Financial Integration of global markets has influenced volatility of stock market of individual countries, which has evinced much interest in identification of clusters of extreme values of financial returns series of specific stock indices. The estimation of extremal index, commonly interpreted as the reciprocal of the mean number of exceedances in a cluster, extends a key role in analyzing the observed volatile behavior of the stock indices. Such analysis was earlier done by estimation of clustered extreme values by a class of processes like GARCH. This paper has applied extremal index approach and compares it with traditional approaches, using simulation from a GARCH process. It studies the financial returns series of NSE CNX Nifty, the leading stock Index of National Stock Exchange of India and assesses empirically the relative performance of the estimators of different methods for identification of clustering of extreme values of NSE CNX Nifty returns series. lt is found that the two threshold method performs better than run estimator method in low level of threshold.

Financial integration of global economies has given rise to excessive volatility in the financial markets causing systematic risk. Generally, the market shocks give rise to a few very big or a few very small movement in the financial returns persisting for a short time and such trait is called as volatility clustering. The extremal index extends Extreme Value Theory to stationary processes and it gauges the level of dependence in the largest observations defined by a threshold sequence {un}. There exists hardly any study on the extremal index approach in analyzing the stock market of India. This paper proposes to contribute to this arena and explores empirically different methods in estimation of extremal index on NSE CNX Nifty, a leading Stock Index of National Stock Exchange of India.

Peculiarity of dependence pattern of financial returns series has evoked numerous empirical research interests which have contributed greatly to the literature of empirical research on volatility of financial markets. To name two outstanding studies are Autoregressive Conditionally Heteroscedastic (ARCH) models introduced by Engle (1982) and generalized ARCH (GARCH) models proposed by Bollerslev (1986).

Applied Finance Journal, Extremal Index, Financial Integration, Global Markets, Stock Market, National Stock Exchange, GARCH Process, Global Economies, Financial Markets, Monte Carlo Simulation, Risk Management, Financial Applications, Autoregressive Conditionally Heteroscedastic Models, ARCH.