Filtering And Forecasting Spot Electricity Prices In The Australian Electricity Market

Modelling and forecasting the volatile spot pricing process for electricity presents a number of challenges. For increasingly deregulated electricity markets, like that in the Australian state of New South Wales, there is a need to forecast the price for a range of derivative securities used for hedging. Any forecasting model that hopes to capture the pricing dynamics within this market must be able to cope with the extreme volatility of the observed spot prices. By applying wavelet analysis, we examine both the price and demand series at different time locations and levels of resolution to reveal and differentiate what is signal and what is noise. Further, we cleanse the data of leakage from high frequency, mean reverting price spikes into the more fundamental levels of frequency resolution. As it is from these levels that we base the reconstruction of our filtered series, we need to ensure they are least contaminated by noise. Using the filtered data, we explore time series models as possible candidates for explaining the pricing process and evaluate their forecasting ability. These models include one from the threshold autoregressive (TAR) class as well as the benchmark linear autoregressive (AR) model. What we find is that models from the TAR class produce m step-ahead forecasts that best appear to capture the mean and variance components of the actual data.