Blazsek, SzabolcsEscribano, ÁlvaroLicht, AdrianUniversidad Carlos III de Madrid. Departamento de Economía2019-10-152019-10-152019-102340-5031https://hdl.handle.net/10016/29030In this paper, new Seasonal-QVAR (quasi-vector autoregressive) and Markov switching (MS) Seasonal-QVAR (MS-Seasonal-QVAR) models are introduced. Seasonal-QVAR is an outlierrobust score-driven state space model, which is an alternative to classical multivariate Gaussian models (e.g. basic structural model; Seasonal-VARMA). Conditions of the maximum likelihood estimator and impulse response functions are shown. Dynamic relationships between world crude oil production and US industrial production are studied for the period of 1973 to 2019. Statistical performances of alternative models are analyzed. MS-Seasonal-QVAR identies structural changes and extreme observations in the dataset. MS-Seasonal-QVAR is superior to Seasonal-QVAR and, and both are superior to Gaussian alternatives.engAtribución-NoComercial-SinDerivadas 3.0 EspañaWorld Crude Oil ProductionUnited States Industrial ProductionDynamic ConditionalScore ModelsScore-Driven Multivariate Stochastic Location And Stochastic Seasonality ModelsMarkov Regime-Switching Modelsworking paperC32C51C52open accessDT/0000001731