Electricity prices are notoriously hard to model due to their exotic behaviors. The extraordinarily high volatilities, strong mean reversion, pronounced cyclical price patterns, and occasional occurrence of price spikes may demand very complicated models. From the empirical analysis with the Dutch and German electricity market data, we find that the two-state regime switching models have a superior performance than a jump diffusion model in terms of replicating the higher moments in the historical data and forecasting electricity prices. The real options method has been used in the valuation of energy assets and the decision-makings in operation and investment in power plants. Our empirical simulation results show that power plant values can be decreased by volumetric risk factors from both the supply side and the demand side. Investment opportunities in power plants can be valued as American options. We value different types of investment options by using both the Hull and White trinomial tree and the Least Squares Monte Carlo method. We confirm the conclusion that the option value to invest in a peak-load power plant is higher than the option to invest in a base-load power plant when the spark spread is lower, and vice versa.
|Award date||5 Oct 2007|
|Place of Publication||Enschede|
|Publication status||Published - 5 Oct 2007|