Abstract: Crude oil is a commodity with both strategic and financial attributes, and its financial attribute affects its price fluctuation to the commodity futures market inevitably. Scholars through the world have studied the impact of crude oil on non energy markets such as food and agricultural products. However, little literature has been done to study the risk conduction relationship between international crude oil and Chinese commodity futures market from risk view.
This paper studies the influence of international crude oil price fluctuation on commodity futures in different historical stages, using the correlation structure breakpoint method proposed by Wied and Kramer(2012, Econometric Theory). The price trend is divided into three historical stages of international crude oil's impact on commodity futures. And the model of risk conduction proposed by Adam and Gluck(2015, Journal of Banking & Finance) is used to study the impact of international crude oil price fluctuation on Chinese commodity futures.
The empirical analysis of three types representative commodity futures, such as rubber, copper and agricultural products from January 2001 to May 2017, is selected. On one hand, the influence of international crude oil to these commodities is clearly divided into three sections by using the correlation structure breakpoint method. The results are as follows:in the first section from 2001-2008, the correlation is maintained below 0.05; in the second section from 2008-2014, the correlation increases to more than 0.22; while in the third section from 2014-2017, the correlation falls below 0.06. The correlation between China's commodity futures and international crude oil has been found cyclical, with a cycle of about seven years. On the other hand, the risk conduction model is used to study the impact of international crude oil on the three commodity futures in three historical sections. The result shows that the correlation between the period of 2008-2014 is easily founded; from the perspective of risk conduction, effect of three futures and the international crude oil have obvious difference and regularity, especially in the high correlation period, the risk conduction trend from high to low showing a "ladder" change, similarly. The conclusion above is not only beneficial to the entity operators to recognize the possible risks from the macro perspective, but also to provide a useful policy basis for the financial supervision department.

Key words: crude oil, commodity futures, correlation structure breakpoint, VaR-quantile regression