Autori
Petrella, LeaLaporta, Alessandro G.Merlo, LucaTitolo
Model Selection for Energy CommoditiesPeriodico
Università degli Studi di Roma "La Sapienza" - Dipartimento di metodi e modelli per l'economia il territorio e la finanza. Working papersAnno:
2018 - Fascicolo:
152 - Pagina iniziale:
1 - Pagina finale:
23In this paper we perform a model selection procedure from a Value at Risk forecasting
point of view for the major energy commodities traded in the markets. We consider several
model specications including GARCH, Generalized Autoregressive Score (GAS) and Con-
ditional Autoregressive Value at Risk (CAViaR) ones. We also propose a Dynamic Quantile
Regression (DQR) framework where the parameters evolve over time following a rst order
stochastic process. The VaR forecasting performances are assessed by using the Model Con-
dence Set procedure which provides a superior set of models by testing the null hypothesis
of Equal Predictive Ability. Subsequently the estimates yielded by each model are pooled
together with a weighted average approach. Our results show that the quantile models i.e.
the CAViaR and the DQR outperform all the others for all the commodities. Moreover, the
VaR aggregation generally produces better results, especially for high level of condence.
Esportazione dati in Refworks (solo per utenti abilitati)
Record salvabile in Zotero