos Siopis and Katerina. The Effects of the Derivative Trading on the Stock Market Volatility: The Case of Athens Stock Exchange [11] M.T. Raju and Anirban Ghosh. Stock Market Volatility An International Comparison. Securities and Exchange Board of India, [12] Micu, Jeanneau Serge Marian. Jeanneau Serge Marian Micu (2003), "Volatility and derivatives turnover: a tenuous relationship. BIS Quarterly Review, March, [13] O.P.Gupta. Effect of Introduction of Index Futures on Stock Market Volatility: The Indian Evidence. UTI Capital Market Conference Paper, [14] Poshakwale and Pok. The impact of the introduction of futures contracts on the spot market

12 42 volatility: the case of Kuala Lumpur Stock Exchange. Applied Financial Economics, [15] Raju and Karande. Price Discovery and Volatility on NSE Futures Market. Securities and Exchange Board of India, [16] Ruchika Gahlot, Saroj K. Datta, Sheeba Kapil. Impact of Derivative Trading On Stock Market Volatility in India: A Study of S&P CNX Nifty. Eurasian Journal of Business and Economics, [17] Shafiqur, Rahman. The introduction of derivatives on the Dow Jones Industrial Average and their impact on the volatility of component stocks. The Journal of Futures Markets, 2001: Vol. 21, No. 7, July, pg [18] Shenbagaraman Premalata. Do Futures and Options trading increase stock market volatility. NSE NEWS, National Stock Exchange of India, [19] Snehal Bandivadekar and Saurabh Ghosh. Derivatives and Volatility on Indian Stock Markets. Reserve Bank of India (Reserve Bank of India), 2005.