Resumen
Petroleum is an important energy commodity, being used in different activities, having a direct or indirect effect on several sectors in the economy. This commodity has unstable prices, as a result of geopolitical shocks as well as market shocks in the perspective of technological innovation in the area of energy and changing consumption patterns. In this work we study the volatility of the main reference oil prices with three models: GARCH; GARCH with regime change (MS-GARCH); and unconditional variance model with regime change (MSIH). The models are compared in terms of predictive performance and value-at-risk outside the estimation sample. We can identify different regimes on oil prices. The models with Markovian Switching are the best models using predictive performance and also the value at risk performance metric.