A Research On Return And Volatility Spillover Effects Between Futures And Spot Markets Of Chinese Commodity Markets

With the development of Chinese commodity markets,more and more commodities become available to commodity futures investors.As for December 31th 2014,there has already been 44 kinds of commodities listed on three major Chinese Commodity Exchanges,namely Shanghai Commodity Exchange,Zhengzhou Commodity Exchange and Dalian Commodity Exchange.Besides,the recent burst of stock market bubble in the middle of 2015 makes commodity futures market more appealing to institutional investors.As a result,in recent years there seems to be a large inflow of capital to commodity futures markets,making futures price volatile.It is also well documented that there exists co-movement among different commodities.This is referred to as the "financialization" of commodity futures market,which has sparked intense debated among researchers and policy makers,arguing whether "speculation" in the commodity futures markets could distort commodity spot prices causing economic instability.With the'financializaiton" of commodity futures market view,the topic of "the relationship between commodity futures price and spot price" is worth looking carefully into.In this paper,24 actively traded commodity futures markets and spot markets price data are used to analyze the relationship between commodity futures price and spot price by implementing two econometric methods,namely Vector Error Correction Model(VECM)and Vector Autoregressive Model(VAR).This paper finds that as for the return relationship between commodity futures market and spot market,three kinds of commodities display bi-directional influence,meaning futures market return affects spot market return at the same time spot market return affects futures market return.These three kinds of commodities are steels(rolled bar and hot rolled coils),metals(copper,aluminum and nickel)and soft commodity(sugar).Besides,unilateral influence from futures market to spot market is found among soy beans,soy oil,rapeseed oil and palm oil,indicating new information being reflected in the futures markets first then transfers to spot markets.As for return volatility relationship or volatility spillover effect,the squared residuals from the two models are used as proxies for return volatility.By implementing BEKK-GARCH(1,1)to the return volatility,this paper finds that there exists bi-lateral influence among steel(rolled bar and hot rolled coil),oil(soy oil,rapeseed oil and palm oil)and soft commodity(sugar).Although there is a strong bi-lateral influence in metals market in terms of return relationship,this paper only finds unilateral influent in these market in terms of volatility spillover effect.Although for oils(soy oil,rapeseed oil and palm oil)only unilateral influence of futures market to spot market is found in terms of return relationship,bilateral influence is found with volatility spillover effects.Finally,for flour unilateral influence of spot market to futures market is found both for return and volatility relationships.Different from previous papers on the relationship between the commodity futures markets and spot markets,this paper stands out in the following aspects:Firstly,the relationship between the two markets is tested on a large scale of commodity kinds.The original data of this paper contains futures and spot price of 41 kinds of commodities.After teasing out the non-actively traded commodities,the return and volatility spillover effects between spot and futures market for each commodity is tested.Non-actively traded commodities is defined as no futures price change in consecutive seven or seven more days.Secondly,previous studies use the squared residual of VAR model as a proxy for return volatility.In this paper,the squared residual of VECM as a proxy for return volatility is used.The benefit of this measure is that it excludes the long-term price relation between the spot and futures markets,yielding a more "pure" volatility measure.Finally,VAR is used as a robustness check,the results by using VAR are similar to the results using VECM,suggesting the robustness of this paper.