A discrete microstructure model based modeling and control method for financial markets

Based on the market microstructure theory and continuous time stochastic nonlinear microstructure model, a discrete time microstructure model is proposed for describing the dynamics of stochastic volatile financial markets. From the discrete microstructure model proposed one can obtain the estimates of two immeasurable state variables representing the market excess demand and liquidity respectively. Based on the estimated excess demand information instead of the prediction of price, one may more effectively control the assets dynamic allocation process. The estimation procedure of the discrete microstructure model using the nonlinear Kalman filter and maximum likelihood method is also presented. A case study on the Japanese Yen/US dollar exchange rate time series modeling and the estimated model-based assets dynamic allocation control shows satisfactory modeling precision and control performance.