Describing And Forecasting Exchange Rate Behavior Based On Order Flow Analysis In Information Heterogeneity Environment

Exchange rate theories are always the core of the international finance, which cover various complex problems. Since 70's of last century, floating exchange rate system has substituted the fixed exchange rate system. The volatility of exchange rate not only affects the behalf of microcosmic main body but also relates to the equilibrium of national economy. The determination of exchange rate is always the hotspot problem .The character of current exchange rate theories is offering a suitable explain of determination of exchange rate. While when we value whether the trend of exchange rate of mid-short term is in reason or not, we meet the determination puzzle which was brought forward by Meese and Rogoff. It incarnates that macro-economic variables cannot explain the fluctuation of exchange rate very well, and we are lack of a fine model which can behave excellent in empirical results. Economic scholars transfer the eyes to the behavior of participants and the microstructure of exchange rate. It presents that we can analyze the proceeds of asset price on the premise of non-efficient market. This paper will use the viewpoint of financial microstructure theory and investigate exchange rate behavior via analyzing order flows.This paper firstly reviews three of the representative traditional exchange rate theories and analyzes the character of the theories. After that the paper continues to analyze the shorts of them and discusses the direction to look for a new method. Then this paper introduces the viewpoint of financial microstructure theory and some important concepts. Based on the theory, the paper puts forward the solution method of information heterogeneity circumstance and the portfolio shift model with order flow. The paper finds that after considering the inter-bank order flow it can improve the model's veracity. Lastly, the paper compares the forecasting results of our model with that of Random Walk. The results proved that our model performs better than Random Walk. Empirical results present that order flow analysis is a good way to describe the volatility of mid-short exchange rate and financial microstructure market plays an important role on explaining the behavior of exchange rate.