Comparison And Empirical Study Of The Linear And Semi Parametric Regression Models Of The Factors Affecting China's Foreign Exchange Reserves

There are many factors that affect the scale of China's foreign exchange reserves.There are always some errors in the application of linear models in practice,but whether the nonparametric regression models belong to the assumed variables need to be studied.All the regression functions are predicted,thus the more suitable sample parameters can be obtained,while the prediction of the data is more accurate.The semi parametric model not only integrates the advantages of the nonparametric model,but also integrates the advantages of the linear model,which has strong explanatory power and can avoid the problem of Curse of dimensionality.Therefore,the semi parametric model has been paid more and more attention by the scholars.In recent years,the linear regression method is used mostly China foreign exchange reserve influence study,semi parametric methods in research and application of the scale of foreign exchange reserves in the factors are rare,and for further research on the kernel estimation method is to be in effect factors of foreign exchange reserves.This paper sets up a linear regression model and a semi parametric regression model of the factors affecting China's foreign exchange reserves from two different aspects.First,the time series analysis theory of the impact of foreign exchange reserves China factors to build a linear regression model based on linear regression and significant lagged variables and the theory of kernel estimation based on a semi parametric regression equation,these two models are Chinese effect on foreign exchange reserves fitting and comparison,the result shows that the latter is better than the semi parametric regression model the linear regression model.secondly,the establishment of FER to the scale of China's foreign exchange reserves as explanatory variables,to import and export balance of XM,DEB,foreign debt RMB dollar exchange rate average gross domestic product EER,GDP as a linear regression model,and considering the Chinese foreign exchange reserve and the consumer price index is highly CPI correlation,regression model and consumer price index to CPI for semiparametric exogenous variables,the two comparison.Found that annual import and export the difference,the yuan against the dollar exchange rate,gross domestic product(GDP)can stimulate the growth of foreign exchange reserves,while inhibiting the growth of foreign exchange reserves,foreign debt balance,the growth rate of import and export the difference changes a unit will cause bigger 1.738% growth in foreign exchange reserves,the yuan against the dollar change a unit annualrate,foreign exchange reserves growth rate 5.76 times bigger,gross domestic product(GDP)growth rate changes by one unit,foreign exchange reserves growth rate 2.25 times bigger,a unit and growth rate of foreign debt balance changes,can cause foreign exchange reserves growth rate by 0.155 times.The results show is the establishment of nonparametric function for semiparametric regression model of consumer price index CPI has more advantages.