Modeling and Forecasting the Demand for Electricity in New Zealand: A Comparison of Alternative Approaches

Models of energy demand in New Zealand have typically been based upon either a partial general equilibrium approach or constructed from spreadsheet models. The results created by such methods predict that electricity is forecast to be the fastest growing energy demanded by households and the industrial sector for the next two decades. Furthermore, aggregate electricity demand is forecast to grow at a constant rate for the next two decades. In this paper we attempt to model and forecast electricity demand using a number of recent econometric approaches including Engle-Granger's Error Correction Model, Phillip and Hansen's (1990) Fully Modified Least Squares, and the AutoRegressive Distributed Lag (ARDL) approach of Pesaran et al. (1996, 1998). We identify the model with the smallest forecasting error using a series of forecasting measures and conclude that the new ARDL approach of Pesaran et al., has better forecasting performance than the other approaches considered.