Empirical Inference

Abstract

The pricing model we propose for the valuation of short-term electricity forwards is based on a factor specification of the spot electricity price and its variance yielding a closed-form solution for the forward curve as given in equation (18.17). This theoretical model can conveniently be transformed into an equivalent empirical state space model whereupon we are able to build a Kaiman filtering algorithm to calibrate our forward pricing model to available empirical data. Prom chapter 3 we know that a state space model consists of two major pillars: the transition and the measurement equation. In the case of our forward pricing model the transition equation captures the dynamics of the state variables. The measurement equation provides us with a link of the spot electricity price to the forward prices of electricity. In the following we first focus on specifying the transition equation by examining the distribution of the state variables.