Title: Real-time forecasts of Henry Hub natural gas prices
Authors: Arthur Thomas - IFP Energies nouvelles-University of Nantes (France) [presenting]
Benoit Sevi - Nantes Universite (France)
Abstract: Using some hand-collected data from the monthly energy review dating back to 1997, a monthly real-time dataset is constructed to generate forecasts of natural gas prices as established at the Henry Hub. We compare the performance of a variety of models including state-of-the-art Bayesian and time-varying-parameters (TVP) models. Considering extensions to possible structural breaks and non-Gaussian distributions along with volatility models, we provide evidence that Bayesian and TVP models help in forecasting the real price of natural gas for horizons up to one year.