The Use of Implied Volatility in Conditional Variance Modeling Can Improve Volatility Forecasting and var and cvar Estimation

Abstract

This paper aims to incorporate the s&p/bmv ipc vix index into the variance equation of the garch,
egarch, figarch and fiegarch models to improve conditional volatility forecasts and var and cvar
estimates for the short and long positions on the Mexican Stock Exchange index. The results of the
statistical test for spa show that the s&p/bmv ipc vix index provides additional information for improving
the accuracy of conditional volatility forecasting, but its value is economically small. Backtesting
results reveal that the var-fiegarchvix, var-egarchvix, cvar-figarchvix and cvar-garchvix
measures are optimal for correctly estimating risk at conventional confidence levels for both financial
positions, although the figarch model is clearly superior for short position cvar forecasts. Findings
have important implications for risk management and financial regulation.
Keywords: Implied volatility, volatility forecasting, value at risk
JEL classification: C22, G1 and G15