Applying wavelet transform techniques for New Zealand electricity market volatility analysis

This paper presents wavelet-volatility analysis technique to measure the variation of wholesale electricity prices. The dynamic character of electricity price is mean-revering, with weekly variation, seasonal fluctuations, and instant jumps. The analysis decomposes historical price signals using wavelet transform in a set of constitutive series with different level/scale of details of the original price signals. A time-varying volatility model is then constructed based on these detailed signals using the GARCH (1, 1) model. Results from the New Zealand Electricity market (NZEM) for year 2004 are presented to illustrate the technique.