A1823
Title: Optimizing wind energy aggregation: a comparative analysis of asset allocation techniques
Authors: Alexios-Ioannis Moukas - Aristotle University of Thessaloniki (Greece)
Alla Petukhina - HTW Berlin (Germany)
Daniel Traian Pele - Bucharest University of Economic Studies, Institute for Economic Forecasting, Romanian Academy (Romania)
Nikolaos Thomaidis - University of the Aegean (Greece)
Vlad Bolovaneanu - Bucharest Academy of Economic Studies (Romania) [presenting]
Abstract: It is well-known that mean-variance efficient portfolios tend to be very sensitive to changes in the mean and covariance estimators and the reliability of sample estimators tends to deteriorate when returns deviate from the elliptical distribution prototype. Alternatives, such as equal risk contribution and hierarchical risk parity, try to focus on uncorrelated assets and better control volatility. Recently, a study proposed a new minimum variance asset allocation model that is robust to heavy-tailed data. Based on experience from the application of these techniques in financial markets, an alternative testbed is explored for asset allocation strategies, the optimal aggregation of wind energy resources. Wind farm production is highly volatile and often characterized by a low reward-to-risk ratio. Although some attempts have been made in the literature to apply concepts from portfolio theory to the selection of energy-generating assets, most of the approaches use simple estimators for the main inputs to the optimization problem. This has potentially adverse effects on the stability of asset weights and the out-of-sample performance of the derived portfolios. Using high-resolution wind capacity factor data for Germany, the goal is to apply a battery of techniques not yet explored in this context and quantify the benefits that these approaches bring to wind energy aggregators.
