The study uses three volatility models of the GARCH family to examine the volatility behavior and in particular volatility persistence or long memory of the return series of four Bombay Stock Exchange (BSE) sectoral indices. The study uses the daily data from January 1, 2002 to December 31, 2013. The results of the standard GARCH model suggest the presence of volatility persistence in the return series of all four indices. The EGARCH results suggest that the leverage effect is present and significant for BSE IT and BSE Bankex only, implying that BSE Metal and BSE PSU can be good investment in anticipation of bad times. The CGARCH estimates indicate that the short-run volatility component is weaker. However, the permanent components of the conditional variance exhibit a high degree of persistence for all return series.

Volatility, a basic and important measurement of risk, is defined as the spread of all likely outcome of an uncertain variable. From risk management, optimal portfolio selection to derivative valuation, volatility plays a central role. Therefore, understanding and forecasting volatility is an active and challenging area of research. Volatility in the price of stock can arise because of several reasons like political scenario, company profits, product demand, budget, general business conditions, etc. Also volatility is closely related to returns. However, too much volatility is considered as a symptom of inefficient stock market, as it affects investment spending and economic growth through the various channels. It can be a major hindrance for attracting investment in small developing economies (Mittal and Goyal, 2012).

Traditionally volatility is measured using a constant one-period variance. But economic time series are found to exhibit periods of large unusual volatility, followed by relative serenity (Engle, 1982). In such circumstance, the assumption of constant variance (homoscedasticity) is inappropriate (Nelson, 1991; and Mittal and Goyal, 2012). Several linear and nonlinear models have been developed to capture this volatility clustering or heteroscedastic behavior of economic (financial) time series. The Autoregressive Conditional Heteroscedasticity (ARCH) and Generalized Autoregressive Conditional Heteroscedasticity (GARCH) models developed by Engle (1982) and Bollerslev (1986) respectively are considered to be basic models to capture volatility clustering behavior for a wide range of financial time series.

Financial Risk Management Journal, Modeling, GARCH, EGARCH, CGARCH, Bombay Stock Exchange (BSE), Conditional Heteroscedasticity , Leverage Effect, BSE IT, Exponential GARCH, BSE Bankex, BSE Sectoral Indices.