本文以波羅地海乾散貨運價指數為分析對象，樣本時間為1999/11/1 至2010/2/28 止，探討價格變幅波動模型的配適績效與預測績效，並且與以收盤價資料為基礎的GARCH 模型相比較。實證結果發現，無論樣本內的配適績效或樣本外的預測績效，價格變幅波動模型皆優於GARCH 模型。此外，在模型中加入與BDI 指數相關的外生變數有助於提升運價指數短期預測績效，但是在長期預測績效方面並未明顯改善。

This study explores the fitting performance and forecasting performance of therange-based price volatility model on the Baltic Dry Index (BDI). The sample periodis from 1999/11/1 to 2010/2/28, and the performance is compared with traditionalreturns-based GARCH model. The empirical results show that range-based model isbetter than GARCH model, in terms of the in-sample fitting performance andout-sample forecasting performance. Besides, the study also adds some adequateexogenous variables into the model to increase its forecasting performance. Theresults demonstrate that exogenous variables could help short-term forecast, but not inthe long-term forecast.