近年來探討商品與能源市場價格動態皆是以均數復歸模型(Mean-Reversion, MR)模型為基礎,在過去文獻Schwartz (1997)認為MR模型較能符合商品市場的價格特性。本研究擴充基本均數復歸模型,加入了跳躍擴散因子與季節效用因子,並以WTI原油為研究標的,透過分析 每日報酬、真實波動度與訂價誤差(pricing error)來衡量這些模型的優劣。實證分析係以在NYMEX交易的WTI 原油為實證對象,將樣本期間區分為金融海嘯與非金融海嘯,觀察不同時期價格波動對於stylized facts模型配適能力的影響。結果發現在金融海嘯期間,具有均數復歸與跳躍擴散因子的模型有較佳的配適能力;而在非金融海嘯期間,則以均數復歸、跳躍擴散與季節效用模型表現較好,其結果顯示不同期間的價格特性,需選擇較適合的模型才能獲得最佳的評價。
This study extends Pilipovic’s (1997) and Schwartz’s (1997) mean reversion (MR) dynamic framework which can capture characteristic of commodity prices. We follow this model and develop a variety of commodity stylized-facts pricing model by including the jump diffusion and seasonality. The study discuss which stylized-facts models is the best to describe the prices dynamics by analyzing the daily returns, realized volatility and pricing errors on an empirical investigation of WTI crude oil traded on the New York Mercantile Exchange. For studying the influence of models’ pricing ability on the different prices fluctuation, we separated our data into the period of financial tsunami and the period of non-financial tsunami. Our results show that the inclusion of mean reversion and jump diffusion model reveals the best pricing ability during the period of financial tsunami but the mean reversion, jump diffusion and seasonality pricing model reveals the best pricing ability out period of financial tsunami.