Emperical Comparative Study of GARCH Family Models for Stock Volatility Forecasting in India’s Pharma and IT Sectors

NA

Abstract

In this paper, we analyse the stock return volatility in the Pharma and IT sectors of India by employing GARCH family models ARCH, GARCH(1,1), TARCH(1,1), EGARCH(1,1), and PGARCH(1,1) to the daily stock returns from 2014 to 2025. The analysis aims to find the best model that accounts for volatility clustering and asymmetries within these sectors. The analysis demonstrates that while all models incorporated volatility clustering to some degree, asymmetrical models such as EGARCH(1,1) and PGARCH(1,1) did markedly better. In the IT sector, EGARCH(1,1) also outperformed all other models, achieving very significant results for INFOSYS (z = -10.47, p = 0.000) and TCS (z = -5.57, p = 0.000), and having lower AIC values (-4.803 for INFOSYS) and Log-Likelihood (6656.31 for TCS) scores. In the Pharma Sector, PGARCH(1,1) did outperform all models, CIPLA's Log-Likelihood of 7478.81 and AIC of -6.031. The results show that for sector-specific volatility forecasting, GARCH models proved more useful and provide better insight for investors and portfolio risk managers as well as supported strategic business decisions.