News Arrival And Jump Dynamics With CBP-GARCH- In Case Of Gold and Oil

• Main Authors: Chen,Hsuehping, 陳學平
• Other Authors: Chiou,Jershiou
• Format: Others
• Language: en_US
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/10654641712762256726

碩士 === 實踐大學 === 財務金融與保險研究所 === 100 === Although gold and crude oil are widely recognized as the two primary elements of the large commodity markets, it is clear that the prices of these two commodities are not completely driven by the most basic market considerations of supply and demand. Indeed, a considerable body of evidence has emerged to suggest the existence of several financial features of the international gold and crude oil markets, as well as the existence of close interactions between the two markets. Despite the considerable wealth of research which has already been undertaken into price dynamics within these two markets, resulting largely in findings of a high correlation between their price movements, very little research appears to have been undertaken into the interactions occurring between the two markets. Furthermore, within the majority of the prior studies focusing on this area, it is traditionally assumed that time-series data follow a smooth and continuous process; however the presence of jumps implies that the statistical specifications of diffusion models can be somewhat erroneous. Accordingly, the present study distinguishes itself from many of the prior studies, insofar as the analyses undertaken in this study utilize a ‘correlated bivariate Poisson-GARCH’ (CBP-GARCH) model, with the focus being placed on the interrelationships between the gold and crude oil markets, whilst also taking into consideration volatility levels within the crude oil market. The results obtained from the CBP-GARCH model reveal no significant correlations between the size of jumps in gold and crude oil returns; in other words, with the occurrence of any substantial jumps in gold returns, there will not necessarily be any simultaneous duplication of such jumps in oil returns. This phenomenon is also reflected in the conditional variances of the CBP-GARCH model. When the variances in returns and jumps are combined, the results reveal that the jump variance provides an appropriate reflection of the trend in returns. In summary, we find that the relationship between the volatility of gold and crude oil is clearly changing over time, since we can now determine that the jump variance in crude oil is much smoother than that for gold, particularly over the more recent periods.