Structured renewable energy derivatives

Giovanni Masala and Amelie Schischke

Abstract

The economic viability of renewable plants increasingly hinges on the ability to hedge simultaneous volume and price risk. This paper develops a framework that forecasts climatic variables and electricity prices. The framework transforms these forecasts into energy production and income scenarios and then prices structured derivatives under a risk-neutral procedure anchored to market forwards. Observations at a given location are modeled using both classical models and advanced machine learning techniques. Rolling backtests show that neural networks are among the top-performing models across the underlying targets and generate scenario distributions with improved downside-risk profiles relative to classical baselines; a scenario-implied minimum-variance hedge based on the average-price forward reduces the variance of the hedged profit and loss by about 70%; and the relative contribution of price-versus-volume sensitivities varies with the horizon, highlighting how dominant risk drivers shift with temporal aggregation. These results provide a bridge from data-driven forecasting to pricing and risk mitigation for hybrid wind–solar plants. This study introduces three original features: a joint forecasting and evaluation framework applied to multiple models over the same drivers, enabling like-for-like model risk comparisons; a forward-anchored scenario engine that propagates forecast uncertainty into joint Monte Carlo paths suitable for risk-neutral pricing and bivariate Greeks; and an empirical hedging module, providing an operational and risk-focused measure of hedge effectiveness.