Improved Hourly Price Forward Curve Forecasting Based on Future Prices for Hydropower Production

Accurate forecasting of the Swissix hourly price forward curve (HPFC) is key to maximizing the revenues of hydropower production portfolios. Current HPFCs used by the market still have significant improvement potential in terms of accuracy. Overlapping seasonalities like day hours, weekdays, and months are not dynamically reflected. This paper provides a new approach to better reflect overlapping seasonalities in HPFC forecasting and shows the accuracy improvement achieved by adding multiplicative factors to each monthly regression as functions of Phelix monthly and quarterly Base Future prices. Evidence from this research shows that the accuracy of HPFC forecasting can be significantly improved in comparison to the HPFC widely used by utilities, as quantified in statistical and monetary terms. Results from this research confirm that with the more precisely forecasted HPFC, a real Swiss hydropower producer would have achieved significantly higher revenues for different time frames over a 12-month period in 2018.