CMStatistics 2018: Start Registration
View Submission - CMStatistics
Title: A multivariate dependence analysis of electricity prices Authors:  Luca Rossini - Vrije Universiteit Amsterdam (Netherlands) [presenting]
Fabrizio Durante - University of Salento (Italy)
Francesco Ravazzolo - Free University of Bozen-Bolzano (Italy)
Angelica Gianfreda - London Business School (United Kingdom)
Abstract: The purpose is to study the joint interdependence between electricity prices, demand and renewable energy sources (RES). Several papers have tried to understand the underling non-linear relationship between electricity prices and fundamental factors. However, only few have considered the dynamics affecting the entire distribution, and more importantly, they generally focused on a bivariate relationship that is looking at price-demand or price-wind interactions. Therefore, we aim at filling this gap inspecting a multivariate dependence structure between all four variables specifying proper marginal distributions later used in multivariate copula models. By using the well-established AR-GARCH framework and two types of copulae, we provide for the first time evidence of a time-varying multivariate dependence structure. Finally, density forecasting performances across copula models are tested and compared to provide operational insights. The interactions between electricity prices, demand and electricity generated by renewable energy sources (wind and solar photovoltaic) are investigated across the 24 hours from 2011 to 2017 in the German market, which has been characterized by an increasing level of RES penetration. Our dataset consists of hourly prices collected from the European Energy Exchange, EEX; and hourly forecasted demand, wind and solar generation, collected from Thomson Reuters.