An estimation of Value at Risk using GARCH models: an application to the Argentine stock market

Abstract

The Value at Risk (VaR) represents the maximum probable loss that an asset may experience in a given time horizon and with a given confidence level. This paper attempts to estimate the most appropriate model to measure the risk of the Argentinean stock market, using the daily series of the S&P Merval index. For this purpose, a parametric VaR model was proposed through GARCH (1,1), GJR-GARCH(1,1) and E-GARCH(1,1) conditional variances together with normal, student's t and skewed student's t distributions. Through backtesting, the suitability of each model was determined. Finally, the most appropriate model for risk management of the Argentine stock market is the parametric VaR with an E-GARCH (1,1) model with student's t distribution.