Using Quantile Regression for Modeling of Electricity Price and Demand
MetadataShow full item record
Article 1: Residual demand, the difference between demand and renewable production, is important variable in predicting the future price and the future need for energy storage for intermittent renewables production. The residual demand represents the load that can not be met by renewable production and must be served by conventional power plant, electricity imports or storage capacity. However, little is known about predicting the residual demand itself as well as its quantiles. We therefore model demand and residual demand using ordinary and linear quantile regression, and thereafter compare the results for the hourly electricity consumption in Germany. We find that that the residual demand is less predictable than demand. Our paper makes two contributions to the literature: (1) unlike other studies it analyses the residual demand by using quantile regression (2) it compares the results of demand and residual demand. Article 2: This paper analysis the relation between several fundamental variables and German day-ahead electricity price for each hour. The study performed quantile regression on the electricity prices and reveals important effects that are missed by ordinary regression. Ordinary regression would assume that the relation to be the same for high and normal electricity prices on a specific hours. While the quantile regression measures the dependence of the extreme event. Examine these extreme event on the price is an important aspect of effective risk management. The results indicate that the effect from the factors on electricity price vary substantially across the quantiles, thus confirming the high complexity of the electricity price.