Aplicación de los modelos Garch a la estimacion del VaR de acciones colombianas

Abstract

The increased volatility in financial markets over recent decades has led researchers, experts, and regulators to design and develop more sophisticated risk management tools. Value at Risk (VaR) has become the standard measure that financial analysts use to measure market risk due to its conceptual simplicity and easy interpretation. In this paper, Value at Risk (VaR) was applied to the returns of the biggest marketability shares of the Colombian stock market and it was calculated by the parametric method with the RiskMetrics approach and the econometric GARCH models. Under the RiskMetrics approach the variance of the series is computed using an IGARCH (1.1) model. For the calculation of VaR with econometric GARCH models, ARIMA methodology is applied to find the model that will help to forecast the returns of the series, this returns does not generally have a constant variance, showing the existence of Heteroscedasticity and should be used generalized autoregressive conditional heteroscedasticity models (GARCH), such as PGARCH, TGARCH, EGARCH, and other models as IGARCH, GARCH-M to find the conditional variance.