Forecasting Volatility and Value-at-Risk of Pakistan Stock Market with Markov Regime-Switching GARCH Models

Abstract

This paper compares and evaluates various generalized autoregressive conditional heteroscedastic (GARCH) models in terms of their ability to forecast volatility and risk of Karachi Stock Exchange (KSE). Linear and nonlinear GARCH and Markov Regime-Switching GARCH (MRS-GARCH) models with normal and fat tails errors are employed to predict 1-day to 1-month ahead forecast of volatility and Value-at-Risk (VaR). The MRS-GARCH model is estimated with two regimes, representing periods of low and high volatility of stock returns. The MRS-GARCH- model for short horizon and EGARCH- model for long horizon are found to predict the volatility and risk of KSE better than the competing models.