The Correlation Analysis Between The Price Of Oil And The US Dollar Index Based On Copulas

The Copula function can be used to satisfy the two conditions which are not linear structure and monotone increasing,This kind of function is aimed at the single variable edge step function and the multi variable joint distribution function, which can effectively solve some of the limitations of the past in the analysis of such data,One of the more obvious characteristics is to describe the correlation between elements and their correlation structure and other aspects of the problem.The historical data of the international crude oil price and the US dollar index from January 3,2013 to December 31,2014 are selected, in the theoretical framework of stationary time series and Copula function, establish Copula model to analyze the correlation between international crude oil prices and the dollar index. The research work of this paper is divided into the following two steps:first, the international crude oil price and the dollar index of the edge distribution, and the stability of the autocorrelation and unit root test; Secondly, select the suitable Copula function, the correlation structure fluctuation between specific description of international crude oil prices and the dollar index. Consequently, this paper first uses the Eviews software to determine the edge of the international crude oil prices and the dollar index distribution, then the appropriate processing and analysis of original data, selects the GARCH model to estimate the marginal distribution of crude oil prices and the dollar index yields, and the fitting result of the evaluation model; In the selection of Copula function, this paper starts with five commonly used Copula functions:the bivariate normal copulas connect, binary t-copula, Frank copula, Clayton copula and Gumbel copula. Firstly, we estimate parameters related to each function and the degree of freedom, and then found out the various copula density function value and distribution function value, and draw out the distribution function and the density function graph. Finally, rounding is not suitable for the copula function. Through a comprehensive comparative analysis, we selected binary t-copula, Frank copula to describe the correlation of crude oil prices and the dollar index structure, calculate the Euclidean distance between oil prices and the dollar index, and finally draw the conclusion:binary t-copula is better than Frank copula to fitting effect.In the part of the summary and outlook of this article, we discussed the practical significance and research contents of a number of shortcomings, and for investors in the firm operation to provide some practical advice, remind investors pay attention to some of the price of crude oil greatly change, so as to avoid danger.