Modelling electricity spot price time Series using coloured noise forces

In this dissertation we develop a mean-reverting stochastic model driven bycoloured noise processes for modelling electricity spot price time series. The deregulation of electricity markets, which were believed to be natural monopolies, has led to the creation of power exchanges where electricity is traded like other commodities. The physical attributes of electricity and behaviour of electricity prices differ from other commodity markets. Electricity spot prices in the emerging power markets experience high volatility, mean-reversion, spikes and seasonal patterns mainly due to the non-storable nature of electricity. Uncontrolled exposure to market price risks can lead to devastating consequences for market participants in the restructured electricity industry. A precise statistical (econometric) model of electricity spot price behaviour is necessary for risk management, pricing of electricity-related options and evaluation of production assets. We therefore formulate and discuss the stochastic approach used to model the spot prices of electricity by using coloured noise forces. Parameter estimation for the model is carried out by the Maximum Likelihood Estimation (MLE) method on a mean-reverting stochastic process. Data used for model calibration were collected from Nord Pool for the period starting from January, 1999 to February, 2009. With the estimated parameters we simulate the model and find that the simulated and real price series have similar trends and cover the same price range. Thus, modelling of electricity spot prices in which the SDE is driven by coloured noise gives a good approximation to real price behaviour and we recommend coloured noise to be used as the driving force in the SDE when modeling the spot prices of electricity.