Valor en riesgo en el sector petrolero: Un análisis de la eficiencia en la medición del riesgo de la distribución α-estable versus las distribuciones t-Student generalizada asimétrica y normal

In the oil sector, value at risk (VaR) can be used to quantify as best as possible the maximum oil price changes, because these have an impact on economic activity and finds evidence of its importance in explaining movements in the stock returns (Sadorsky, 1999). With this purpose, in this paper we quantify the VaR of three types of oil (Brent, WTI and MME) and analyze the performance of the one-day VaR estimation by Kupiec test considering GARCH models with three alternative distributions in the innovation process: stable, Student-t generalized and normal in a period of high volatility. The results of the performance evaluation of the model based on the Kupiec statistic indicate that the VaR-stable model is a more robust and accurate model for both confidence levels than those based on the generalized asymmetric and normalized Student t-distributions. This result is crucial in the financial sector, because it directly impacts the provision of reserves necessary to face potential losses.