MODELLING AND FORECASTING OF STOCK PRICE VOLATILITY OF SELECTED NIFTY 50 COMPANIES IN INDIA

Abstract

Price volatility, particularly in the stock exchange market, is a crucial topic that both practitioners and theorists are concerned about. This study helps to see how GARCH family models may be used to predict and forecast stock price volatility for chosen NIFTY 50 companies in India. Only 2% of India's population invests in the stock market, and many investors find it difficult to choose their investment stock due to a lack of knowledge about the securities risk and return. In this study two major sectors like financial services and information technology were considered for the financial year from 1/04/2011 to 31/03/2021. The research design used for the study is analytical research design which helps to shape the research problem and the information related to the research were not manipulated. The sampling method used for the study was purposive sampling by which the researcher selected the sample based on his knowledge which is reliable for the study. The database of the selected 14 companies under two major sectors was listed on the NSE and indexed under the NIFTY 50 index. The models of forecasting like GARCH (1,1), asymmetric GARCH models like Exponential GARCH–EGARCH (1,1) and Threshold GARCH–TGARCH (1,1) have been considered in this study. To see the ARCH effect, Heteroskedasticity Test like - The Lagrange Multiplier (LM) test for ARCH were used to see the presence of Heteroskedasticity in residual of the return series. Because if ARCH effect is present, we can use ARCH/GARCH models. The stationarity test like Augmented Dicky-Fuller test were conducted to see whether the return series are stationary. Two forecasting error statistics, Root Mean Square Error and Mean Absolute Error, were employed to assess the performance of these GARCH family models. Overall, the TGARCH (1,1) model outperformed, and it is regarded as the best-fitting model.