Change Point Identification And Impact Analysis Of Economic Growth Data

The quantile regression model has good robustness for data fitting with outliers and heavy-tailed distribution characteristics.Moreover,in reality,the coupling relationship of variables is often not linear,making the discussion of change-point quantile regression a practical work.This paper adopts the continuous piecewise linear quantile regression model at the change point,and fits the nonlinear relationship through piecewise linearity,which not only maintains the simplicity of the model,but also makes the model more explanatory.Furthermore,the change point quantile model parameters are estimated using the bootstrapped restart iterative segmented quantile algorithm,which makes the explanatory power of the model parameters stronger with a small increase in the amount of computation.The number of change points is estimated using the reverse elimination algorithm.By iterating the number of change points,the optimal model under the Bayesian information criterion is found.In addition,on the basis of the selection consistency and limit distribution theory of the iterative piecewise quantile algorithm,the change point effect test and the confidence interval of the change point parameter are investigated.In the empirical research,this paper uses the change point quantile regression model to analyze the differential characteristics of economic growth.First,this paper focuses on the heterogeneity-promoting role of capital factor levels in the process of economic growth.Based on the Cobb-Douglas production function and the neoclassical economic growth theory,the change-point quantile regression problem with per capita capital stock as the independent variable and per capita output as the dependent variable is studied.Using the actual macroeconomic data of major countries in the world,we conduct change point identification and quantile regression analysis on the 0.1,0.3,0.5,0.7,and 0.9 quantiles respectively.The results show that there are change points in other quantiles except the median,and there are obvious differences in the coefficients of each quantile regression model.This chapter also performs quantile regression without change points on the 0.1,0.3,0.5,0.7,and 0.9 quantiles respectively.When comparing the quantile regression model without change points and the quantile regression model with change points,it is found that Quantile models with change points describe the data more delicately,and have stronger explanatory power to the reality behind the data.Further consider the differences in the impact of technological progress on economic growth.In Chapter 4,the technological progress rate is regarded as a variable in the production function,and the difference between the economic growth rate,the labor growth rate,and the technological progress rate is used as the quantile of the change point The response variable of the numerical regression,the difference between the capital growth rate and the labor growth rate is used as the independent variable of the change point quantile regression,and the change point quantile regression analysis is performed on the 0.1,0.3,0.5,0.7,and 0.9 quantiles respectively.The results show that there are change points in all quantiles except the 0.1 quantile,and there are obvious differences in the coefficients of each quantile regression model.At the same time,it is also found that the main reason for the data to fall into a certain quantile is the rate of technological progress.When a country has a relatively stable rate of technological progress,the data will stably fall near a certain quantile.Selecting some of these countries and analyzing the data and economic development,it is found that the model has a good explanatory power to the economic reality behind the data.