Korean Energy Economic Review ISSN:1599-7057(Print) 에너지경제연구

This study is intended to develop various types of forecasting models and to compare their predictability using the forward prices of bituminous coals. For this purpose, the conventional four criteria are adopted and the test methodology by AGS(1980) is used to check the statistical significances among the forecast errors of various forecasting models. The forecasting models used include various types of seven time-series models including the simple OLS models. The empirical results are summarized as follows. The forecasting errors turn out to be different by forecasting period and model. On average, the forecast deviations in the first half of out-of-sample period are larger than those of the second half. This implies that the deviations of forecasts decreases as they move to the second half of out-of-sample period. On the other hand, the deviations of forecasts increase as forecasting periods are extended. In spite of forecasting results, it is hard to make a conclusion that one forecasting model consistently outperforms the other in terms of predictability. Consequently, the forward prices of bituminous coals would provide valuable information about future demand and supply conditions and prices.