Electricity Price Forecast: a Comparison of Different Models to Evaluate the Single National Price in the Italian Energy Exchange Market

Abstract

In the last decades, electricity markets throughout the Eurozone have undergone substantial changes. The deregulation of electricity markets stimulated investments in the production and distribution of energy, but there are large risks associated with these investments due to price volatility. The paper in the introduction describes the algorithm that governs the operation of the Day-Ahead Market in the Italian Power Exchange and proposes an econometric model for short-term forecasting (six months or a year) of the daily Single National Price (Prezzo Unico Nazionale, PUN) of electricity. The model includes constants, regressors, moving averages, weekly and seasonal dummies, autoregressive and heteroschedastic variables. The results show a significant decrease in error of the short-term forecast of the analyzed time series, in comparison with the method of linear least squares, traditionally used in literature. An analysis on the influence of different variables on PUN such as brent, solar radiation and weather has been reported. A comparison of the different models with specific indices have been performed and discussed. Keywords: Electricity prices; Day-Ahead Market; Italian Power Exchange; ARMA–GARCH model; Forecasting. JEL Classifications: C5; C51; L; L1; L11