The Volatility of Pakistan Stock Market : A Comparison of GARCH Type Models With Five Distribution

This study conducts empirical analyses modeling the volatility of Pakistani stock market over the period of 1st January 2008 to 30th June 2018 via different GARCH type Model; Symmetric (GARCH & GARCH-M) and Asymmetric (EGARCH & TGARCH) with five different Distribution Techniques such as Normal Distribution (Norm), Student's t Distribution (Std.), Generalized Error Distribution (GED), Student's t Distribution with fix the degree of freedom (Std. with fix DOF) and Generalized Error Distribution with fix parameters (GED with fix parameters). The results are shown in GARCH (1, 1) lagged conditional variance and squared disturbance which effects conditional variance is significant in all distribution. GARCH-M (1, 1) depicts a positive significant at 1% results in Std. and GED which indicates the existence of risk premium and insignificant in rest of the distribution on. EGARCH and TGARCH both are found to leverage effect significant at 1% level. In determining the accuracy and adequacy of forecasting density and choice of volatility model the results on simulated data indicates choice of conditional distribution appear as a more dominant factor. EGARCH model with Student's t the distribution technique is delivered satisfactory results as compare to other models which censored by statistical tools of maximum Log Likelihood, minimum AIC, and SIC. The previous study of Pakistani Stock Market is limited to GARCH family models with one or two distributions. This study covers the limitations and also contributes existing literature in this regard. This research is considered important for investors, policymakers, and researchers