原油、黃金與美元/英鎊外匯市場關係之研究-機率分配、週期效應與動態條件相關

The Relationships among Crude Oil, Gold and US Dollar/ British Pound Markets: Three Essays on Probability Distributions, Weekday Effects and Dynamic Conditional Correlations

石東立(TUNG-LI SHIH)

指導教授 ： 俞海琴

中原大學/商學院/管理研究所/博士(2008年)

https://doi.org/10.6840/cycu200900640

摘要

原油價格為影響全球經濟之重要因素，油價持續不斷上漲，不僅對原油市場的交易造成影孿，金融市場也受到衝擊。有鑒於高油價可能導致通貨膨脹的產生，以及美元持續走貶，將引起社會的動盪與不安，民眾為了保值，會將資金轉移至黃金等實質資產。因此原油、黃金與美元／英鎊外匯市場關係之研究為重要的課題 · 本文旨在探討原油、黃金與美元／英鎊外匯市場的機率分配、週間效應及動態條件相關，提供投資人及經理人於多角化投資組合管理、風險配置、避險策略、資產配置決策之參考。本研究共分為三個子議題：第一篇為「原油、黃金與美元／英鎊外匯市場：風險與報酬之機率分配」，本研究利用機率分配方法探討原油、黃金與美元／英鎊外匯市場風險與報酬之統計特性，利用高斯分配配適報酬復，可以得到報酬高低依序為原油、黃金、美元／英鎊外匯市場；利用對數常態分配配適風險復，可以發現美元／英鎊外匯市場波動最小，而原油市場波動最為劇烈。第二篇為「原油、黃金與美元／英鎊外匯市場是否存在週間效應？機率分配方法之應用」，本研究利用機率分配方法探討原油、黃金與美元／英鎊外匯市場是否存在週間效應，研究發現原油市場投資人縮短投資期間（從星期三到星期五）以規避風險；黃金市場投資人延長持有期間（從星期四到下星期斗）以提高獲利；而美元／英鎊外匯市場星期四（星期一）為一週當中報酬最高（低）的一天。第三篇為「原油、黃金與美元／英鎊外匯市場：動態條件相關與風險門檻值之機率分配」，本研究應用 Engle (2002)提出動態條件相關多變量 GARCH 模型探討原油、黃金與美元／英鎊外匯市場之關係，研究發現相關係數與時俱變的現象；市場之相關性及風險值會隨著波動性之提高而上升；以及高、低波動天期之動態條件相關之機率分配存在明顯差異。本研究應用 Kasch and CaPorin (2007)提出風險門檻值動態條件相關模型，研究發現原油與黃金市場存在風險門檻， 1990 年波灣戰爭與 2001 年 911 恐怖攻擊事件。

關鍵字

原油市場 ； 黃金市場 ； 美元／英鎊外匯市場 ； 機率分配 ； 週間效應 ； 動態條件相關

並列摘要

Oil prices remain an important determinant of global economic performance. The oil
prices have climbed up steadily recently, and it has not only shocked the crude oil market, but
also influenced the financial markets. Owing to the fear of inflation brought on by higher
crude oil, and the depreciation in the value of the US dollar, more and more investors put
money into gold market. Thus the investigation of the relationship among the crude oil, gold
The third thesis named as “The Realized Distributions of Dynamic Conditional
Correlation and Volatility Thresholds among Crude Oil, Gold, and US Dollar/ British Pound
Markets” applies Engle (2002) dynamic conditional correlation model to analyze the
relationship among the crude oil, gold and US Dollar/ British Pound markets. Both the timevarying
correlations and high contemporaneous dynamic conditional correlation across
volatilities are explored, and some differences between low volatility days and high volatility
days in realized distributions are found. Applying Kasch and Caporin (2007) volatility
threshold dynamic conditional correlation model, we find that the volatility thresholds, the
First Gulf War in 1990 and the 911 terror attack in 2001.

and US Dollar/ British Pound market is an important issue. This paper investigates probability
distributions, weekday effect, and dynamic conditional correlation among the crude oil, gold
and US Dollar/ British Pound market. All these findings are important to market traders and
hedging strategies, these have important implications for international investors, multinational
firms, and risk managers and so on. They consider the impact of commodity return and
volatility on portfolio diversification and on management, risk assessment, pricing and
hedging, asset allocation decisions. This doctoral dissertation is divided into three parts.
The first thesis named as “Probability Distribution of Return and Volatility among Crude
Oil, Gold, and US Dollar/ British Pound Markets” applies new methodology of probability
distributions to analyze the statistical properties of daily returns and volatility in crude oil,
gold and US Dollar/ British Pound market. After fitting the data into probability distributions
and estimating the parameters of the Gaussian distribution, we find that crude oil market
shows the highest return, followed by gold market and US Dollar/ British Pound market.
After estimating the peak and width of the volatility of the log-normal distribution, the US
Dollar/ British Pound market seems to have the least volatility in the log normal distribution,
and the crude oil market is the most unstable and volatile market.
The second thesis named as “Does Weekday Effect Exist in Crude Oil, Gold, and US
Dollar/ British Pound Markets? An Analysis from Probability Distribution Approach”
reshapes the data into a panel style from Monday to Friday to investigate whether or not the
weekday effect exists in the intraday return of crude oil, gold, and US Dollar/ British Pound
market. We document traditional weekend effect had been lengthened (from Thursday to the
next Tuesday) in gold market. This weekend effect disappears given that the distribution
shifted leftward in oil market; instead, the weekday effect (from Wednesday to Friday) seems
to appear significantly. As for the US Dollar/ British Pound market the lowest negative
returns appear on Mondays, while the highest returns appear on Thursdays rather than Fridays.

並列關鍵字

Crude Oil Market ； Gold Market ； US Dollar/ British Pound Market ； Probability Distribution ； Weekday Effect ； Dynamic Conditional Correlation

參考文獻

realized stock return volatility, Journal of Financial Economics, 61, 43-76.

realized exchange rate volatility, Journal of the American Statistical Association, 96, 453,

3. Becker, K. G., J. E. Finnerty, and M. Gupta, 1990, The intertemporal relation between

4. Bollerslev, T., 1986, Generalized Autoregressive conditional heteroscedasticity, Journal

of Econometrics, 31, 307-327.