The purpose of this paper is to develop a model for energy demand forecasting in Taiwan. Such a forecast can serve as (1) the foundation of energy supply planning; (2) the base for calculating strategic energy reserves; and (3) the core of national energy demand management policy and regulation.
Time-series models are selected as the main forecasting methods because they provide better forecast performance than other traditional models, such as econometric models. Another benefit of using time-series models is their relatively low computing cost. The Bayesian VAR model is an appropriate choice for the long-term energy demand forecasting model. To observe the structural change in the economy, GDP has been put into the VAR model as an endogenous variable. The only limitation of the time-series models is their need for a long series of data to generate meaningful results. The information on energy demand forecasting provided by this study will be useful in overall economic planning for Taiwan.
