构建根据现货价格预测的JEPX远期价格模型

Motivated by the full liberalization of the Japanese retail electric power market which started in April 2016, trading of wholesale electricity via the Japan Electric Power Exchange (JEPX) has been attracting much more attention recently. In addition, the Tokyo Commodity Exchange (TOCOM) is now planning to list electricity futures/forward contracts on the JEPX spot prices, and a variety of electricity derivatives using those contracts are expected to be available in the near future. However, the construction of forward/future models is considered difficult due to the lack of sufficient observed data despite its importance in introducing such derivatives. Also, it should be noted that electricity needs to be consumed as soon as it is produced so that future/forward contracts cannot be replicated using the spot electricity contracts, further exacerbating the pricing problem. In this paper, we develop a prediction based forward pricing model by extending our previously developed technique for estimating predicted values of JPEX spot prices. To the end, we decompose the log- process of JEPX spot prices into a trend function modeled by a spline regression and the residual term represented by state space equations formula. Then, we apply a financial engineering technique known as the Esscher transformation for the daily forward pricing problem to derive forward prices with arbitrary delivery periods (e.g., from one week to six months). An empirical analysis illustrates the accuracy of predicted forward prices estimated by our proposed technique and the risk aversion coefficient implied by realized forward prices in JEPX.