Analysis Of Asset Pricing Model With Time Varying Price Regression Rate And Extrapolation Strength

The traditional financial model agrees that traders are heterogeneous.According to the price forecasting rules,they are generally divided into fundamentals and technical analysts,and the main parameters(price regression speed and trend extrapolation inten-sity)are fixed,when the two types of traders predict the price of the next period.In this paper,we think that the main parameters are time-varying,and then establish the asset pricing model with time-varying price regression speed and extrapolation intensity.We use the difference equation theory to analyze the dynamic system,get the equilibrium solution,local asymptotically stable region and branch boundary.We take parameters near the stable region and the Neimark branch to simulate and discuss the different char-acteristics of price fluctuation and return autocorrelation structure.At the end of this paper,we first take the Shenzhen composite index as an empirical sample,and based on the Monte Carlo simulation,compare and analyze the volatility aggregation,asymmetry and long memory of the real market,this model and the original model.It shows that the model can fit the real market well,and it is better than the original model in fitting long memory.Then we take the Shenzhen A shares as an empirical sample,compare and analyze the non normality,stability and autocorrelation of the Shenzhen A shares,the model and the original model,indicating the validity and the optimality of the model.