Simulation of Transaction Behavior and Price Volatility in Chinese Soybean Futures Market Using MVAR Model

Traditional vector autoregressive (VAR) modeling theory has the defect that it can not effectively utilize the multiple time scale information contained in the inner of variables. In order to discuss multiscale behavior among economic variables and capture variables’ information in different time scale, multiresolution VAR model which can also be called as MVAR model has been established in the paper by combining multiscale analysis and theory of VAR modeling to overcome the defect of traditional model, which can also capture the relationship between variables in different time scale in detail. Taking soybean futures on Dalian Commodity Exchange for example, the paper studies the relationship between transaction behavior of agricultural futures investors and volatility of futures price using MVAR model.