A1163
Title: Forecasting spot price in the UK natural gas market
Authors: Chih-Yueh Huang - University of Kent (United Kingdom) [presenting]
Ekaterini Panopoulou - University of Essex (United Kingdom)
Stella Hadjiantoni - University of Kent, UK (United Kingdom)
Abstract: Natural gas is one of the most widely used energy resources in the UK as it is the main energy commodity for domestic heating and power generation. The pricing of natural gas prices is crucial for generators, power and gas suppliers, manufacturing plants, real estate agency and other downstream users since the movements of gas prices affect their revenue or costs. Different popular forecasting models, such as forecast combination, penalised regression, stepwise regression and principal components regression, are employed to analyse and forecast UK gas spot price. The result shows that more advanced techniques do not improve the forecasting performance based on Mean squared errors (MSE) ratio and success ratio in the out-of-sample periods. Then, we compare the energy procurement costs of three purchasing strategies, namely spot only, futures only and mixing strategy. The empirical evidence shows that comparing the spot forecast and the front-month futures price helps energy users reduce purchasing costs.
