Abstract: The role of electric power exchanges is to offer platforms of meeting sell and buy orders for delivering fixed amount of electricity in certain area during a given time interval. In such exchanges, the “spot electricity” is considered as a basic product, whose orders are executed one day (or several hours) before the delivery with a “spot price” determined in the market. However, since the spot electricity price fluctuates and is uncertain until several hours before the delivery period, there is a strong need to introduce other products called “forwards,” which are traded in advance, e.g., 1 month before the delivery, between sellers and buyers by fixing electricity prices with the so-called “forward prices.” The objective of this research is to develop a methodology to derive a theoretical forward price based on spot price prediction and risk-adjustment techniques. Although the relation between the spot and the forward prices is actually similar to the one in financial exchange markets such as a stock price and its future price (the terminology of futures is usually used in financial markets instead of forwards in electricity markets), an important difference is that electricity cannot be stored and has to be consumed as soon as it is produced. This fact makes the forward pricing problem difficult for electricity market, where the future price may be created using the current stock price almost directly in the case of stock market.

Our approach is based on the intersection between control and financial engineering applied to forward pricing problem. At first, we model the spot price dynamics using discrete time state space equations formula, where we decompose the log-price process of electricity spot price into trends and residual, where the trend part is modeled by the generalized additive model (GAM) and the residual by a state space representation. Then, we set the forward price using a predicted value of spot price, where we apply one of probability measure transformations used in the area of financial engineering and derive risk-adjusted forward price with arbitrarily long horizon and delivery period. Finally, we apply our proposed technique for Japanese electricity market using empirical data of the Japan Electric Power Exchange (JEPX), where we illustrate the accuracy of prediction and the behavior of time-series data implied by observed forward price information in the JEPX.

Biosketch: Yuji Yamada is a professor in the University of Tsukuba, Japan, and starting from April 2018, he serves as a Dean of the Faculty of Business Sciences, Tokyo-Campus. He holds an undergraduate degree in Engineering from University of Chiba, a Master’s degree and PhD in Engineering from Tokyo Institute of Technology. Before joining University of Tsukuba, he served as a postdoctoral scholar at Control and Dynamical Systems, Caltech. At University of Tsukuba, he teaches courses in Mathematical Finance and Financial Engineering. His current research interests are at the intersection between finance, optimization, and control theory.